Cubois in the Enterprise Collaboration Industry

Cubo

cubo

Recently, I gave a talk at in a panel called in front of around 100 university students, entrepreneurs, and startup founders. The panel itself was facilitated by the wonderful , Assistant Professor at UC Boulder, , and co-author of the paper . Boulder Startup Week Exit to Community, Journey of 2 Startups Nathan Schneider Hacker Noon contributor "Exit To Community: Strategies for Multi-Stakeholder Ownership in the Platform Economy" You can watch (my presentation starts at around 11:15), or read the transcript (note-taking style) below. Learn more about via Nathan's article on Hacker Noon . the talk here Exit to Community here Framing of Session: As founders, we build startups to serve communities, and as users, we value the startups that help us, so in 2019 we asked: what if communities become owners of the startups? What if founders centered users in their exit strategy? An exit strategy is how we sell and transition out of business. The two dominant options for startups are going public or getting acquired, but both exits benefit investors in ways that leave everyone else out of decisions about the startup they depend on. We decided to organize around a third option, . an Exit To Community But a bold idea leads to many questions: How can we make finance, legal structures, and community leadership work together? Is the risk of a new exit strategy really worth it? Is it possible to make everyone happy? There is no one clear E2C approach. Startups are trying to build something sustainable, and they’ve stumbled or jumped into this idea in a dozen different ways. This session is part of a long series -- 7 webinars on specific topics, a 10-week cohort -- but this is the first time we hear from startups making the road by walking, figuring out how to make an exit to the community in their own way. Presentations Collaborative Notes : Usually, I wouldn’t talk about my own education in this kind of blurb, but it feels relevant. I went to this idealistic school called United World College for 2 years at 16 and basically believed that it’s possible to change the world for the better by attending the same environment with 200 students from almost 100 different countries. Then I went to Brown, majoring in international development and education, liberal af, as they say. About Linh Dao Smooke I’m a business woman now and a lot more skeptical about the state of the world but I think in my heart of hearts I'm still that 16-year-old Vietnamese who wants to believe that capitalism is not and should not be the future of humanity. Why I think it’s relevant to mention my past: My high school friend , who I haven't spoken to in over a decade, actually co-authored the with Nathan about the 3 options startups could use to successfully funnel ownership from greedy VCs to actual stakeholders. The universe works in such a strange, mysterious way! Morshed E2C paper Now on to the presentation... What is Hacker Noon? Hacker Noon publishes stories and writes software! We are a super active community- 30 published and 60+ submitted each day- 220 years worth of reading total. We also built our own platform for publishing, from scratch- which enabled us to remove Medium software. Our software could be used for licensing one day too. We don’t require anyone to log in or go through a paywall. We are profitable as of now and will be profitable by the end of the year, all thanks to a robust community and a large editorial team and collection of writers too! The Medium Fuckup (as in, they fuck us up) Our shift in 2019, Medium made a drastic move of bringing curators/publishers in and trying to remove the middle layer – that’s us – and keep the writers, a very effective bait & switch. We were one of the first collaterals of Medium's business shift (and it's not the first or last time they did it, as we soon later found out). My husband and I were proud of doing well, but they cut off our advertising model built over the years… and we felt betrayed. Importantly… we realized all of our writers were HN writers, they identified with us, and all we needed Medium for was the software. Our options: One - not sustain HN; Two raise some money from VCs and follow whatever direction they dictate; or Three: retain who we are – our writers – and build our own tech to move off of Medium. The community rallied behind us, and we raised over $1,000,000! That was the maximum we could raise under equity crowdfunding under the 2015 JOBS Act. The idea came from one of our contributing writers! The Transition We had a 6 months runway New software, lots of bugs, but we love it - lots of pain and anguish. Now, 11% of the company is owned by around 1,300 co-investors. Lessons Learned They already know what value they get from HN, and they care the most, so they deserve to keep getting it and caring for it. 1) The people who deserve to own the platform the most are the users who use the platform. We were approached by 4 VCs and bigger companies that in the short term would have made more sense financially, but in the long term, we won’t control our own destiny. Now my husband and I control ~70%; crowdfunding control 11%; 8% are owned by a strategic partner, Coil, which helped with the second round of funding. 2) Independently owned with less money is better than owned by more money with many strings attached. If E2C grows, people will figure out how to avoid “having too many cooks in the kitchen”. We don’t defer to users entirely! 3) Listen to community concerns and feedback without losing sight of your own direction. The Future We may continue to explore an E2C… perhaps tokenizing user activity? Learn more about Exit to Community here.

How Hacker Noon Leans on Our Community -  a #CUBoulder Talk by Linh Dao Smooke

Authors:
(1) Hamed Alimohammadzadeh, University of Southern California, Los Angeles, USA;
(2) Rohit Bernard, University of Southern California, Los Angeles, USA;
(3) Yang Chen, University of Southern California, Los Angeles, USA;
(4) Trung Phan, University of Southern California, Los Angeles, USA;
(5) Prashant Singh, University of Southern California, Los Angeles, USA;
(6) Shuqin Zhu, University of Southern California, Los Angeles, USA;
(7) Heather Culbertson, University of Southern California, Los Angeles, USA;
(8) Shahram Ghandeharizadeh, University of Southern California, Los Angeles, USA. Table of Links Abstract and Introduction Flying Light Specks User Interaction Multimedia Systems Challenges Related Work Conclusions and Current Efforts, Acknowledgments, and References ABSTRACT Today’s robotic laboratories for drones are housed in a large room. At times, they are the size of a warehouse. These spaces are typically equipped with permanent devices to localize the drones, e.g., Vicon Infrared cameras. Significant time is invested to fine-tune the localization apparatus to compute and control the position of the drones. One may use these laboratories to develop a 3D multimedia system with miniature sized drones configured with light sources. As an alternative, this brave new idea paper envisions shrinking these room-sized laboratories to the size of a cube or cuboid that sits on a desk and costs less than 10K dollars. The resulting Dronevision (DV) will be the size of a 1990s Television. In addition to light sources, its Flying Light Specks (FLSs) will be network-enabled drones with storage and processing capability to implement decentralized algorithms. The DV will include a localization technique to expedite development of 3D displays. It will act as a haptic interface for a user to interact with and manipulate the 3D virtual illuminations. It will empower an experimenter to design, implement, test, debug, and maintain software and hardware that realize novel algorithms in the comfort of their office without having to reserve a laboratory. In addition to enhancing productivity, it will improve safety of the experimenter by minimizing the likelihood of accidents. This paper introduces the concept of a DV, the research agenda one may pursue using this device, and our plans to realize one. 1 INTRODUCTION A Dronevision, DV, is a 3D display to design and implement next generation multimedia applications using Flying Light Specks, FLSs. An FLS is a miniature sized drone with one or more light sources with adjustable brightness [25–27]. A swarm of cooperating FLSs will render 3D shapes, point clouds, and animated sequences in a fixed volume, a 3D display. This Brave New Idea (BNI) envisions shrinking today’s room sized robotics laboratories to the size of a cuboid that fits on a user’s desk for multimedia content creation and rendering. Figure 1 shows a DV setup configured for visual display only. Four glass panes enhance the user’s safety by preventing rogue FLSs from coming to contact with a user and injuring them. Once the rendering of an illumination is verified to be safe, the user may remove the glass panes. The DV may illuminate the rose beyond the size of its base, see Figure 2. Moreover, a user may touch and interact with the illumination. For 3D media content, a DV is a low cost alternative to an expensive robotics laboratory shared by different research groups. It will provide an experimenter with a dedicated 3D testbed to debug and evaluate their hardware and software solutions. It will enable a content creator to design and implement diverse multimedia applications that range from healthcare to entertainment. With healthcare, a designer may develop a 3D FLS display to illuminate 3D MRI scans of a patient and their organs, empowering a physician to separate the different organs (illuminations) and examine them in real time. With entertainment, the designer may illuminate characters of multi-player games such as Minecraft and Fortnite to come alive and interact with one another on a table top. In addition to improving productivity, a DV enhances the safety of both the experimenter and the content creator. Figure 1 shows a DV displaying an illuminated rose with a falling petal; its original may be captured using an Intel RealSense depth camera[1]. The top panel of the DV consists of wireless charging stations. FLSs with a low remaining flight time fly through the square hole and land on a charging coil to charge their battery. Hundreds of FLSs may land on a charging coil and simultaneously charge their battery. The floor of the DV consists of a Hub, Tiles, Terminus, and a conveyor belt. Each rod on the side acts as an elevator that delivers fully-charged FLSs to the base of the DV, where they are stored on the black tiles for future deployment. The Hub is comparable to today’s servers and provides both wired and wireless connectivity to peripheral devices such as a depth camera and earbuds. Black and white tiles are modular pieces that can be plugged into other tiles to construct DVs of different lengths and depths. Black tiles serve both as a hangar and a dispatcher of FLSs, which are moved up and down using an elevator, and white tiles are rigid pieces that keep the floor together. Black tiles containing FLSs may be stacked below the white tiles, see Figure 1. When an empty black tile is pushed below a white tile, a black tile with FLSs is pushed onto the elevator to bring its FLSs to the floor of the DV; these FLSs are then dispatched to illuminate virtual objects. Failed FLSs will fall onto a conveyor belt, which is a garbage collector that deposits failed FLSs to a Terminus. The Terminus is an empty row on the DV floor, e.g., the row closest to the user (the back panel) in Figure 1 (2). The conveyor belt protects FLSs sitting on black tiles from falling FLSs that have failed. The Hub stops the conveyor belt before deploying one or more FLSs atop a black tile. A deployed FLS flies below the conveyor belt and past the Terminus to navigate to its coordinates in the DV’s display volume. Stopping the conveyor belt minimizes the likelihood of a failed FLS hitting and damaging a deployed FLS. Similarly, by placing the charging stations at the top of the DV, the likelihood of failed FLSs hitting and damaging a functional FLS that is in transit to a charging station is minimized. A DV will be both modular and configurable. Once an experimenter is satisfied with the safety of their setup in Figure 1, they may remove the glass panes and place the top as a back panel in front of the Terminus, enabling the DV to render illuminations that extend beyond the length and width of its base and allowing the user to physically interact with the illuminations (Figure 2). The haptic feedback provided by the illumination will include a combination of kinesthetic and tactile sensations. The kinesthetic (force) feedback will enable a user to experience the shape and hardness of an illuminated object. For example, a user may grasp a rose petal by its sides using their fingers. The user may press their fingers together to bend the rose petal to experience its stiffness. The tactile (skin-based) feedback will enable a user to experience the object’s surface properties. For example, FLSs with electrotactile actuators may provide the sharp sensation a user would expect when running their finger over a thorn on the virtual rose of Figure 2. The contributions of this paper include: • DV as an experimental 3D platform to design and implement next generation multimedia applications using FLSs (Sections 1 and 2). • Use of DV for haptic interactions (Section 3). • Two multimedia systems research challenges raised by a DV (Section 4). We present related work in Section 5 and brief conclusions and our current efforts to realize a DV in Section 6. This paper is available on arxiv under CC 4.0 license. [1] The use of a camera is for illustration purposes only. The shown rose is a mesh file consisting of 65K points created using Blender [3, 26]. Authors: (1) Hamed Alimohammadzadeh, University of Southern California, Los Angeles, USA; (2) Rohit Bernard, University of Southern California, Los Angeles, USA; (3) Yang Chen, University of Southern California, Los Angeles, USA; (4) Trung Phan, University of Southern California, Los Angeles, USA; (5) Prashant Singh, University of Southern California, Los Angeles, USA; (6) Shuqin Zhu, University of Southern California, Los Angeles, USA; (7) Heather Culbertson, University of Southern California, Los Angeles, USA; (8) Shahram Ghandeharizadeh, University of Southern California, Los Angeles, USA. Authors: Authors: (1) Hamed Alimohammadzadeh, University of Southern California, Los Angeles, USA; (2) Rohit Bernard, University of Southern California, Los Angeles, USA; (3) Yang Chen, University of Southern California, Los Angeles, USA; (4) Trung Phan, University of Southern California, Los Angeles, USA; (5) Prashant Singh, University of Southern California, Los Angeles, USA; (6) Shuqin Zhu, University of Southern California, Los Angeles, USA; (7) Heather Culbertson, University of Southern California, Los Angeles, USA; (8) Shahram Ghandeharizadeh, University of Southern California, Los Angeles, USA. Table of Links Abstract and Introduction Abstract and Introduction Flying Light Specks Flying Light Specks User Interaction User Interaction Multimedia Systems Challenges Multimedia Systems Challenges Related Work Related Work Conclusions and Current Efforts, Acknowledgments, and References Conclusions and Current Efforts, Acknowledgments, and References ABSTRACT Today’s robotic laboratories for drones are housed in a large room. At times, they are the size of a warehouse. These spaces are typically equipped with permanent devices to localize the drones, e.g., Vicon Infrared cameras. Significant time is invested to fine-tune the localization apparatus to compute and control the position of the drones. One may use these laboratories to develop a 3D multimedia system with miniature sized drones configured with light sources. As an alternative, this brave new idea paper envisions shrinking these room-sized laboratories to the size of a cube or cuboid that sits on a desk and costs less than 10K dollars. The resulting Dronevision (DV) will be the size of a 1990s Television. In addition to light sources, its Flying Light Specks (FLSs) will be network-enabled drones with storage and processing capability to implement decentralized algorithms. The DV will include a localization technique to expedite development of 3D displays. It will act as a haptic interface for a user to interact with and manipulate the 3D virtual illuminations. It will empower an experimenter to design, implement, test, debug, and maintain software and hardware that realize novel algorithms in the comfort of their office without having to reserve a laboratory. In addition to enhancing productivity, it will improve safety of the experimenter by minimizing the likelihood of accidents. This paper introduces the concept of a DV, the research agenda one may pursue using this device, and our plans to realize one. 1 INTRODUCTION A Dronevision, DV, is a 3D display to design and implement next generation multimedia applications using Flying Light Specks, FLSs. An FLS is a miniature sized drone with one or more light sources with adjustable brightness [25–27]. A swarm of cooperating FLSs will render 3D shapes, point clouds, and animated sequences in a fixed volume, a 3D display. This Brave New Idea (BNI) envisions shrinking today’s room sized robotics laboratories to the size of a cuboid that fits on a user’s desk for multimedia content creation and rendering. Figure 1 shows a DV setup configured for visual display only. Four glass panes enhance the user’s safety by preventing rogue FLSs from coming to contact with a user and injuring them. Once the rendering of an illumination is verified to be safe, the user may remove the glass panes. The DV may illuminate the rose beyond the size of its base, see Figure 2. Moreover, a user may touch and interact with the illumination. For 3D media content, a DV is a low cost alternative to an expensive robotics laboratory shared by different research groups. It will provide an experimenter with a dedicated 3D testbed to debug and evaluate their hardware and software solutions. It will enable a content creator to design and implement diverse multimedia applications that range from healthcare to entertainment. With healthcare, a designer may develop a 3D FLS display to illuminate 3D MRI scans of a patient and their organs, empowering a physician to separate the different organs (illuminations) and examine them in real time. With entertainment, the designer may illuminate characters of multi-player games such as Minecraft and Fortnite to come alive and interact with one another on a table top. In addition to improving productivity, a DV enhances the safety of both the experimenter and the content creator. Figure 1 shows a DV displaying an illuminated rose with a falling petal; its original may be captured using an Intel RealSense depth camera[1]. The top panel of the DV consists of wireless charging stations. FLSs with a low remaining flight time fly through the square hole and land on a charging coil to charge their battery. Hundreds of FLSs may land on a charging coil and simultaneously charge their battery. The floor of the DV consists of a Hub, Tiles, Terminus, and a conveyor belt. Each rod on the side acts as an elevator that delivers fully-charged FLSs to the base of the DV, where they are stored on the black tiles for future deployment. The Hub is comparable to today’s servers and provides both wired and wireless connectivity to peripheral devices such as a depth camera and earbuds. Black and white tiles are modular pieces that can be plugged into other tiles to construct DVs of different lengths and depths. Black tiles serve both as a hangar and a dispatcher of FLSs, which are moved up and down using an elevator, and white tiles are rigid pieces that keep the floor together. Black tiles containing FLSs may be stacked below the white tiles, see Figure 1. When an empty black tile is pushed below a white tile, a black tile with FLSs is pushed onto the elevator to bring its FLSs to the floor of the DV; these FLSs are then dispatched to illuminate virtual objects. Failed FLSs will fall onto a conveyor belt, which is a garbage collector that deposits failed FLSs to a Terminus. The Terminus is an empty row on the DV floor, e.g., the row closest to the user (the back panel) in Figure 1 (2). The conveyor belt protects FLSs sitting on black tiles from falling FLSs that have failed. The Hub stops the conveyor belt before deploying one or more FLSs atop a black tile. A deployed FLS flies below the conveyor belt and past the Terminus to navigate to its coordinates in the DV’s display volume. Stopping the conveyor belt minimizes the likelihood of a failed FLS hitting and damaging a deployed FLS. Similarly, by placing the charging stations at the top of the DV, the likelihood of failed FLSs hitting and damaging a functional FLS that is in transit to a charging station is minimized. A DV will be both modular and configurable. Once an experimenter is satisfied with the safety of their setup in Figure 1, they may remove the glass panes and place the top as a back panel in front of the Terminus, enabling the DV to render illuminations that extend beyond the length and width of its base and allowing the user to physically interact with the illuminations (Figure 2). The haptic feedback provided by the illumination will include a combination of kinesthetic and tactile sensations. The kinesthetic (force) feedback will enable a user to experience the shape and hardness of an illuminated object. For example, a user may grasp a rose petal by its sides using their fingers. The user may press their fingers together to bend the rose petal to experience its stiffness. The tactile (skin-based) feedback will enable a user to experience the object’s surface properties. For example, FLSs with electrotactile actuators may provide the sharp sensation a user would expect when running their finger over a thorn on the virtual rose of Figure 2. The contributions of this paper include: contributions • DV as an experimental 3D platform to design and implement next generation multimedia applications using FLSs (Sections 1 and 2). • Use of DV for haptic interactions (Section 3). • Two multimedia systems research challenges raised by a DV (Section 4). We present related work in Section 5 and brief conclusions and our current efforts to realize a DV in Section 6. This paper is available on arxiv under CC 4.0 license. This paper is available on arxiv under CC 4.0 license. available on arxiv [1] The use of a camera is for illustration purposes only. The shown rose is a mesh file consisting of 65K points created using Blender [3, 26].

Dronevision: An Experimental 3D Testbed for Flying Light Specks

Authors:
(1) Ritesh Sharma, University Of California, Merced, USA rsharma39@ucmerced.edu;
(2) Eric Bier, Palo Alto Research Center, USA bier@parc.com;
(3) Lester Nelson, Palo Alto Research Center, USA lnelson@parc.com;
(4) Mahabir Bhandari, Oak Ridge National Laboratory, USA bhandarims@ornl.gov;
(5) Niraj Kunwar, Oak Ridge National Laboratory, USA kunwarn1@ornl.gov. Table of Links Abstract and Intro Related Work Methodology Experiments Conclusion & Future work and References 2 Related Work Floor plans are crucial for many applications. Software approaches to floor plan creation depend on data availability and data format. Our work builds on previous research on data collection and floor plan computation. Data collection. Indoor environments can be captured in many formats, including RGBD images, point clouds, and triangle meshes. Zhang et al. [29] uses panoramic RGBD images as input and reconstructs geometry using structure grammars, while [21] uses a 3D scan to extract plane primitives and generates models using heuristics. A single image is used in some deep learning methods [8,11,13,14,30] to generate cuboid-based layouts for a single room. Detailed semi-constrained floor plan computations for a complete house require processing a 3D scan of the house [15]; the complete scan increases accuracy, but increases computing requirements and time. Pintore and Gobbetti [23] proposed a technique to create floor plans and 3D models using an Android device camera, leveraging sensor data and statistical techniques. Chen et al. [7] introduced an augmented reality system utilizing the Microsoft Hololens for indoor layout assessment, addressing intuitive evaluation and efficiency challenges. In our approach, we begin with a triangle mesh from a HoloLens 2, using its Spatial Mapping software [18], which has been surveyed by Weinmann et al. [27]. Floor plan computations Early methods [1,3,22,28] relied on image processing techniques, such as histograms and plane fitting, to create floor plans from 3D data. While [22] creates a floor plan by detecting vertical planes in a 3D point cloud, [3] uses planar structure extraction to create floor plans. These techniques rely on heuristics and were prone to failure due to noise in the data. There has been much progress in floor plan computation using graphical models [4,9,12]. Such models [10] are also used to recover layouts and floor plans from crowd-sourced images and location data. One interactive tool [16] creates desirable floorplans by conforming to design constraints. Pintore et al. [24] characterizes several available input sources (including the triangle meshes that we use) and output models and discusses the main elements of the reconstruction pipeline. It also identifies several systems for producing floor plans, including FloorNet [15], and Floor-SP [6]. Monszpart et al. [19] introduced an algorithm that exploits the observation that distant walls are generally parallel to identify dominant wall directions using k-means. Our approach also utilizes k-means, but does so to identify walls in all directions, not just the dominant ones. Cai et al. [5] uses geometric priors, including point density, indoor area recognition, and normal information, to reconstruct floorplans from raw point clouds. In contrast to Arikan et al. [2], which employed a greedy algorithm to find plane normal directions and fit planes to points with help from user interaction, our approach is automatic. It also differs from the work of [20], which focuses on removing clutter and partitioning the interior into a 3D cell complex; our method specifically divides the building into separate walls. Our work is related to [22] and [26]. In [22], floor plan generation starts with a laser range data point cloud, followed by floor and ceiling detection using a height histogram. The remaining points are projected onto a ground plane, where a density histogram and Hough transform are applied to generate the line segments that form a floor plan. In projecting to 2D, their method risks losing information that may be useful for creating 3D models or detailed floor plans. Similarly, [26] utilizes a histogram-based approach to detect ceilings and floors. Their method involves identifying taller wall segments to create a 2D histogram, and then employing heuristics based on histogram point density to compute the floor plan. Our approach differs from [22] and [26] by aligning the mesh with global coordinate axes and not relying on laser data or a point cloud. Working primarily with 3D data throughout the pipeline, it benefits from enhanced information and generates both a 3D model and a floor plan. This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. Authors: (1) Ritesh Sharma, University Of California, Merced, USA rsharma39@ucmerced.edu; (2) Eric Bier, Palo Alto Research Center, USA bier@parc.com; (3) Lester Nelson, Palo Alto Research Center, USA lnelson@parc.com; (4) Mahabir Bhandari, Oak Ridge National Laboratory, USA bhandarims@ornl.gov; (5) Niraj Kunwar, Oak Ridge National Laboratory, USA kunwarn1@ornl.gov. Authors: Authors: (1) Ritesh Sharma, University Of California, Merced, USA rsharma39@ucmerced.edu; (2) Eric Bier, Palo Alto Research Center, USA bier@parc.com; (3) Lester Nelson, Palo Alto Research Center, USA lnelson@parc.com; (4) Mahabir Bhandari, Oak Ridge National Laboratory, USA bhandarims@ornl.gov; (5) Niraj Kunwar, Oak Ridge National Laboratory, USA kunwarn1@ornl.gov. Table of Links Abstract and Intro Abstract and Intro Related Work Related Work Methodology Methodology Experiments Experiments Conclusion & Future work and References Conclusion & Future work and References 2 Related Work Floor plans are crucial for many applications. Software approaches to floor plan creation depend on data availability and data format. Our work builds on previous research on data collection and floor plan computation. Data collection. Indoor environments can be captured in many formats, including RGBD images, point clouds, and triangle meshes. Zhang et al. [29] uses panoramic RGBD images as input and reconstructs geometry using structure grammars, while [21] uses a 3D scan to extract plane primitives and generates models using heuristics. A single image is used in some deep learning methods [8,11,13,14,30] to generate cuboid-based layouts for a single room. Detailed semi-constrained floor plan computations for a complete house require processing a 3D scan of the house [15]; the complete scan increases accuracy, but increases computing requirements and time. Pintore and Gobbetti [23] proposed a technique to create floor plans and 3D models using an Android device camera, leveraging sensor data and statistical techniques. Chen et al. [7] introduced an augmented reality system utilizing the Microsoft Hololens for indoor layout assessment, addressing intuitive evaluation and efficiency challenges. In our approach, we begin with a triangle mesh from a HoloLens 2, using its Spatial Mapping software [18], which has been surveyed by Weinmann et al. [27]. Data collection. Floor plan computations Early methods [1,3,22,28] relied on image processing techniques, such as histograms and plane fitting, to create floor plans from 3D data. While [22] creates a floor plan by detecting vertical planes in a 3D point cloud, [3] uses planar structure extraction to create floor plans. These techniques rely on heuristics and were prone to failure due to noise in the data. Floor plan computations There has been much progress in floor plan computation using graphical models [4,9,12]. Such models [10] are also used to recover layouts and floor plans from crowd-sourced images and location data. One interactive tool [16] creates desirable floorplans by conforming to design constraints. Pintore et al. [24] characterizes several available input sources (including the triangle meshes that we use) and output models and discusses the main elements of the reconstruction pipeline. It also identifies several systems for producing floor plans, including FloorNet [15], and Floor-SP [6]. Monszpart et al. [19] introduced an algorithm that exploits the observation that distant walls are generally parallel to identify dominant wall directions using k-means. Our approach also utilizes k-means, but does so to identify walls in all directions, not just the dominant ones. Cai et al. [5] uses geometric priors, including point density, indoor area recognition, and normal information, to reconstruct floorplans from raw point clouds. In contrast to Arikan et al. [2], which employed a greedy algorithm to find plane normal directions and fit planes to points with help from user interaction, our approach is automatic. It also differs from the work of [20], which focuses on removing clutter and partitioning the interior into a 3D cell complex; our method specifically divides the building into separate walls. Our work is related to [22] and [26]. In [22], floor plan generation starts with a laser range data point cloud, followed by floor and ceiling detection using a height histogram. The remaining points are projected onto a ground plane, where a density histogram and Hough transform are applied to generate the line segments that form a floor plan. In projecting to 2D, their method risks losing information that may be useful for creating 3D models or detailed floor plans. Similarly, [26] utilizes a histogram-based approach to detect ceilings and floors. Their method involves identifying taller wall segments to create a 2D histogram, and then employing heuristics based on histogram point density to compute the floor plan. Our approach differs from [22] and [26] by aligning the mesh with global coordinate axes and not relying on laser data or a point cloud. Working primarily with 3D data throughout the pipeline, it benefits from enhanced information and generates both a 3D model and a floor plan. This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. available on arxiv

Using Scanned Mesh Data for Auto-Digitized 3D Modeling: Related Work

Let's learn about via these 33 free stories. They are ordered by most time reading created on HackerNoon. Visit the to find the most read stories about any technology. Image Recognition /Learn Repo 1. Using Image Classification for Fitness and Dieting Apps Body management is desired yet hard. Knowing how much the daily take-in is can also be a challenge. Read on to use ML tech to overcome these challenges. 2. How Computer Vision Turns Images Into Arrays How Images are turned into arrays in Computer Vision 3. Universal Data Tool Introduction: Weekly Update 2 If you haven’t heard of the Universal Data Tool, it’s an open-source web or desktop program to collaborate, build and edit text, image, video, and audio datasets with labels and annotations. 4. An Intro to AI Image Recognition and Image Generation The ability of artificial intelligence to recognize and generate images or even videos is assisting businesses to enhance the speed and quality of their work. 5. How AI Is Helping To Diagnose Cancer Read how artificial intelligence is used in digital pathology for cancer diagnosis and reveal MobiDev’s research on WSI processing using machine learning models 6. This Is How Easy it Is to Read and Display Images Working with images is more and more popularized; we are going to take dive into this field, first with the basics and starting with an open CV library. 7. Small Object Detection in Computer Vision: The Patch-Based Approach How to carry out small object detection with Computer Vision - An example of finding lost people in a forest. 8. 10 Best Image Classification Datasets for ML Projects To help you build object recognition models, scene recognition models, and more, we’ve compiled a list of the best image classification datasets. These datasets vary in scope and magnitude and can suit a variety of use cases. Furthermore, the datasets have been divided into the following categories: medical imaging, agriculture & scene recognition, and others. 9. What Makes "Good" Fashion Image Tagging Fashion image tagging is infamously tedious for eCommerce. But, how can AI help create accurate tags--and go a step beyond in understanding fashion information? 10. How to Build Neural Network that Recognizes People Wearing Masks CDC officially recommends wearing face masks (even though not everyone complies). Meanwhile, governments in European countries like Spain, Ukraine, or certain regions in Italy require everyone, big or small, to wear masks all the time, when shopping, walking a dog, or plainly going outside. Breaking the requirements could result in a hefty fine. 11. Image Annotation Business Models [Reviewed] In the rise of robotics, computer vision and image processing cameras, image annotation comes as the first step to get the right AI training data for Deep Learning models. Whether you build an app to allow users to snap fashion items at the store as a new omni-channel sales or use machine vision installed at edge device at the industrial facility to monitor anomalies: it starts with training massive image data sets. 12. What's The Best Image Labeling Tool for Object Detection? An image labeling or annotation tool is used to label the images for bounding box object detection and segmentation. It is the process of highlighting the images by humans. They have to be readable for machines. With the help of the image labeling tools, the objects in the image could be labeled for a specific purpose. The process of object labeling makes it easy for people to understand what is in the image. The labeling tool helps the people to mark the items in an image. There are several image labeling tools for object detection, and some of them use varied techniques for detection of the object, like a semantic, bounding box, key-point, cuboid, semantic and many more. In this article, we will talk about image labeling and the best image labeling tools. 13. Hacking Your Mind with AI: The Case of Addictive TikTok Discover how the AI algorithms used in TikTok make you addicted to the app.  Find out how companies use them to improve their marketing efforts! 14. ICDAR 2021 Competition: Detecting Tables Using Image Recognition To participate go to: https://competitions.codalab.org/competitions/26979 15. Unwrapping Wine Labels - How We Trained A Neural Network To Do It In the previous article, it was described a six-point method to unwrap wine labels. Finding anchor points were performed with Hough transform. It gave fair results for good labels, but for many real cases it was quite unstable, and the efforts to tune it didn’t help much. It became clear at some point, Hough transform itself wasn’t capable of handling the variety of label forms, so the next step was training a neural network. 16. Experiencing Seattle's Amazon Go Grocery Capitol Hill, Seattle. That’s the location of the most futuristic supermarket on the planet, the Amazon Go Grocery store. Today it opened its doors to the general public for the first time, and this was my experience. 17. The “Magic of Tidying for Photos” App — Introduction This post shares about the idea of applying Kon Marie organising method to your digital photos, and I developed an IOS app  with AppleML and PhotoKits to solve the problem. I share my reflection on lean product development strategy and my journey of making the app end-to-end using a minimum, iterable tech stack and strategy. 18. Creating an Image Recognizer on Golang Telegram Bot In this article, we will go over a project on image recognition using Go. We will also create a Telegram bot, through which we can send images for recognition. 19. Major Image Recognition And Annotation Trends Image recognition and annotation technologies are evolving. New techniques that allow you to solve a wide variety of tasks quickly appear. We are happy to present five major trends in image recognition and annotation. 20. Facial Recognition On Drone A facial recognition demonstration using Keras, Tensorflow, python and the drone Tello, from DJI. 21. Are Your Photos Exposing You? You may be accidentally sharing personal information in your photos 22. Clearview AI Fined in UK for Illegally Storing Facial Images In this thread, our community discusses Clearview's endeavor and the backlash that followed. 23. Can I Avoid Facial Recognition At The Airport? Technically, yes, but it is not always easy 24. High-Resolution Photorealistic Image Translation in Real Time You can apply any design, lighting, or graphics style to your 4K image in real-time using this new machine learning-based approach 25. Image Analysis using AWS Rekognition via Lambda Function In this blog, I am going to show you how we can use rekognition for image analysis using lambda function.we will be going to perform label detection and object detection for an image so basically we are performing image analysis in this blog. 26. Using Sparse R-CNN As A Detection Model Today, we are going to discuss a method proposed by researchers from four institutions one of which is ByteDance AI Lab (known for their TikTok App). 27. How Artificial Intelligence Can Improve Modern Radiology Artificial Intelligence in Radiology - Advantages, Use Cases & Trends. Read about the challenges radiology AI is facing on the way to deployment 28. Use Cases and Roll-Out Tips for Image Recognition in Retail Heavily shattered by the pandemic, the retail sector is on the lookout for innovation. 29. Image Analysis using AWS Rekognition via AWS CLI In this blog, I am going to show you how we can use rekognition for image analysis using AWS CLI. So you can test the image analysis on your local machine. We will be going to perform label detection and object detection for an image so basically we are performing image analysis in this blog by using AWS CLI. 30. How To Identify Trees with Deep Learning Idea / inspiration 31. How Synthetic Data is Accelerating Computer Vision In the spring of 1993, a Harvard statistics professor named Donald Rubin sat down to write a paper. Rubin’s paper would go on to change the way that artificial intelligence is researched and practiced, but its stated goal was more modest: analyze data from the 1990 U.S. census, while preserving the anonymity of its respondents. 32. What is Human Pose Estimation? Part of the broader artificial intelligence and computer vision realms, human pose estimation (HPE) technology has been gradually making its presence seen in all kinds of software apps and hardware solutions. Still, human pose estimation seemed to be stuck at the edge, failing to cross into mainstream adoption. 33. Manipulate Images Using Text Commands via this AI Manipulate Real Images With Text - An AI For Creative Artists! StyleCLIP Explained Thank you for checking out the 33 most read stories about Image Recognition on HackerNoon. to find the most read stories about any technology. Visit the /Learn Repo

33 Stories To Learn About Image Recognition

278 Stories To Learn About Machine Learning

320 Stories To Learn About Deep Learning

306 Stories To Learn About Artificial Intelligence

Let's learn about via these 76 free stories. They are ordered by most time reading created on HackerNoon. Visit the to find the most read stories about any technology. Vc /Learn Repo 1. Valuing a Startup with Negative EBITDA Learn how to value a startup with negative EBITDA a startup valuation tool that stands for Earnings Before Interest, Taxes, Depreciation, and Amortization. 2. What Will the Next Era of Marketplaces Look Like? 3. Li Jin: My Path to Venture Capital 4. Community-as-a-Service Concept: How You Can Monetize Access 5. Airtable: 26 Firms Offering Non-Dilutive Funding Options to Founders Raise money without giving up a piece of your company 6. Advice for Founders During Uncertain Times Having been in the VC business for 22 years now, we have had the misfortune of living through at least two major downturns. While each has its own unique characteristics, there are a few lessons that may be relevant in responding to the latest crisis. 7. Consider a Loan Instead of VC Money “If you’re starting a business and you take out a loan, you’re a moron,” said firebrand entrepreneur Mark Cuban in an interview with Bloomberg. 8. Being A Creator: Definition, Economy, Tools and More 9. Crypto VC Behavior Reveals Market Trends for 2023 and Beyond Crypto VCs scaled down funding significantly in 2022 due to the macro uncertainty and systemic risks such as Luna's collpase and the recent FTX debacle. 10. 10 First Principles from Investing Legends for Better Decision Making Now We're living in crazy times. Coronavirus, lockdowns, recession and riots. It's enough to make even the most seasoned investor wonder what to do next. Not to mention what it all does to an average person. 11. Invest in People, Not Companies: Identity-Centric Investment Model According to Paul A. Gompers and Anna Kovner’s research, serial entrepreneurs are the key driving forces in the modern business world. Those who succeeded once have significantly higher chances of succeeding in their next venture by 30% when compared to the first-timers. That makes the past performance of an entrepreneur is a strong indicator of a future success.[1] 12. What Will VC Investment Look Like in the Year Ahead? The crises of 2022, which came in the form of geopolitical instability, economic recession, and inflation, have taken a toll on the venture capital investment. 13. What'll Happen to Crypto After Black Monday 2020? Bitcoin was explicitly called “digital gold” back in 2011, and a 2010 InfoWorld article says: 14. Aves Lair Announces Three Startups Comprising Inaugural Accelerator Cohort New York, NY, (December 14, 2020) – Aves Lair, a New York-based venture capital and frontier technology accelerator, announced today the debut of three innovative startups in its Winter 2020 cohort. 15. Pecan.ai Raises 11 Million to Bring Machine Learning to Business Analysts Pecan.ai has just come out of stealth, raising an $11M Series A, to enable business analysts to build machine learning models automatically. Dell Capital led the round, joined by S capital and bringing the total funding of the company to $15M. 16. Bootstrapping A Marketplace: "Fake It Till You Make It" Strategy 17. Passion Economy: The "Influencer Stack" and Chinese Influencer Incubators [Tweetstorm] 18. How 100 True Fans Can Change Everything More than a decade ago, Wired editor Kevin Kelly wrote an essay called “1,000 True Fans,” predicting that the internet would allow large swaths of people to make a living off their creations, whether an artist, musician, author, or entrepreneur. Rather than pursuing widespread celebrity, he argued, creators only needed to engage a modest base of “true fans”—those who will “buy anything you produce”—to the tune of $100 per fan, per year (for a total annual income of $100,000). By embracing online networks, he believed creators could bypass traditional gatekeepers and middlemen, get paid directly by a smaller base of fans, and live comfortably off the spoils. 19. What is the Biggest Risk For Passion Economy Startups? Creators want a direct relationship with their own users, which is against the platforms’ DNA. E.g. FB restricts ability for large group admins to message members. Agree that platforms adding in direct monetization
features is a big risk though to new startups in the Passion Economy.
Here’s some tactics for startups to hedge against incumbents: 20. Future of Technology in Modern Education: Hallway Conversation 21. Why Do Investors Say "No"? All startups hunt for capital, and in order to attract it from investors or funds they need, first of all, to contact them. 22. Passion Economy: Not All Communities Are Created Equal 23. Why Hardware Will Be Sexy to Mainstream VC (in 5 Slides) If you're a software engineer looking to become a founder, consider hardware. There's a shift coming. 24. How I Managed a Fund that Gave No Money, Took No Equity and Had an Exit I managed a fund that gave no money, took no equity, and had an exit. Here's that story. 25. Passion Economy: Future of Social Products [Tweetstorm] I like this article by Sam Lessin on the future of social, but think there’s a big omission. Namely, one of the major opportunities in social is to first
focus on helping creators monetize and own their audiences, then layering in social elements on top. 26. Private Equity in Your Pocket: How Digitization Changed VC Investment Market With further simplification of access to the VC market, people will start massively investing in new technologies and shape the image of tomorrow. 27. An insight into Asia's Tech Startup Scene With the boom in technology, tech startups are booming all over the world, especially in the Asian region. In recent years, it has been noticed that startups are growing the track record of success stories in the business world. 28. Investment Trend in Generative AI Shows No Signs of Slowing Down Generative AI refers to the use of machine learning to create new content. It has the potential to revolutionize industries 29. The European Tech Champions Initiative Aims to Fend Off Non-EU Investors The European Tech Champions Initiative (ETCI) aims to fend off the threat of non-European investors acquiring such businesses as they become more established. 30. Meet the Writer: HackerNoon's Contributor Eran Elhanani, Co-Founder of BullPerks and GamesPad My name is Eran Elhanani, and I’m the Co-Founder of BullPerks, a decentralized VC and multichain launchpad, currently ranking as the #2 IDO investment platform. 31. Pros and Cons of Running a Managed Marketplace 32. Naming Your Startup: 6 Considerations for First Time Founders Naming your startup can be a tricky and time-consuming process. This article breaks down some tips and stories  from VC Partners and CEOs to get you started. 33. What Are Crypto Venture Capitalists (VCs) Investing in This Bearish Market? The value of a cryptocurrency can be influenced by various factors, including news about the cryptocurrency and its underlying technology, regulations etc 34. 5 Things That Make Ecommerce Startups Work 35. Notion : A Product That Users Love, and VCs Can't Invest Into If you somehow navigate the mysterious path to reach the Notion HQ at 1:00 PM PST on a Friday, you would find an empty office with two golden poodles and a mutt running around. The entire team would not be far though, just shy of a mile away sitting at Barzotto and eating Extra-Long Noodles pasta over a glass of sparkling white wine. Add some soft serve gelato to that. This is just one of the many idiosyncrasies you would find in this 20 member start-up that has captured over a million users with their sleek product. 36. Data Leak of $1B “Tech Unicorn” Gorillas Shows How Far Venture Capital Has Fallen VC money is spilling out onto the streets, and despite millions of dollars of investment tech companies still neglect their security. 37. Raising Money? Prioritize Smart Money When raising capital, not all startups have options—they only receive one term sheet from one interested investor. And that should be celebrated. However, some startups are “hot” enough to attract multiple VCs. 38. For Founders: How the Venture Capital Investors Deal With Your Pitch Deck А useful guide for founders on how to prepare a pitch deck and how we look on it as a VC investor. 39. Over 90% of Startups Fail, So it is Difficult for Banks to Commit Millions BBVA launched BBVA Spark in July 2022, its comprehensive financial services proposal for innovative companies in their various stages of growth. 40. How Aggregators Are Replacing Brands [Tweetstorm] 41. Your Complete Guide To Venture Capital Funding For Startups Venture Capital is an institutional investment made into early-stage startup companies. If you want the investment, you need to work on a growth plan. 42. How Technology is Helping Venture Capitalists Make Investment Decisions From stone tools to the invention of the wheel to steam engines to the internet, technology has come a long way, it has overtaken nearly all major aspects of our lives. Just take this article for example, you wouldn’t be reading it if it wasn’t for technology. 43. Critical Necessities of the Venture-Backed Startup Once you’ve obtained the capital required to launch your startup, it’s time to look at fulfilling your technical needs. Do you have a secure web server? An IT security specialist? A reliable computer setup? These are all important, but there’s more. 44. 3 Best Tips To Attract Investors Being An Outsider Unshackled Ventures founding partner Nitin Pachisia explains three ways he built a strong network during his fundraising journey. 45. Passion Economy is a Consumer Enterprization Movement 46. List of Top VC Firms Investing Into Promising HR Startups [Part 4] Hi there! My name is Vitalii Berdichevskyi, I’m a Product Manager. In this essay of the series I will share with you List of top VC's invested in HR Tech. Previous parts of the series you can find below: 47. Startup Metrics: The Dynamics of Network Effects [This essay was co-authored by my colleague D’Arcy Coolican (@dcoolican) and me. We also compiled a list of metrics to measure your network effects.] 48. [Announcement] Wise Raises $12M To Accelerate Embedded Banking Partnerships Wise, the leading embedded banking FinTech, that offers business banking via partnerships, announced today that it has closed a $12M Series A. 49. How Crypto Fund Iconium is Leading The Charge for Web3 In this post we’ll be looking at how crypto investment fund Iconium is leading the charge for development of Web3 infrastructure. 50. How To Raise Capital for Blockchain Projects (and other startups) Venture capitalists get pitched hundreds of times a year. And, even though countless articles have been written on the topic of how to present a startup to professional investors, entrepreneurs consistently fail to address check points needed by the investor to come to a decision. 51. The Hedge Fundization of VC The Hedge Fundization of VC 52. Rock, Paper, Blockchain Most data from government and financial institutions is still siloed. Public blockchains may be used by anyone and are permissionless. But how far off are we? 53. 263 Meetings, 12 Months and $128,700 in Startup Funding Later Is traction enough to raise funding for a Startup or do you need more than that? Here's how we raised $128,700 at Alpe. 54. Passion Economy: How to Organically Grow While Maintaining Paid Groups 55. Walnut Raises $2.5M Seed To Make Product Demos Easier For the Post-Pandemic World Seed round is being led by NFX with additional participation from other top SF Angels 56. Growing Your Influencer Brand Into a Sustainable Business 57. Networking at VC Events: How to Gain More Value? The pandemic and lockdowns have proven to the world the simple truth that we are social beings. 58. What Rowing For The U.S. National Team And Building A Startup Have In Common Saying “it’s not fair” won’t get you there. To be the best, you have to beat everyone. 59. Passion Economy Shall Be The Disruptor of 2021 and Beyond Clayton Christensen's disruption theory and what it means for the Passion Economy 60. The Secret Behind Successful Marketplaces [Tweetstorm] Thread: The best consumer marketplaces end up
supply-constrained, because they tap into an incredible amount of demand. For instance, Airbnb, Lyft, and Uber have seemingly infinite demand since the product/market fit is so strong–and this demand puts pressure on supply. 61. 7 Ways to Stand Out at a Tech Conference After more than 60 meetings at a major tech conference in Dubai Gitex Global 2022, I figured out what helps projects stand out and get stuck in the memory. 62. The State of Startups in the Southeast Fourth annual report reveals continued market maturation; $34.5 billion invested in southeastern startups since 2015 63. How to Choose the Right Partners in a Venture Capital Fund When you are searching for the right venture capital fund, there are a lot of factors that can go into deciding which fund to choose. 64. How Hacker Noon Leans on Our Community -  a #CUBoulder Talk by Linh Dao Smooke Transcript of COO Linh Dao Smooke's talk on how Hacker Noon leans on their community as a fundraising mechanism as opposed to traditional VCs. 65. Takeaways from Sequoia Capital's Black Swan Memo Soothsayer Sequoia Capital warned its portfolio companies in a memo—how can startups stay healthy? 66. Disruption Theory for the Passion Economy: 4 Major Implication (Part 2) Passion Economy platforms are rooted in a value proposition that deeply appeals to workers—they lower the barriers to earning income 67. The Passion Economy’s Worklife Balance The top-earning writer on the paid newsletter platform Substack earns more than $500,000 a year from reader subscriptions. The top content
creator on Podia, a platform for video courses and digital memberships,
makes more than $100,000 a month. And teachers across the US are
bringing in thousands of dollars a month teaching live, virtual classes
on Outschool and Juni Learning. 68. The Future of Marketplaces: Reinventing the $10 Trillion Service Economy [I co-wrote this essay with my a16z partner Andrew Chen. We’re excited about the future of marketplaces and the service economy. Hope you enjoy this glimpse into our thinking, and let us know if you have any feedback!] 69. Live Shopping Features On Social Media And Interview With Chinese Social Commerce Tycoon How new models of e-commerce can enable the Passion Economy 70. How We Closed Our First Pre-Seed Round in 2 Weeks We got accepted into YC on the 9th of December 2019. Originally, we wanted to raise a pre-seed round before starting YC on the 6th of January. But at the same time, we didn’t want to incorporate the company in 2019 so we wouldn’t have to deal with the taxes for 2019. Call us lazy! One would say lazy is not bad, as long as it’s smart lazy! 71. How to Fix Entrepreneurship in Europe I Studied Entrepreneurship in 4 Countries. Here's How To Fix it in Europe. 72. What to Learn from the Power User Curve The importance of power users Power users drive some of the most successful companies — people who
love their product, are highly engaged, and contribute a ton of value to the network. In ecommerce marketplaces it’s power sellers, in ridesharing platforms it’s power riders, and in social networks it’s influencers. 73. Startup Metrics: 16 Ways to Measure Network Effects [This post was co-authored by me and my colleague at a16z, D’Arcy Coolican (@dcoolican). This is a companion to our other essay exploring the dynamics of network effects. If you’re an operator, we’d love to hear your thoughts on how you measure your own product’s network effects!] 74. How to Build a Company in Highly Regulated Marketplaces: 8 Things to Consider What can founders / operators do when it comes to navigating the complexities that come with the territory in regulated industries? Here are some opportunities. 75. Bigger picture with the Decentralized identifiers This article summarizes DIDs and VCs' fundamental principles, describes the use in the real world, and provides bonus materials for tech-savvy readers. 76. How to Invest Intelligently in the Podcast Ecosystem In the world of podcasting, the flywheel is spinning: new technologies including AirPods, connected cars, and smart speakers have made it much easier for consumers to listen to audio content, which in turn creates more revenue and financial opportunity for creators, which further encourages high-quality audio content to flow into the space. There are now over 700K free podcasts available and thousands more launching each week. Thank you for checking out the 76 most read stories about Vc on HackerNoon. to find the most read stories about any technology. Visit the /Learn Repo

76 Stories To Learn About Vc

There are various methods of visualizing three-dimensional objects in two-dimensional space. For example, most 3D graphics engines use perspective projection as the main form of projection. This is because perspective projection is an excellent representation of the real world, in which objects become smaller with increasing distance. But when the relative position of objects is not important, you can use parallel projections for a better understanding of the size of objects. They are more common in engineering and architecture, where it is important to maintain parallel lines. Since the birth of computer graphics, these projections have been used to render 3D scenes when 3D rendering hardware acceleration was not possible. Recently, various forms of parallel projections have become a style choice for digital artists, and they are used to display objects in infographics and in digital art in general. The purpose of this article is to show how to create and manipulate isometric views in SVG, and how to define these objects using the library. To illustrate SVG’s capabilities in creating parallel projections, we will use isometric projection as an example. This projection is one of the dominant projection types because it allows you to maintain the relative scale of objects along all axes. JointJS Content Overview Isometric projection SVG isometric view transformations SVG isometric view transformations Isometric grid rendering Creating a custom isometric SVG element Z-index calculation in isometric diagrams The JointJS demo Isometric projection Let’s define what isometric projection is. First of all, it is a parallel type of projection in which all lines from a “camera” are parallel. It means that the scale of an object does not depend on the distance between the “camera” and the object. And specifically, in isometric (means “equal measure” in greek) projection, scaling along each axis is the same. This is achieved by defining equal angles between all axes. In the following image, you can see how axes are positioned in isometric projection. Keep in mind that in this article we will be using a left-handed coordinate system. One of the features of the isometric projection is that it can be deconstructed in three different 2D projections: top, side, and front projections. For example, cuboid can be represented by three rectangles on each 2D projection and then combined into one isometric view. The next image represents separate projections of an object using the left-handed coordinate system. Then we can combine them into one isometric view: The challenge with is that it contains 2D objects which are located on one XY-plane. But we can overcome this by combining all projections in one plane, and then separately applying transformation to every object. SVG SVG isometric view transformations In 3D, to create an isometric view, we can move the camera to a certain position, but SVG is purely a 2D format, so we have to create a workaround to build such a view. We recommend reading that presents a method for creating isometric representations from 2D object views — top, side, and front projections. Based on the article, we need to create transformations for each 2D projection of the object separately. Cody Walker’s article First, we need to rotate our plane by 30 degrees. And then we will skew our 2D image by -30 degrees. This transformation will align our axes with the axes of the isometric projection. Then we need to use a scale operator to scale our 2D projection down vertically by 0.8602. We need to do it due to the fact of isometric projection distortion. Let’s introduce some SVG features that will help us implement isometric projection. The SVG specification allows users to specify a particular transformation in the of an SVG element. This attribute helps us apply a linear transformation to the SVG element. To transform 2D projection into an isometric view, we need to apply scale, rotate and skew operators. transform attribute To represent the transformation in code, we can use the object, which is a browser API to represent the transformation matrix. Using this interface, we can create a matrix as follows: DOMMatrixReadOnly const isoMatrix = new DOMMatrixReadOnly()
    .rotate(30)
    .skewX(-30)
    .scale(1, 0.8602); This interface allows building a transformation matrix using our values, and then we can apply the resulting value to the attribute using the function. transform matrix In SVG, we can present only one 2D space at a time, so for our conversion, we will be using top projection as a base projection. This is mostly because axes in this projection correspond with axes in a normal SVG viewport. To demonstrate SVG possibilities, we will be using the JointJS library. We defined a rectangular grid in the XY-plane with a cell width of 20. Let’s define SVG for the elements on the top projection from the example. To properly render this object, we need to specify two polygons for two levels of our object. Also, we can apply a translate transformation for our element in 2D space using DOMMatrix: // Translate transformation for Top1 Element
const matrix2D = new DOMMatrixReadOnly()
    .translate(200, 200); 
<polygon joint-selector="body" id="v-4" 
  stroke-width="2" stroke="#333333" fill="#ff0000" 
  fill-opacity="0.7" points="0,0 60,0 60,20 40,20 40,60 0,60" 
  transform="matrix(1,0,0,1,200,200)">
</polygon>


<polygon joint-selector="body" id="v-6" 
  stroke-width="2" stroke="#333333" fill="#ff0000" 
  fill-opacity="0.7" points="0,0 20,0 20,40 0,40" 
  transform="matrix(1,0,0,1,240,220)">
</polygon> Then we can apply our isometric matrix to our elements. Also, we will add a translate transformation to position elements in the right place: const isoMatrix = new DOMMatrixReadOnly()
    .rotate(30)
    .skewX(-30)
    .scale(1, 0.8602);

const top1Matrix = isoMatrix.translate(200, 200);
const top2Matrix = isoMatrix.translate(240, 220); ‍ For simplicity, let’s assume that our element’s base plane is located on the XY plane. Therefore, we need to translate the top view, so it will be viewed as it is located on the top of the object. To do it, we can just translate the projection by its Z coordinate on the scaled SVG space as follows. Top1 element has elevation 80, so we should translate it by (-80, -80). Similarly, Top2 element has elevation 40. We can just apply these translations to our existing matrix: const top1MatrixWithHeight = top1Matrix.translate(-80, -80);
const top2MatrixWithHeight = top1Matrix.translate(-40, -40); ‍ In the end, we will have the following attributes for and elements. Note that they differ only in the two last values, which represent the translate transformation: transform Top1 Top2 // Top1 element
transform="matrix(0.8660254037844387,0.49999999999999994,-0.8165000081062317,0.47140649947346464,5.9,116.6)"

// Top2 element
transform="matrix(0.8660254037844387,0.49999999999999994,-0.8165000081062317,0.47140649947346464,26.2,184.9)" To create an isometric view of side and front projections, we need to make a so we can place all projections on 2D SVG space. Let’s create a net by attaching side and front views similar to the classic cube net: net Then we need to side and front projections by 45 degrees. It will allow us to align the Z axis for all projections. skewX After this transformation, we will get the following image: Then we can apply our isoMatrix to this object: In every projection, there are parts that have a different 3rd coordinate value. Therefore, we need to adjust this depth coordinate for every projection as we did with the top projection and its Z coordinate. In the end, we will get the following isometric view: SVG isometric view transformations allows us to create and manipulate such objects with ease due to its elements framework and wide set of tools. Using JointJS, we can define and control isometric objects to build powerful isometric diagrams. JointJS Remember the basic isometric transformation from the beginning of the article? const isoMatrix = new DOMMatrixReadOnly()
    .rotate(30)
    .skewX(-30)
    .scale(1, 0.8602); In the JointJS library, we can apply this transformation to the whole object which stores all SVG elements, and then simply apply the object-specific transformations on top of this. Isometric grid rendering JointJS has great capabilities in the rendering of custom SVG markup. Utilizing JointJS, we can generate a path which is aligned to an untransformed grid, and have it transformed automatically with the grid thanks to the global paper transformation that we mentioned previously. You can see the grid, and how we interpret the coordinate system in the demo below. Note that we can dynamically change the paper transformation which allows us to change the view on the fly: https://codepen.io/jointjs/pen/poQzZJZ/f8ea3c83fc17b9f438bb1e95d91d11a9?embedable=true Creating a custom isometric SVG element Here we show a custom SVG isometric shape in JointJS. In our example, we use the property to store information about a third dimension, and then use it to render our isometric object. The following snippet shows how you can call the custom function to alter object properties: isometricHeight createIsometricElement const element = createIsometricElement({
    isometricHeight: GRID_SIZE * 3,
    size: { width: GRID_SIZE * 3, height: GRID_SIZE * 6 },
    position: { x: GRID_SIZE * 6, y: GRID_SIZE * 6 }
}); In the following demo, you can see that our custom isometric element can be moved like an ordinary element on the isometric grid. You can change dimensions by altering parameters of the function in the source code (when you click “Edit on CodePen”): createIsometricElement https://codepen.io/jointjs/pen/eYPxLgo/72e6d9213d8ebab5f16fcfe34c6a99e9?embedable=true Z-index calculation in isometric diagrams One of the problems with an isometric view is placing elements respective to their relative position. Unlike in a 2D plane, in an isometric view objects have perceived height and can be placed one behind the other. We can achieve this behavior in SVG by placing them into the DOM in the right order. To define the order in our case, we can use the JointJS attribute which allows sending the correct element to the background, so that it can be overlapped/hidden by the other element as expected. z We decided to sort the elements using the . The algorithm consists of two steps. First, we need to create a special graph, and then we need to use a depth-first search for that graph to find the correct order of elements. topological sorting algorithm As the first step, we need to populate the initial graph — for each object we need to find all objects behind it. We can do that by comparing the positions of their bottom sides. Let’s illustrate this step with images — let’s, for example, take three elements which are positioned like this: We have marked the bottom side of each object in the second image. Using this data, we will create a graph structure that will model topological relations between elements. In the image, you can see how we define the points on the bottom side — we can find the relative position of all elements by comparing and points. We define that if the and coordinates of point are less than the coordinates of point , then object is located behind object . aMax bMin x y bMin aMax b a Comparing the three elements from our previous example, we can produce the following graph: After that, we need to use a to find the correct rendering order. A depth-first search allows us to visit graph nodes according to the visibility order, starting from the most distant one. variation of the depth-first search algorithm Here is a library-agnostic example of the algorithm: const sortElements = (elements: Rect[]) => {
    const nodes = elements.map((el) => {
        return {
            el: el,
            behind: [],
            visited: false,
            depth: null,
        };
    });

    for (let i = 0; i < nodes.length; ++i) {
        const a = nodes[i].el;
        const aMax = aBBox.bottomRight();

        for (let j = 0; j < nodes.length; ++j) {
            if (i != j) {
                const b = nodes[j].el;
                const bMin = bBBox.topLeft();
                if (bMin.x < aMax.x && bMin.y < aMax.y) {
                    nodes[i].behind.push(nodes[j]);
                }
            }
        }
    }

    const sortedElements = depthFirstSearch(nodes);
    return sortedElements;
};

const depthFirstSearch = (nodes) => {
    let depth = 0;
    let sortedElements = [];

    const visitNode = (node) => {
        if (!node.visited) {
            node.visited = true;

            for (let i = 0; i < node.behind.length; ++i) {
                if (node.behind[i] == null) {
                    break;
                } else {
                    visitNode(node.behind[i]);
                    delete node.behind[i];
                }
            }

            node.depth = depth++;
            sortedElements.push(node.el);
        }
    };

    for (let i = 0; i < nodes.length; ++i) {
        visitNode(nodes[i]);
    }

    return sortedElements;
}; This method can be implemented easily using the JointJS library — in the following CodePen, we use a special JointJS event to recalculate z-indexes of our elements whenever the position of an element is changed. As outlined above, we use a special property of the element model to specify rendering order and assign it during the depth-first traversal. (Note that the algorithm’s behavior is undefined in the case of intersecting elements, due to the nature of the implementation of isometric objects.) z https://codepen.io/jointjs/pen/ZEqPyyX/e06f0a38d57651c6eb3502a6ce6dd1e2?embedable=true The JointJS demo We have created a JointJS demo that combines all of these methods and techniques and also allows you to easily switch between 2D and isometric SVG markup. Crucially, as you can see, the powerful features of JointJS (which allow us to move elements, to connect them with links, and to create tools to edit them, among others) work just as well in the isometric view as they do in 2D. You can see the . JointJS isometric diagram demo here Throughout this article, we used our open-source JointJS library for illustration. Also published here.

Creating Isometric Diagrams with SVG

Computer vision is a rapidly developing field of artificial intelligence, particularly in the area of 3D. This overview will consider an applied task: transitioning between 2D and 3D environments. DIRECT TASK To begin with, we will analyze how to solve a direct problem of computer graphics, namely creating a 2D image using a 3D model, and get acquainted with the basic concepts. is the process of moving from a 3D model to its 2D projection. You’ve probably heard of some of them: Rendering is one of the earliest and fastest rendering methods. Rasterization treats the model as a grid of polygons. These polygons have vertexes embedded with information such as position, texture, and color. These vertices are then projected onto a plane perpendicular to the perspective. Rasterization has problems with overlapping objects: if the surfaces overlap, the last part drawn will be reflected during rendering, which will cause the wrong object to be displayed. This problem was solved using z-buffering (in fact, the z-buffer is a depth map). Rasterization Unlike rasterization, the potential problem of overlapping surfaces does not occur during raycasting. Ray casting, as the name suggests, directs rays at the model from the camera’s point of view. Rays are output to each pixel on the image plane. The surface that it hits first will be shown during rendering, and any other intersection after the first surface will not be drawn. Ray casting. . Despite the advantages of ray casting, the technique still lacks the ability to correctly model shadows, reflections, and refractions. The ray tracing method was developed to help resolve these issues. Ray tracing works in a similar way to ray casting, except that it displays light better. Basically, the primary rays from the camera’s point of view are directed at the models to produce secondary rays. After hitting the model, shadow rays, reflected rays, or refractive rays will be emitted, depending on the surface properties. Ray tracing Now that we’ve considered the direct problem of building a 2D image from a 3D model, let’s look at ways to solve the inverse problem: building a 3D model from a 2D image. INVERSE PROBLEM A two-dimensional photograph is a projection of a three-dimensional scene. A 3D scene is a collection of 3D meshes, vertices, faces, texture maps, and a light source viewed from a camera or viewpoint. For simplicity, let’s limit the scene to a single 3D object. If we were able to restore the original 3D scene from which the 2D photo was created, we should be able to verify this by projecting the 3D object onto 2D using the same point of view that was used to create the input 2D photo. To reconstruct an object, you need to calculate all possible combinations of vertices, faces, light sources, and textures, which, when projected in 2D, should give an equivalent image in 2D, given the input image, provided that the camera position is the same. This is essentially a search problem. But the problem with brute forcing is that there are a huge number of combinations of vertices, faces, texture maps, and lighting that can be created, so we can’t solve this problem by brute force. Let’s look at the existing approaches to solving this problem! DIB-R DIB-R is a differential renderer that models pixel values using the differentiable rasterization algorithm. It has two methods for assigning pixel values. One for foreground pixels, and the other for background pixels. Here, in contrast to standard rendering, where the pixel value is assigned to the nearest face covering the pixel, foreground rasterization is considered as an interpolation of vertex attributes. On each foreground pixel, we perform a z-buffering test and assign it to the nearest covering face. Each pixel is affected exclusively by this face. So foreground pixels are calculated as an interpolation of the nearest three neighboring vertices using a weight for each vertex. For background pixels, i.e. pixels that are not covered by any face of the 3D object, the value is calculated based on the distance from the pixel to the nearest face. DIB-R can generate images with realistic lighting and shading effects that are difficult to achieve with traditional rendering. Official paper Im2Struct: SMN+SRN A structural masking network (SMN) creates an object mask based on the input 2D image at different scales. This is a multi-layer convolutional neural network (CNN). Its task is to save information about the form while viewing irrelevant information: background and textures. A structure-restoring network (SRN) recursively reconstructs the hierarchy of object details in the form of a cuboid structure. The SRN receives input data from the SMN, adds CNN characteristics of the 2D image, and then passes these functions to the recursive neural network (RvNN) for decoding into a 3D structure. At the output, we get an image in the form of three-dimensional cuboids with a plausible spatial configuration. Im2Struct has several advantages over traditional 3D scanning methods, as it can recover the 3D structure of an object from a single 2D image, which is often faster and less expensive than scanning an object from multiple viewpoints. Official paper Link to the solution ATLAS The method takes as input a sequence of RGB images of arbitrary length. The internal characteristics and pose are known for each image. These images are passed through a 2D CNN backbone for feature extraction. The objects are then projected back into the 3D voxel volume and accumulated using the current average value. Once the image elements are combined in 3D, we regress the TSDF directly using the 3D CNN. ATLAS is useful in a variety of industries, including manufacturing, engineering, and archaeology. One limitation of ATLAS 3D is that it requires the object being scanned to be stationary, which may not always be feasible in certain applications. Additionally, the system may struggle to capture fine details and textures on objects with highly reflective or transparent surfaces. Official paper Link to the solution Mesh R-CNN The framework uses a two-stage approach: in the first stage, it detects and segments the object in the image using a convolutional neural network (CNN), similar to the popular Mask R-CNN framework. In the second stage, it regresses a set of 3D vertices for each object instance using a mesh prediction network. One of the main advantages of Mesh R-CNN is its ability to reconstruct detailed 3D meshes of objects, including their fine-grained geometry and texture. This makes it useful for applications such as virtual reality, augmented reality, and 3D printing. Official paper Link to the solution We’ve covered several state-of-the-art solutions for solving the inverse graphics problem. All of these solutions can help you solve a wide variety of tasks, like reconstructing a room, creating a 3D local map, reconstructing a 3D scene from a single image, or even estimating the height and depth of crops or terrain to guide planting, harvesting, and irrigation decisions. Keep in mind that all these solutions are based on different approaches to rendering, voxel prediction, mesh prediction, and so on. But they all have a common need to build or predict a depth map in one form or another. That’s why I also propose to separately consider the problem of constructing a depth map. DEPTH ESTIMATION There are several ways to get a depth map: Use a stereo pair of RGB images. Use an RGB-D camera. Teach a model to predict the depth map based on one RGB image Let’s look at several state-of-the-art solutions for predicting depth maps. Monocular depth estimation — GLPN A new architecture with global and local feature paths through the entire network. The overall structure of the framework is next: the transformer encoder enables the model to learn global dependencies, and the proposed decoder successfully recovers the extracted feature into the target depth map by constructing the local path through a skip connection and the feature fusion module. Official paper Link to the solution Dense depth model For the encoder, the RGB input image is encoded into a vector of objects using a DenseNet-169 network pre-trained in ImageNet. This vector is then fed into a sequential series of layers with increased sampling to build a final depth map with a resolution equal to half of the input. These upsampling layers and their associated bandwidth connections form the decoder. Official paper Link to the solution Midas The architecture is represented by a visual multi-connected transformer as the basis. The overall encoder-decoder structure that has been successful for prediction in the past is preserved. The input image is converted to tokens either by extracting non-overlapping sections and then the linear projection of their smoothed representation (DPT-Base and DPT-Large) or by applying ResNet-50 (DPT-Hybrid). Image embedding is supplemented with positional embedding and a patch-independent token. Tokens go through several stages of conversion. Tokens are collected from different stages as a multi-resolution image (Reassemble). The Fusion modules gradually merge and upsample views to produce a detailed forecast. Link to the solution All of these solutions can help you get the depth map and use it in your own way. For instance, you might want to build 3D scenes with PyTorch-3d. Official paper THE HAND OF CROWDSOURCING In the latter case, MIDAS was able to achieve its result by linking new data sources, which no one had implemented before. There is a difficulty in collecting diverse depth datasets at scale, so a tool has been introduced to combine complementary data sources. In addition, a new dataset based on 3D movies provides reliable information about various dynamic scenes. Thus, I wanted to focus on the problem of data for the 3D direction. Every developer faces this problem and has to somehow dodge it, including architecturally. All these solutions I described were using almost the same scroll of the open datasets. It was not enough because it’s not that simple to collect such complex and high-quality data due to various reasons, such as occlusions, poor lighting conditions, and limited viewpoints. When there is not enough data available, it becomes difficult to accurately estimate the depth and structure of the scene, leading to inaccurate or incomplete 3D reconstructions. Crowdsourcing can be used as a potential solution to address the problem of needing more data for 3D reconstruction. By leveraging the collective effort of a large number of individuals, crowdsourcing can provide additional data and perspectives on a scene, which can improve the accuracy and completeness of the 3D reconstruction. For example, a crowdsourcing platform could be used to collect multiple images of a scene taken from different viewpoints by a large number of contributors. These images could then be processed using multi-view stereo or structure-from-motion techniques to create a more accurate 3D reconstruction of the scene. This is exactly what was implemented in the Neatsy project to partially compensate for a lack of 3-D data. Neatsy develops AI software for virtually sizing shoes. They used the Toloka crowdsourcing platform for additional data collection (more than 50 thousand new photos) and made improvements to the model’s metrics. Their software creates a 3D model of your feet using around 50 different measurements and helps you find the perfect pair of sneakers. The project has moved on and now they can also diagnose health problems in feet, all thanks to data from people in the crowd. This is just one example of the strong potential of developing 3D technology. SUMMARY There are many state-of-the-art solutions available for forward and inverse graphics, as well as for predicting the depth map. We looked at the approaches of each in practical applications and also noted the limitations caused by the lack of data for 3D. Crowdsourcing platforms have the potential to solve the data collection problem and support the development of 3D technologies for real-life computer vision applications.

Single Image 3D Scene Reconstruction: A Review of Recent Advances

THE GIANT CHILDREN

Hey there👋, Nowadays the browsers are becoming more powerful then in the past and now they can render different complex animations with ease. You might have seen lot of websites rendering 3D model without any performance issues. In this article we are going to use to render an iPhone model in a React application. Three.js Before we start this tutorial let me show you a demo website which renders this model. Here is the demo link: https://apple-iphone14.netlify.app/ For this website I have used Three.JS along with the GSAP to add smooth scrolling animations. If you want to learn to make this responsive 3D landing page you can watch the following tutorial👇. https://youtu.be/cT160dOzpGY Let's render an awesome 3D model in our React application. Setup and Installation For this tutorial we will be using (create-react-app) template. Open your project folder and use the following command to install CRA template. CRA npx create-react-app 3d-model-in-reactjs You can name your application as you wish for this tutorial I have kept it as "3d-mode-in-reactjs". Here the npx is a utility that comes when you install the npm, It helps us to run any npm package available on npm registry without installing that package. Let's install required libraries to render 3D model.
Change the directory to "3d-mode-in-reactjs" using the following command. cd 3d-mode-in-reactjs Use the following command to install two libraries the and . @react-three/fiber @react-three/drei npm install @react-three/fiber @react-three/drei : It is a React renderer for threejs. @react-three/fiber : It is a collection of useful helpers and fully functional, ready-made abstractions for @react-three/fiber. @react-three/drei For this tutorial we are going to use a 3D model of Apple iPhone 13 Pro Max. You can get this model from the following link👇. by DatSketch is licensed under . Before you use this model in any project make sure to check it's license. "Apple iPhone 13 Pro Max" Creative Commons Attribution Open the given link and download the model in format.
Why GLTF? As per the threejs gltf documentation glTF is focused on runtime asset delivery, it is compact to transmit and fast to load. Once you download the file, create one folder called inside the folder and extract all the model file under folder called . assets src 3D-Model Open the file and replace all code with the following css. App.css .App {
  width: 100vw;
  height: 100vh;
  display: flex;
  justify-content: center;
  align-items: center;
} Now we are ready to render the model. Let's understand the basics of rendering a 3D model Now there are 3 things that you have to keep in mind while rendering a model. scene: Here you can setup all objects, lights and cameras. camera: It is used to view the 3D model through different angels. renderer: It is used to display/render a scene on an HTML element called . canvas If you are directly using threejs in a plain JavaScript project then you might have to setup all these things from the basics, but in this tutorial we are using and so we don't have to do all the things from scratch. @react-three/fiber @react-three/drei Open the file, clean up the default code and let's import the element from . Checkout the following code snippet. App.js canvas @react-three/fiber import "./App.css";
import { Canvas } from "@react-three/fiber";

function App() {
  return (
    <div className="App">
      <Canvas>{/* Here we will render out model */}</Canvas>
    </div>
  );
}

export default App; Here the component sets up the and the and renders every frame. Canvas scene camera scene Now copy the following code inside the component. <Canvas /> <Canvas>
  <mesh>
    <boxGeometry />
    <meshStandardMaterial />
  </mesh>
</Canvas> Let's understand each of these elements. : A is a class that inherits from Object3d and it represents polygon objects. In the mesh you can instantiate the polygon object by using and . <mesh> mesh Geometry Material : It is a geometry class for a rectangular cuboid with a given 'width', 'height', and 'depth' <boxGeometry /> : It is a standard physically based material. <meshStandardMaterial /> For now you might see one black square in the middle that does not look like a 3D object. Let's add controls so that we can move the camera angle around this 3D object. Import the from the by using the following line. <OrbitControls /> @react-three/drei import { OrbitControls } from "@react-three/drei"; Now render this inside the components and after the . Refresh the development server page and try to grab the object. Now you might be able to see this square as an 3D object. This allow the camera to orbit around a target. <OrbitControls  /> <Canvas> <mesh> <OrbitControls  /> Now the object is total black because there is no source of light. So let's add lights in our scene. Inside the component add the following line before the . <Canvas /> <mesh> <ambientLight /> Here this globally illuminates all objects in the scene equally. Now you are able see the off-white color of the mesh material. We can also change the intensity of ambient light through props as shown in the following code. ambientLight <ambientLight intensity={1.25} /> You can also change the color of 3D object by using prop in the . checkout the following line. color <meshStandardMaterial /> <meshStandardMaterial color="green"/> Now the color of the 3D object is green. There are lot of different lights and elements are available which you can use to create whole model. For that you can refer the documentation of . That's it for the basics not let's render our iPhone. threejs Converting model to JSX component Open the model files and you will see different files like , , etc. We will convert into a JSX component so that we can easily render it in our React application. textures scene.gltf scene.bin scene.gltf Now to convert any GLTF files to JSX component we will use one simple command line tool called as . As per the documetation the is a small command-line tool that turns GLTF assets into declarative and re-usable react-three-fiber JSX components. gltfjsx gltfjsx Now open the terminal/cmd and go to the directory where you have stored all the model files, in my case it is inside the folder which is in the folder. So I will use the following command to change the directory. 3D-Model assets cd src/assets/3D-Model Now use the following command. npx gltfjsx scene.gltf This command will take few seconds and returns the file which is the JSX component. Scene.js NOTE: This gltfjsx command might give you an error if you are using latest NodeJS version. I'm currently using 17.2.0 and this command is working fine. If you face any error please downgrade your NodeJs version and try this code again. You can use tools like to use different versions of the NodeJs. nvm Important! Before we render the model from file we have to copy the model files into directory since if you look in the file it uses hook which loads nodes and materials from file. So copy , and into folder of your project. Scene.js public Scene.js useGLTF /scene.gltf textures scene.bin scene.gltf public Rendering the 3D model Open the file and remove the . Let's import the from file. App.js <mesh> Model Scene.js import Model from "./assets/3D-Model/Scene"; Now render the inside the component, before the . <Model /> <Canvas /> <OrbitControls /> <Canvas>
  <ambientLight intensity={1.25} />
     <Model />
  <OrbitControls />
</Canvas> Refresh the output page. Now you should be able to see the iphone rendering on the screen, you can zoom in and zoom out and move the model. Let's increase the view of the model. You can directly access the camera from the canvas element. Add the camera inside canvas as the following code snippet. <Canvas camera={{fov: 18}}> Here the stands for the "field of view". After setting the refresh the page to see the output. For now we have set the to 18. You can try different numbers and see the output. fov fov fov For now there is no reflection on the model. To add reflection on the model you have to set images around this model so that these images can reflect on the surface of this model. Instead of adding these images let's use from the library. <Environment /> component @react-three/drei Let's import the , <Environment /> import { Environment, OrbitControls } from "@react-three/drei"; Now after the model, let's add the Environment. <Environment preset="sunset" /> Here the preset sets different types of environment surrounding the model. For now we have set it to "sunset". You can check the different presets available from . here After adding this environment you should see the beautiful reflection on our iPhone. Let's render out model inside the component of the React so that it can load asynchronously. Suspense <Suspense fallback={null}>
   <Model />
</Suspense> Here is the final code. https://gist.github.com/codebucks27/6defde4db7e290d95d7ff1ae39078a06 Final Thoughts That's the end of this tutorial. You can use to read the documentation of the and and try different elements. If you want to create full fledge website using the 3D model and animate it's colors then you can follow the video tutorial given at the start of this lesson. https://docs.pmnd.rs/ @react-three/fiber @react-three/drei If you like this tutorial then you should checkout awesome tutorials I have on my YouTube channel called . You can check it from . CodeBucks here You might also like these website templates: A beautiful portfolio template in ReactJS => here NFT collection landing page in ReactJS => [here](- A beautiful portfolio template in ReactJS => ) here Thank you for reading this tutorial. Have a great day! Also published . here

How to Render 3D Model in a React JS Application

Cómo renderizar un modelo 3D en una aplicación React JS

Comment rendre un modèle 3D dans une application React JS

रिएक्ट JS एप्लिकेशन में 3D मॉडल कैसे प्रस्तुत करें

React JS アプリケーションで 3D モデルをレンダリングする方法

Como renderizar um modelo 3D em um aplicativo React JS

Cách hiển thị mô hình 3D trong ứng dụng React JS

如何在 React JS 应用程序中渲染 3D 模型

In this article, we will be talking about the practical use cases for ENS domains and how they will simplify Ethereum and blockchain technologies for the average Joe. ENS domains are the readable domain names for Ethereum wallet addresses. So the long address 0x6fg5y555cns gets translated into 123.eth. The main topics that will be covered are risk management on NFT mints, identity management for simplified logins across multiple platforms, IP and MAC address replacements via ENS, stabilizing gas fees during large events, verifying documents, and protecting owners of sellable assets. ENS will simplify how we navigate web 3. The exact same way readable IP addresses via DNS gave us the internet we know and love will be the same way readable ETH addresses via ENS will give us the web 3 we will end up easily integrating into. How will ENS help the regular person who is new to web 3? How will ENS help improve safety going forward? How will ENS help validate ownership? Minting Protection. Let me ask you a question. Were you ever looking at getting into a minting process for an NFT, but did not want to connect your wallet to the minting website because it might be a scam website looking at stealing your assets? If the answer is yes there is a handy tool to mitigate the risk of getting your assets wiped. This tool is called the Ethereum Name Service. Hypothetically you wanted to mint a Bored Ape before it blew up. Now instead of connecting to Yuga Labs website what hypothetically you could do is send the wETH (wrapped ether) to “mint.bayc.eth.” Sending tokens to an ENS website mitigates the risk of not actually connecting your wallet. Because connecting your wallet opens the risk of you signing a smart contract that you actually did not read through. Also, it protects you from fraud by knowing the only valid mint address is “mint.bayc.eth.” Identity Management. ENS domains offer risk management and also identity management. With web 2 some of the main nuisances are the fact that you have to remember so many passwords for so many different websites. Signing in with Google is somewhat helpful, but Google is centralized and can delete you from the internet at any time thus wiping all of the accounts you use to sign in with it. With Ethereum, it is on a blockchain and the only way you get wiped out is if you consent to it. With an ENS domain, you can add all of your web 2 accounts to the list of addresses within the ENS domain. (see image below) Think of ENS domains as if they are decentralized calling cards. You have full ownership of these domains and the only way you lose access to them is if you sign them away, or if you lose access to the internet. We will eventually live in an age where the only thing you need to contact someone whether it be calling, SMS, video chat, and/or e-mail will be with an ENS domain. Another aspect of identity management in relation to ENS is cybersecurity. ENS domains and smart contracts will kill many jobs in IT and cybersecurity. ENS domains and along with some basic education on smart contracts will make cybersecurity seem like child's play in the coming years. Imagine a world where you cannot spoof IP or MAC addresses? With ENS everything is a certainty it cannot be spoofed nor duplicated (unless you fall for the hidden character scam). With ENS the validators will be much more transparent and it will be easy to know if an unauthorized user is trying to gain access to your router, switch, host, and/or server. It will be easy to validate users on a network. In a nutshell; Wireshark, Ettercap, and Fing will be useless in the near future. ENS Domains Will Replace IPv4, IPv6, and MAC Addresses. The problem with our current IP addressing system is that EVERY network device requires an individual IP address and MAC address. Now imagine using one address for every single device you have. Instead of having an IPv4 of 192.168.0.1 for your router and then 192.168.0.3 for your phone, and 192.168.0.4 for your laptop you can simply have “simple.eth” for all devices that can receive an Ethereum blockchain connection. Are you seeing how ENS will kill lots of IT jobs? Here is an example of an IPv4 and IPv6 replacement. You own the ENS domain “john.eth.” You buy a router and an internet subscription. In the near future, all routers will come with Ethereum receivers instead of fixed octets (IPv4) or fixed hexadecimal (IPv6). Your IP address of the router becomes “john.eth” and you can change it out at any time and use the same ENS domain for multiple network devices. It merges MAC addressing and IP addressing into one. ENS is basically unlimited in the number of addresses it can have. ENS Domains Will Help Stabilize High ETH Gas Fees Think of people who own gas stations. Some own 7/11s, some own Shells, and some own Chevron. Anyone who owns a three-letter or three-number ENS domain can lease out subdomains e.g. “sub.123.eth” When people register ENS domains you see three addresses at the top, the registrant (owner), controller (daily operations manager), and the resolver (translates wallet addresses into words and vice-versa). The resolver is basically the one who controls whether your transaction can be translated or not. Now imagine using a resolver that can also stabilize the gas fees? Right now the safest bet is to use the Public Resolver that the ENS team offered, but once the demand for mainstream Ethereum use comes about, then there will be a massive need to stabilize gas fees. Now imagine having a collection of 46,656 three-character alpha-numeric ENS domains (only maximum possible) and leasing out subdomains in them and then the utility you offer is a fixed and stable gas fee that facilitates the transaction. This is where private resolvers will come in (gas stations). Unlike actual petrol gas stations, the private resolvers of ENS will be forced to stick to a fixed rate, FOR LIFE. For example, they could charge 0.02% per transaction on the basis you send a minimum of 1x10^-14 ETH, which would mean the fee would be 2x10^-18 ETH. Ethereum can be divided 18 times. The 0.02% comes from the fact that the private resolver gets 0.01% and the owner of the 3-character alpha-numeric gets the other 0.01%. This can only work if there are multiple private resolver competitors who allow the owners of 3-character alpha-numerics to use each resolver simultaneously, because if there is only one then they are just like the public resolver. Having a Mainnet for Ethereum is practical, but expensive since everyone uses it causing high gas fees and failed transactions. With private resolvers, the only ones who can access the fast and cheap network are the ones who have bought subdomains that have been contracted to private resolver networks. What a private resolver can do is negotiate with the Mainnet and be like “yes we charge very little, but we still have more ETH than you can dream of, so you will have no choice but to approve our transactions.” This can only be possible if people use subdomains within 3-character alpha-numeric ENS domains. Using a domain with a private resolver that offers fixed rates and guaranteed transactions will be huge for high-demand mints going forward. In late April of 2022, Yuga Labs launched the Otherside NFT which strangled the Ethereum network on a level never seen before. People lost lots of money due to failed transactions on the network during the highly anticipated mint. For high-demand mints, in the future, they will need to contract with private resolvers who lease their resolver addresses to owners of 3-character alpha-numeric ENS domain and be like hey “if you have a subdomain with this resolver on the 123.eth domain then you will be able to mint the NFT.” This will reduce the stress on the Mainnet and allow those who are not participating in the mint to use the network without the consequences of someone else’s demand. Soul Bound Tokens In the world today you have problems with fake degrees, fake passports, and fake identities. In the future, we will have Soul Bound Tokens (SBT). SBTs are NFTs that cannot be transferred, sold, and/or altered under any circumstance. If you go to CU Boulder then their ENS name will be “CUBoulder.eth,” and if you get a degree from there let’s say in the business department they will hand out an SBT under the subdomain “business.cuboulder.eth” and then add your name into the subdomain and then add the SBT in the subdomain that has details about you. And then they will freeze the metadata of that SBT making sure the metadata always leads back to them and you. Paper documents will be a thing of the past. SBTs can also be given out as passports. Instead of forging a fake passport, you make an SBT passport whose contents cannot be altered or transferred making sure you are dealing with the real person. An example is “USA.eth” Has the subdomains of the years' people are born and within each subdomain gets added and SBT that is linked to the person, DNA sequence, time, and place of birth of said individuals. SBTs will make individualism a less expensive process. Validating Ownership Of Sellable/Transferrable Assets You own a home and someone forges documents to sell your own home without your knowledge and the problem only gets resolved once the forgery has been proven; how do you even prevent that from happening in the first place? Like this; You own a home, and the company you bought it from has the ENS domain “property.eth” In the subdomain of “property.eth” is your name “john.property.eth.” This subdomain contains the NFT that validates your ownership, so if a scammer wants to borrow against that house, then he will need to provide ownership of the NFT that is linked to the owner's SBT passport. The NFT can be altered, but ownership cannot be changed unless John validates to remove his ownership from the NFT by signing the smart contract. But even then there will be barriers in place, where if John gets scammed into signing away his NFT, the company he bought it from (property.eth) will have a right-of-first-refusal in the smart contract and then will explain to John how he got scammed and why the first-refusal clause in the smart contract is there in the first place. Even if the scammer screenshots the JPEG and tries to delete the metadata that shows his forgery the lenders, buyers, and sellers will have a way of knowing who the NFT actually belongs to before any money is borrowed, sold, or lent. Right now good talkers know how to weasel past the system and get money on assets that don’t belong to them, but in the near future, they will need to prove conclusively that the asset belongs to them. There will be transaction validators that can scan the blockchain to see if the NFT is real or just a screenshot, and then notify the property manager and homeowner. In conclusion, we are only in the beginning stages of web 3. Ethereum will be the undisputed mainstream operating system of web 3. Without ENS the average person will not be able to understand how web 3 will work, just like how they did not understand the internet before the DNS standard came out. Once ENS becomes mainstream people will begin understanding the use cases of NFTs and how they validate authenticity, identity, and ownership. Without ENS, NFTs and many of these web 3 projects will seem useless as history has shown that once people can attach a readable name to things then only will they begin to understand and embrace them. The solution to solving the mystery of web 3 will be ENS plain and simple. Ask yourself how simple and safe web 3 would be if you had readable addresses to direct yourself to? Making people feel safe and secure with new technology is always the first step toward mass adoption and ENS will give people that security of knowing where to turn to. People need directions in order to do something securely and ENS is the most simplified way to navigate new users on web 3.

Practical Use Cases For ENS Domains

Looking for the best AR software for your app? This guide will help you choose from the 10 best AR programs that lead in 2022. Keep reading to find also our tips on what to consider while selecting an augmented reality platform. 4 Top Augmented Reality Platform Technologies Before deciding on the best AR programs, let’s first discuss the most popular AR technologies that you can implement into your app. Marker-Based AR This type of augmented reality is also known as image recognition or recognition-based AR. Augmentation happens when the camera recognizes a required marker. It could be an image, an object, a text, or a sound. A simple example is an AR book. Characters go live when you point a smartphone camera at the book. Face AR works almost the same way as marker-based, but the marker is a face. Face AR You've surely used this augmented reality platform technology. Snapchat filters and are the most vivid examples. Instagram face effects employs face recognition and tracking technology to let users apply 3D filters, Snapchat-like lenses, real-time makeovers, face masks, and avatars to augment user appearances. Face AR software This way, you drive user engagement, increase sessions, and boost in-app sales. This AR software type is a perfect match for the communication, e-commerce, education, advertising, and gaming industries. Let's imagine you run an e-commerce business. You integrate face augmented reality software and offer users a . virtual try-on feature This makes them excited as they can try before they buy. In turn, it boosts user engagement, increases session time, and drives sales. Want to Boost Sales with AR Software?Get a 14-Day Free Trial Right Now Location-Based AR This type of AR software solutions uses a smartphone’s GPS, camera, compass, and accelerometer to tie an object to a specific area. Pokemon GO is the best example. Users could spot virtual creatures at specific locations displayed on an in-app map. Projection-Based AR Frequently referred to as spatial AR, this technology allows the rendering of virtual objects in a physical space on or around the user. Activation happens when a user touches a surface lit by a projector to interact with virtual objects that are integrated into the environment. Users and virtual objects can move within a specified zone in the area where the projector and camera are placed. 10 Best AR Software Solutions in 2022 Meishe Meishe is one of the best Beijing-based augmented reality platforms. It offers face tracking and AR software along with a video editing SDK. The key features of Meishe's AR toolkits are 2D/3D masks, background separation, triggers, beautification, gesture recognition, etc. The core offerings by Meishe: Efficient and accurate face detection Rich sticker effects Beauty makeup Gesture detection Face prop making tools. What Brands Trust Meishe? Asus, Xiaomi, Oppo, Splice, and Vivo are using the AR software solutions by Meishe. How Much Does It Cost? The augmented reality platform offers a free trial period and requires contacting for getting pricing details. What Platforms Are Supported? Meishe's augmented reality SDK supports mobile, web, and desktop platforms. Banuba Face AR SDK Banuba is among the top face AR software focused on face-related augmented reality. The key product is called Face AR SDK. It is an advanced augmented reality platform with multiple features and a list of high-profile users. Face AR is about empowering your business with increased sessions, , and grown sales. better user engagement The 4 key modules of our SDK are: AR face filters Unity face tracking functionality AR beautification features Web AR software. They all are about real-time facial animated Snapchat-like 3D filters. Using face-tracking technology, we offer >1,000 ready-made AR effects. They include makeup, hair color, face morphing, background segmentation, beautification, game filters, LUTs, and . many others What's more, Banuba is among the few commercially available augmented reality kits with a . Try it right now in your browser. background segmentation for web apps What Brands Trust Banuba AR Software? Meta, Gucci, Samsung, Voicemod, Paltalk, and Smilegate are already using augmented reality platforms by Banuba. How Much Does It Cost? fully depends on the set of features you need. The more features you need, the higher is the subscription fee. Our pricing What Platforms Are Supported? Our augmented reality toolkit is compatible with iOS, Android, Windows, Mac, Web, and Unity. DeepAR DeepAR enters the list of the 14 best AR programs with its advanced augmented reality SDK. Since 2014, the company has been employing scientists from MIT, ESIEE Paris, and other institutions to create next-gen AR software. They've created and released multiple augmented reality tools including: AR face filters Beautification & makeup tools Hair and background segmentation Emotion detection Body tracking. The company offers hundreds of ready-made 3D masks that you can try in real-time . here What's more, you can utilize Creator Studio - an all-in-one hub to combine masks, special effects, and filters. Simply put, it's a place where you can generate AR effects yourself. It improves user-generated content and increases user engagement along with in-app sessions. Also, the company offers the advertising augmented reality toolkit. It allows you to utilize premium AR ad units for web, iOS, Android, and Unity platforms. What Brands Trust DeepAR? Google, Coca-Cola, Lego, Tencent, and RayBan are using DeepAR's augmented reality platforms to drive sales. How Much Does It Cost? The product's pricing is based on monthly active users (MAUs). You can use its AR toolkits for free having up to 10 MAUs. What Platforms Are Supported? Supports mobile platforms, desktop, and web. ARGear/Seerslab Seerslab is a Korean AR software company serving augmented reality business needs since 2014. Its main product is ARGear SDK - an augmented reality toolkit for developers. It includes 3 main components: with key features AR SDK to create, manage and monitor your AR content Advanced console with >1,000 AR filters. Augmented reality content Simply put, it's an all-in-one toolbox for your AR business needs. From creating content to managing and monitoring its efficiency. ARGear focuses on 6 use cases: to drive adoption 5G to boost the shopping experience Commerce to drive a fun and unique streaming experience Live Streaming to increase engagement and interactivity Video Calls to empower your app with AR features AR Camera to change the way we see, imagine, and learn. Education What are the core features of AR SDK by ARGear? Facial Landmark Engine Facial Landmark Trigger 2D Animation Sticker / Effects Face Mask 3D Animation Sticker 3D Avatar Virtual Background What Brands Use ARGear? Samsung, eBay, LG, Pepsi, Codac, and Amway are utilizing easy to use augmented reality software by Seerslab. How Much Does It Cost? You can use it for free with limitations. The pricing depends on MAUs, content, and network usage if you need more. What Platforms Are Supported? The AR SDK supports Android, iOS, and Unity platforms. Vuforia One of the best software for augmented reality by Qualcomm, Vuforia serves AR needs successfully. Just to know, renowned Qualcomm was a leading company in building wireless communication means and SoC electronic microchips within its powerful R&D labs. Vuforia specializes in releasing AR software solutions for manufacturing, service, and marketing needs. BAE Systems, Toyota, Volvo Group, Howden, and Cannondale are well-known brands that use Vuforia's augmented reality SDK. How does this SDK work? It uses a QCAR library in C++ as the core element with real objects (image cube, cuboid, cylinder) and target rendering functionality. Simply put, QCAR recognizes real objects and lets your app array virtual objects in the required locations and proportions.  The key SDK features include: Extended Tracking & Smart Terrain Vuforia Object Scanner to recognize 2D/3D objects Text recognition with in-built 100,000-word English vocabulary VuMarks to detect barcodes that can contain data or act as markers Virtual buttons that can turn any surface into a touchscreen Smart glasses support Storing data on the device or in the Cloud. How Much Does It Cost? It has free and paid subscription plans with a free trial: Classic – $499 per time, Cloud – $99 per month, Pro for commercial use, Individual pricing. What Platforms Are Supported? Vuforia AR software solutions support Android, iOS, Universal Windows Platform, and Unity. Wikitude Wikitude is among the best AR programs and frameworks to create augmented reality apps. First launched back in 2008 by an Australian development company, Wikitude holds its top position in serving market AR needs. The company constantly enhances its AR SDK with the latest release in 2020. of the augmented reality toolkit are: The main features Object & scene tracking Instant image tracking Multiple image targets Cloud recognition Geo AR technologies AR bridge for integrations. The platform offers high flexibility for engineers to generate customized solutions and render objects. The one and only limitation is to use JavaScript code along with Wikitude JS or Native SDK. What Platforms Are Supported? The AR SDK supports Android, iOS, and Windows for tablets. How Much Does It Cost? Wikitude provides a free trial with watermarks and paid subscription plans (starting from €1,990 per app). ARKit ARKit is perhaps one of the most popular augmented reality software kits first introduced at WWDC 2017 by Apple. The leader of AR toolkits was designed specifically to create mobile AR apps for the iOS platform. This means ARKit is a perfect choice if you launch apps for iPhones and iPads. What's more, the product supports Arkit Unity with a unique feature: you no longer need to connect additional markers like iBeacon to orientate and route. The key AR SDK features include: 2D image detection and tracking 3D objects and space recognition; Multi shared AR experience for creating multiplayer AR games; Horizontal planes detection Lighting estimation Facial tracking. How Much Does It Cost? The ARKit augmented reality platform is open-source and free. ARCore ARCore is a top AR SDK software and Google's response to Apple's ARKit. The product was released back in 2018 and provides a great difference from its main competitor. ARCore supports both Android and iOS platforms and offers 3 core technologies: Motion tracking Environmental recognition Lighting recognition. They all help you embed virtual content into real environments. The AR SDK software's core is an advanced 3D model. It reacts to any changes in lighting and creates shades. What's more, mobile devices don't affect the model operation in space. This means generated targets are automatically attached to any environmental surfaces. of the augmented reality SDK include: The key features Phone’s motion and position tracking Lighting estimation Surface, size, and location detection to understand an environment Placing texts and virtual objects in a certain spot and sharing them with other app users; Working with Java/OpenGL, Unreal, and Unity; Creating 3D objects with Tilt Brush and Blocks VR building tools. What Platforms Are Supported? One of the best AR programs supports Android 7.0+, iOS 11+, Android NDK, Unity for Android, Unity для iOS, Unreal Engine. How Much Does It Cost? ARCore is completely open-source and free. MaxST MaxST AR SDK software is yet another powerful augmented reality toolkit for mobile development. The product focuses on industrial AR needs serving such global giants as KIA, Hyundai, Samsung, Nokia, LG, and others. MaxST AR software provides 2 different AR SDK software: a 2D toolkit for image tracking and 3D kit for scanning and recognizing surroundings. The latter offers : 8 key functions to place desired content in the space Instant tracker to extract feature points, track surroundings, and save them Visual SLAM to track targets and augment virtual objects on a map Object tracker to augment virtual contents by marker recognition & tracking Marker tracker to recognize target images and augment virtual contents Image tracker for a cloud-based image recognition service Cloud recognizer to augment virtual contents QR code tracker for fast and accurate results. QR/Barcode reader What Platforms Are Supported? MaxST supports Android, iOS, Windows, and macOS platforms. How Much Does It Cost? The augmented reality platform is available for free (with watermarks) and includes paid subscription EasyAR EasyAR is among the top yet underestimated AR software solutions released by an Asian company. It offers 3 AR products: AR Cloud, EasyAR Sense, and CRS. AR Cloud is a cloud-based augmented reality platform that empowers travel, real estate, education, and business marketing domains. EasyAR Sense is an advanced AR SDK to track motion, surface, 3D objects, and planar images. EasyAR Cloud Recognition Service (CRS) is a web platform offering cloud-based image recognition technologies and large databases. Compared to the above AR software solutions, the product doesn't provide an all-in-one built-in object manager. This means you can identify targets by regenerating previously defined 3D models. This way, objects are created in runtime. The key features of EasyAR include: image recognition 3D objects identification Cloud recognition Smart Glass employment AR-based screen recording Transition of files to storages export, import, and integration of assets. What Platforms Are Supported? The augmented reality tools support Android, iOS, UWP, Unity Editor, Windows, macOS. What Brands Use EasyAR? KFC, Vogue, and Pepsi utilize EasyAR solutions to drive their business. How Much Does It Cost? The augmented reality SDK is free but requires a generated key. 5 Tips to Choose Your Augmented Reality Platform In most cases, AR development is both pricey and time-consuming. These reasons make you often choose ready-to-use SDKs as they save resources on integrating AR into your products or services. Now we want to go through 5 tips and things you should consider while choosing AR software solutions. Let's take our Face AR SDK as an example. Launch Date If you need a quick product launch to occupy a niche or gain a competitive advantage in the market, then a proprietary SDK could be the best option. Our AR SDK software saves your development time. The matter is that our face AR features can be quickly integrated into your app. On the other hand, in-house development can take months or even years before these features go live. Development Costs Some open-source solutions like provide you with a set of ready-to-use features and modules. They can give you a good and fast start.But during the development process. OpenCV they require more investments In contrast, you know proprietary SDK licensing costs in advance. They are fixed and let you plan investments and activities transparently. What's more, you can integrate such AR features when you need them without limitations. AR Content Using open-source AR software could be enough for introducing basic AR technologies. But what happens if you need to create original AR content? It may include AR masks, filters, avatars, furniture, or virtual surface enhancements? This takes additional effort and may lead to higher costs and time spent by your business. This is especially true for face AR development. It requires solid niche experience and expertise. In this case, you can benefit from the native object libraries and tools for creating custom AR content like Face Filters SDK. Platforms Platform compatibility is another crucial aspect. For sure, you may want to run your apps on multiple platforms. The more platforms you cover, the more audience you get. But many AR software solutions focus solely on supporting 1 or 2 platforms. This brings limitations to your growth strategy. In contrast, proprietary solutions like Banuba SDK optimize technology for each platform. Feature Stack The more features an SDK offers, the more you can scale your app. But scalability may become an issue if AR software solutions solve narrow tasks only. This means you should choose a large-scale AR SDK with potential features you may want to add in the future. For sure, you may initially overpay for those features you don't currently use. But you will win anyway. Replacing your infrastructure and changing the AR software provider is way more pricey. Conclusion Augmented reality allows extending, enhancing, and modifying real-world objects around us via technology. And we hope our list of the 10 best AR programs helped you choose the most easy to use augmented reality software for your business. The blog post was originally published at . Banuba Blog

10 Best AR Software for Your Business App in 2022

Meet the Winners of Startups in North and Central America

Introduction to 3D Javascript Library: Three.js Basics

Amazon SageMaker is a popular and full-managed service by Amazon that allows developers and data scientists to build, train and deploy machine learning models in a quick and hassle-free manner. This platform offers multiple advantages to the end-users such as removal of heavy lifting during every step of the machine learning process which makes it easy to deploy the quality machine learning models: •	One can build his own model for labeling workflows •	The Amazon SageMaker Studio provides the first fully integrated development environment for machine learning •	The platform can be leveraged by the data scientists and ML Engineers to build and deploy their own MXNet or TensorFlow model •	The marketers can automatically create high-quality machine learning models with the help of Amazon SageMaker Autopilot •	One can also select the ML Instances on the Fly Using Amazon SageMaker, one can build, train and deploy machine learning models with the help of the XGBoost ml algorithm. Amazon SageMaker is a fully managed service that allows developers and data scientists to build, train and deploy machine learning models quickly. Training ML models from conceptualization to production is often complex and time-consuming. One has to manage large chunks of data to train a model using the best of algorithms for training and has also to manage the compute capacity while training the model and then deploy the model in an appropriate production environment. The traditional machine learning development process is a very complex one. It's expensive and iterative and is also very tough because of the absence of any tools for the entire ML workflow. Stitching the tools and the workflows together is time-consuming and also error-prone. Amazon SageMaker reduces the complexity by making it easier to build and deploy the ML models. SageMaker also provides all the components used for machine learning in a single toolset and allows the models to get to faster production with much less effort and at much lower costs. The users can choose the right algorithms and frameworks from a wide array of choices available. SageMaker manages all the innate and fundamental infrastructure required to train a model at a petabyte scale and helps deploy the model to production. Advantages of Building Machine Learning Models Using SageMaker 1. Reduced Cost of Data Labeling With Amazon SgaeMaker Ground Truth With the help of Amazon SageMaker Ground Truth, marketers can build and manage highly accurate training datasets quickly. Amazon claims that SageMaker Ground Truth allows marketers to . One can use exclusive labelers or use vendors recommended by Amazon through AWS Marketplace. Additionally, continuously learns from labels provided by humans to make high-quality, automatic annotations that significantly lower the costs of labeling. reduce data labeling costs by up to 70% This fully managed data labeling service by Amazon allows marketers to easily label training data for machine learning at scale. Marketers can build highly precise training datasets to get started with their machine learning models in minutes through the SageMaker Ground Truth console. This can be done by using custom or built-in data labeling workflows. These workflows support a wide array of use cases such as 3d point clouds, video, images, and text. As part of workflows, labelers have access to assistive labeling features such as automatic 3D cuboid snapping, removal of distortion in 2D images, and auto-segment tools to reduce the time required to label datasets. Moreover, Ground Truth offers automatic data labeling that utilizes machine learning models to label your data. 2. Models Can Be Built Using a Single Web Interface With the help of Amazon SageMaker, users can easily build machine learning (ML) models at scale and can train them. One gets access to everything needed to quickly load data, create new notebooks, and use built-in algorithms and frameworks. Amazon SageMaker Studio is the first fully integrated development environment for machine learning. Users can quickly upload data, create new notebooks, can train and tune models, move back and forth between steps to adjust experiments and compare results, and can deploy all the models to production in one place. In order to automatically generate models, one can use Amazon SageMaker Autopilot along with SageMaker Studio to automatically generate models. SageMaker Autopilot is the first automated machine learning capability provider that gives users complete control and visibility into their ML models. SageMaker Autopilot automatically derives inferences from raw data, applies feature processors, picks the best set of algorithms, and trains and tunes multiple models. SageMaker Autopilot also tracks the performance of different models and then ranks the models within just a few clicks. Amazon SageMaker Notebooks are one-click, sharable Jupyter notebooks that can be quickly spun up. One can choose from dozens of pre-built notebooks within SageMaker or from hundreds of algorithms and pre-trained models that are available in AWS Marketplace. 3. Amazon SageMaker Supports the Leading Deep Learning Frameworks With the help of Amazon SageMaker, users can use some of the most popular open-source Deep Learning Frameworks for their use cases and can get the best results. Some of these frameworks include TensorFlow, PyTorch, mxnet, Chainer, Keras, Gluon, sckit-learn, HDROVOD, and DeepGraphLibrary. Users can run their models on Amazon SageMaker by putting up their code in a docker file right. That’s actually bringing your own docker file putting your code inside that docker file and then registering that on ECR. Another way to train models on SageMaker is by using script mode. Building a Sample Machine Learning Model For A Bank to Predict Whether a Customer Will Enroll for A Certificate Such a model can be trained on a . This dataset has information on customer demographics, responses to marketing events, and external factors. The data has been labeled for the convenience of the users and a column in the dataset identifies whether a customer has enrolled for a product offered by the bank. This dataset is from the Machine Learning Repository curated by the University of California, Irvine. Bank Marketing Dataset publicly accessible For building this model, a user needs to and create an account. Thereafter the following steps need to be followed: sign up for AWS Step 1: Data Preparation by Creating AWS SageMaker Notebook Instance In this step, the users create the notebook instance that they can download and process their data. As a part of the data preparation process, users create an Identity and Access Management (IAM) role that allows Amazon SageMaker to access data in Amazon S3. Once you log into the , in the top right corner select your preferred . For example, US West (Oregon) can be taken up as one sample region. Amazon SageMaker console AWS Region In the left navigation pane, select , and then choose to . Notebook instances Create notebook instance Once on the page, in the box, one needs to fill up the following fields:
a)	For , type .
b)	For , choose .
c)	For , one must keep the default section as . Create notebook instance Notebook instance setting Notebook instance name SageMaker-Tutorial Notebook instance type ml.t2.medium Elastic inference none Choose , in the section, for . The dialog box has an option that needs to be chosen, and then needs to be selected. Those who already have a bucket that they’d like to use can choose and then can specify the bucket name. to Create a new role Permissions and encryption the IAM role Create an IAM role Any S3 bucket Creates role Specific S3 buckets Amazon SageMaker creates the *-***role*. AmazonSageMaker-ExecutionRole Users can keep the default settings for the remaining options and choose . to Create a notebook instance In the section, the new instance is displayed under the status: “Status of Pending.” The notebook is ready when the Status changes to . Notebook instances SageMaker-Tutorial In-Service Step 2: Data Preparation In this step, marketers use the Amazon SageMaker notebook instance to preprocess the data that you need to train your machine learning model and then upload the data to Amazon S3. After the users’ notebook instance status changes to , they can choose the option. SageMaker-Tutorial InService Open Jupyter In , marketers can choose New and then can choose . the Jupyter notebook conda_python3 Marketers in a new code cell on their Jupyter notebook should run the code above and choose . Run This code imports the required libraries and then defines the environmental variables that the users need to prepare the data, train their ML models and deploy them. Marketers can create the S3 bucket to store their data. They can run the code above into the next code cell. Marketers must ensure that they replace the bucket_name your-s3-bucket-name with a unique S3 bucket name. If one doesn't receive a success message after running the code, one must change the bucket name and try again. Users can download the data into a data frame and can run the code above into the next code cell. Marketers can shuffle and split the data into the training data and test data by running the code above. The is used by 70% of the customers during the model training loop and then can use gradient-based optimization to iteratively refine the model parameters. Gradient-based optimization allows users to find model parameter values that minimize the model error, using the gradient of the model loss function. The remaining 30% of the customers comprise the test data. This data is used to evaluate the performance of the model and measure how well the trained model generalizes to unseen data. training data Step 3: Train the ML Model This step is all about marketers using their training dataset to train their machine learning models. In a new code cell on their Jupyter notebook, marketers can copy and paste the following code and choose Run. This code reformats the header and the first column of the training data and then loads the data from The S3 bucket. This step is required to use the Amazon SageMaker pre-built XGBoost algorithm. In order to set up the Amazon SageMaker session, marketers must create an instance of the XGBoost model (an estimator), and define the model’s hyperparameters. Then the marketers must copy and paste the code shown above and choose Run. To start the training job, marketers must copy and paste the following code into the next code cell and choose Run. This code trains the model using gradient optimization on an ml.xlarge instance. After a few minutes, marketers can witness the training logos that are being generated in their Jupyter notebook. Step 4: Deploying the Model In this step, marketers deploy the trained model to an endpoint, reformat and load the CSV data and then run the model so that the predictions can be created. Users must copy and paste the following code in a new cell on their Jupyter notebook and then must choose . Run This code deploys the model on a server and creates a SageMaker endpoint that anyone can access.
The entire process may take a few minutes to complete. In order to predict whether customers in the test data enrolled for the bank product or not, the marketers can copy the following code into the next code cell and choose . Run Step 5: Evaluate Model Performance This step allows marketers to evaluate the performance and accuracy of their machine learning models. The users can run the following code on their Jupyter notebook. This code enables the comparison of actual vs. predicted values in a table called . confusion matrix Based on the prediction, the marketers can conclude whether a predicted customer will enroll for a certificate of deposit accurately for 90% of the customers in the test data, with a precision of 65% (278/429) for enrolled and 90% (10, 785/11,928) for didn’t enroll. Step 6: Data Cleansing In this step, users terminate the resources used so far in the tab. Terminating the resources that haven't been effectively and optimally used reduces cost and is a best practice. Not terminating the resources causes changes in your account. – Users can run the following code. Deletion of Your Endpoint Next up the users need to delete their training artifacts and S3 bucket. They must run the following code: After the above step, the users must stop & delete their SageMaker Notebook by following the steps below:
Step 1: Open the .
Step 2: Choose Notebook Instances from the option “Notebook.”
Step 3: Choose the notebook instance that you have used for creating the predictive analytical insights and then choose . The instance may take up several minutes to stop. When the changes to , move on to the next step.
Step 4: From SageMaker Console Actions, Stop Status Stopped Actions choose Delete. Wrap Up This blog article clearly helps comprehend how users can use Amazon SageMaker to easily build, train and deploy ML models. This powerful service by AWS helps users with its exemplary modular design. SageMaker is used for model training in pre-configured environments and also facilitates model training by hyper-parameter tuning. Amazon's innate pre-built models have been highly optimized to run on AWS services. These models are pre-built and nobody needs to do much to build and check the model. Marketers can pre-build XGBoost or LDA or PCA or Seq to Seq models all of which are available in SageMaker. Moreover, SageMaker is also used for model deployment and has multi-model endpoints. Users can easily build ML models by providing everything that they need to. They can quickly connect to their training data and can select the best algorithm and framework for their application. As all of the underlying infrastructures are managed by AWS, the models can be trained at a petabyte scale.

Building Machine Learning Models Using AWS SageMaker 

is one of the most important tasks in computer vision. With numerous applications, computer vision essentially strives to give a machine eyes – the ability to see and interpret the world. At times, machine learning projects seem to unlock futuristic technology we never thought possible. AI-powered applications like augmented reality, automatic speech recognition, and neural machine translation have the potential to change lives and businesses around the world. Likewise, the technologies that computer vision can give us ( , , ) are extraordinary. Image annotation autonomous vehicles facial recognition unmanned drones However, none of these amazing computer vision technologies would be possible without image annotation. This article will explain what image annotation is and five image annotation services provided by numerous training data companies around the world. What is Image Annotation? Image annotation is the human-powered task of annotating an image with labels. These labels are predetermined by the AI engineer and are chosen to give the computer vision model information about what is shown in the image. Depending on the project, the amount of labels on each image can vary. Some projects will require only one label to represent the content of an entire image (image classification). Other projects could require multiple objects to be tagged within a single image, each with a different label. How Does Image Annotation Work? To create annotated images you need three things: Images Someone to annotate the images A platform to annotate the images on Most image annotation projects begin with sourcing and training annotators to perform the annotation tasks. AI is a very specialized field, but doesn’t always have to be. While you need higher education in machine learning to be able to create a self-driving car, you don’t need a master’s degree to draw boxes around cars in images (bounding box annotation). Thus, most annotators don’t have degrees in machine learning. AI training data annotation However, these , as every company will have different requirements. Once the annotators are trained on how to annotate the data, they will get to work annotating hundreds or thousands of images on a platform dedicated to image annotation. This platform is a software that should have all the necessary tools for the specific type of annotation being performed. annotators should be thoroughly trained on the specifications and guidelines of each annotation project 5 Common Image Annotation Services 1. 2D and 3D Bounding Boxes With 2D bounding boxes, annotators must draw a box around the object they want to annotate within the image. Sometimes these target objects will be the same, i.e. “Please draw boxes around every bicycle in this image.” Other times, there may be more than one target object, “Please draw boxes around every car, pedestrian, and bicycle in this image.” In those cases, after drawing the box, the annotator would then have to choose from a list of labels to attribute to the object within the box. Also known as cuboids, 3D bounding boxes are almost the same as 2D bounding boxes except that they also can show approximate depth of the target objects being annotated. Similar to 2D bounding box annotations, annotators draw boxes around the target objects, making sure to place anchor points at the object's edges. Sometimes a portion of the target object may be blocked. In such cases, the annotators would approximate the location of the target object’s blocked edge(s). 2. Image Classification Whereas bounding boxes deal with annotating multiple objects in an image, Image classification is the process of associating an entire image with just one label. A simple example of image classification is labeling types of animals. Annotators would be given images of animals and asked to classify or categorize each image based on the animal species. Feeding this annotated image data to a computer vision model would teach the model the visual characteristics unique to each type of animal. In theory, the model would then be able to categorize new unannotated animal images into the proper species categories. 3. Lines and Splines As their title suggests, lines and splines annotation is the labeling of straight or curved lines on images. Annotators would be tasked with annotating lanes, sidewalks, power lines, and other boundary indicators. Images annotated with lines and splines are mainly used for lane and boundary recognition. As well, they are also often used for trajectory planning in drones. From autonomous vehicles and drones to robotics in warehouses and more, lines and splines annotations are useful in a variety of use cases. 4. Polygons Sometimes target objects with irregular shapes can’t be easily annotated with bounding boxes or cuboids. Polygon annotation allows annotators to plot points on each vertex of the target object. This annotation method allows all of the object’s exact edges to be annotated, regardless of its shape. Like bounding boxes, the pixels within the annotated edges would then be tagged with a label to describe the target object. 5. Semantic Segmentation Bounding boxes, cuboids, and polygons all deal with the task of annotating individual objects in an image. However, semantic segmentation is the annotation of every pixel within an image. Instead of giving annotators a list of objects to be annotated, they are given a list of segment labels to divide the image into. A good example is semantic segmentation in traffic images for autonomous vehicles. A typical semantic segmentation task could ask annotators to “segment the image by vehicles, bicycles, pedestrians, obstacles, sidewalks, roads, and buildings”. Each segment is usually indicated by a unique color code. Annotators would draw lines around the pixels they want to annotate and select the appropriate label. The end result would look something like this: Image via medium.com/intro-to-artificial-intelligence Hopefully, this article helped you understand the basics of five in-demand . If you’re looking for more reading on image annotation and AI, be sure to check out: image annotation services in machine learning Image Annotation - an overview Notes on Image Annotation Image Annotation for Video Games

Stories, News & Mentions about CUBO

Cubo’s Stories on HackerNoon