John Fury has claimed the only man who would get him in the ring is the legendary Mike Tyson - but warned he will need 'big boy money' to face the fighter he named his son after. John's son Tyson ...

John Fury claims Mike Tyson is the ONLY man who would get him in the ring but warns he would need 'big boy money' to face the fighter he named his son after... as he predicts ...

P.S Got a problem with your brand? Get it sorted quickly with the newly launched 'Brand Power Hour.' Squark Design’s Brand Power Hour Banner. Get a suction to your branding problems FAST!

NEW TOPIC: Make your Visual Content Accessible for everyone

The Brand Identity Essentials Package Elevate Your Brand Consistency on a Budget! Give your brand the consistency it needs without breaking the bank.Our limited-time ‘Brand Identity Essentials ...

Your Brand Identity Essentials Package is here!

This week, we're diving into the importance of maintaining a consistent and cohesive visual style for your brand imagery. We will briefly explore why this matters and the role of high-quality ...

Maintaining a consistent and cohesive visual style

So far in our Brand Identity topic we have looked at colour and brand personality. This week we will look at another important area in creating a memorable brand - and that is… When it comes to ...

Typography: The silent ambassador of your brand

Accent colour: The third colour used to complement the main and secondary colours. Michelle Jones, Squark Design Founder and Visual Brand Storyteller P.S Did you get the 3 well known logo brands ...

Would you know a brand just by it’s colour?

We study the effect of squark-generation mixing on production and decays of squarks and gluinos at LHC in the Minimal Supersymmetric Standard Model (MSSM). We show that the mixing effects can be very large in a significant range of the squark-generation ...

Impact of squark generation mixing on the search for squarks and gluinos at LHC

Keep up to date with all the matches around the county grounds

County cricket: Gloucestershire v Glamorgan, Essex v Kent, and more – live

Traditional finance is betting big on bitcoin this year Bob Diamond, former Barclays PLC (LSE:BARC) chief and part of traditional finance’s old guard, took to CNBC’s Squark Box on Wednesday to give his take on the recent insurgence into the ...

Former Barclays boss weighs in on TradFi’s crypto coup

Rebecca is an assistant news editor for Express.co.uk. She started her career in local newspapers where she stayed for four years before moving to digital newspapers in 2014.

Rebecca Perring

squark

**SINGAPORE, Singapore, October 3rd, 2024/CyberNewsWire/--**At DEF CON 32, the SquareX research team delivered a hard-hitting presentation titled Sneaky Extensions: The MV3 Escape Artists where they shared their findings on how malicious browser extensions are bypassing Google’s latest standard for building chrome extensions: Manifest V3 (MV3)’s security features, putting millions of users and businesses at risk. SquareX’s research team publicly demonstrated rogue extensions built on MV3. The key findings include: Extensions can steal live video streams, such as those from Google Meet and Zoom Web, without requiring special permissions.
The rogue extensions can act on a user’s behalf to add collaborators to private GitHub repositories.
The extensions are capable of hooking into login events to redirect users to a page disguised as a password manager login.
Extensions built on MV3 can steal site cookies, browsing history, bookmarks, and download history with ease, like their MV2 counterparts.
The rogue extensions can add pop-ups to the active webpage, such as fake software update prompts, tricking users into downloading malware. Browser extensions have long been a target for malicious actors — a Stanford University report estimates that 280 million malicious Chrome extensions were installed in recent years. Google has struggled to address this issue, often relying on independent researchers to identify malicious extensions. In some cases, Google has had to manually remove them, such as the 32 extensions taken down in June last year. By the time they were removed, these extensions had already been installed 75 million times. Most of these issues arose because the Chrome extension standard, Manifest Version 2 (MV2), was riddled with loopholes that granted extensions excessive permissions, and allowed scripts to be injected on the fly, often without users' knowledge. This allowed malicious actors to easily exploit these vulnerabilities to steal data, inject malware, and access sensitive information. MV3 was introduced to address these problems by tightening security, limiting permissions, and requiring extensions to declare their scripts beforehand. However, SquareX’s research shows that MV3 falls short in many critical areas, demonstrating how attackers are still able to exploit minimal permissions to carry out malicious activity. Both individual users and enterprises are exposed, even under the newer MV3 framework. Today's security solutions, such as endpoint security, SASE/SSE, and Secure Web Gateways (SWG), lack visibility into installed browser extensions. There is currently no mature tool or platform capable of dynamically instrumenting these extensions, leaving enterprises without the ability to accurately assess whether an extension is safe or malicious. SquareX is committed to the highest level of cybersecurity protection for enterprises and has built key innovative features to solve this problem, which include; Fine grained policies to decide which extensions to allow / block and parameters include extension permissions, creation date, last update, reviews, ratings, user count, author attributes etc
SquareX blocks network requests sent by extensions at run time – based on policies, heuristics and machine learning insights
SquareX is also experimenting with dynamic analysis of Chrome Extensions using a modified Chromium browser in its cloud server These are part of SquareX’s Browser Detection and Response solution which is being deployed at medium-large enterprises and is effectively blocking these attacks. Vivek Ramachandran, Founder & CEO of SquareX, warned about the mounting risks: “Browser extensions are a blind spot for EDR/XDR and SWGs have no way to infer their presence. This has made browser extensions a very effective and potent technique to silently be installed and monitor enterprise users, and attackers are leveraging them to monitor communication over web calls, act on the victim’s behalf to give permissions to external parties, steal cookies and other site data and so on.” “Our research proves that without dynamic analysis and the ability for enterprises to apply stringent policies, it will not be possible to identify and block these attacks. Google MV3, though well intended, is still far away from enforcing security at both a design and implementation phase,” said Vivek Ramachandran. About SquareX SquareX helps organizations detect, mitigate and threat-hunt client-side web attacks happening against their users in real time. SquareX's industry-first Browser Detection and Response (BDR) solution, takes an attack-focused approach to browser security, ensuring enterprise users are protected against advanced threats like malicious QR Codes, Browser-in-the-Browser phishing, macro-based malware, malicious extensions and other web attacks encompassing malicious files, websites, scripts, and compromised networks. With SquareX, enterprises can also provide contractors and remote workers with secure access to internal applications, enterprise SaaS, and convert the browsers on BYOD / unmanaged devices into trusted browsing sessions. Contact Head of PR Junice Liew SquareX junice@sqrx.com This story was distributed as a release by CyberNewsWire under HackerNoon’s Business Blogging Program. Learn more about the program here **SINGAPORE, Singapore, October 3rd, 2024/CyberNewsWire/--**At DEF CON 32, the SquareX research team delivered a hard-hitting presentation titled Sneaky Extensions: The MV3 Escape Artists where they shared their findings on how malicious browser extensions are bypassing Google’s latest standard for building chrome extensions: Manifest V3 (MV3)’s security features, putting millions of users and businesses at risk. SquareX SquareX SquareX’s research team publicly demonstrated rogue extensions built on MV3. The key findings include: Extensions can steal live video streams, such as those from Google Meet and Zoom Web, without requiring special permissions. The rogue extensions can act on a user’s behalf to add collaborators to private GitHub repositories. The extensions are capable of hooking into login events to redirect users to a page disguised as a password manager login. Extensions built on MV3 can steal site cookies, browsing history, bookmarks, and download history with ease, like their MV2 counterparts. The rogue extensions can add pop-ups to the active webpage, such as fake software update prompts, tricking users into downloading malware. Extensions can steal live video streams, such as those from Google Meet and Zoom Web, without requiring special permissions. The rogue extensions can act on a user’s behalf to add collaborators to private GitHub repositories. The extensions are capable of hooking into login events to redirect users to a page disguised as a password manager login. Extensions built on MV3 can steal site cookies, browsing history, bookmarks, and download history with ease, like their MV2 counterparts. The rogue extensions can add pop-ups to the active webpage, such as fake software update prompts, tricking users into downloading malware. Browser extensions have long been a target for malicious actors — a Stanford University report estimates that 280 million malicious Chrome extensions were installed in recent years. Google has struggled to address this issue, often relying on independent researchers to identify malicious extensions. report report In some cases, Google has had to manually remove them, such as the 32 extensions taken down in June last year. By the time they were removed, these extensions had already been installed 75 million times. Most of these issues arose because the Chrome extension standard, Manifest Version 2 (MV2), was riddled with loopholes that granted extensions excessive permissions, and allowed scripts to be injected on the fly, often without users' knowledge. 32 extensions 32 extensions This allowed malicious actors to easily exploit these vulnerabilities to steal data, inject malware, and access sensitive information. MV3 was introduced to address these problems by tightening security, limiting permissions, and requiring extensions to declare their scripts beforehand. However, SquareX’s research shows that MV3 falls short in many critical areas, demonstrating how attackers are still able to exploit minimal permissions to carry out malicious activity. Both individual users and enterprises are exposed, even under the newer MV3 framework. Today's security solutions, such as endpoint security, SASE/SSE, and Secure Web Gateways (SWG), lack visibility into installed browser extensions. There is currently no mature tool or platform capable of dynamically instrumenting these extensions, leaving enterprises without the ability to accurately assess whether an extension is safe or malicious. SquareX is committed to the highest level of cybersecurity protection for enterprises and has built key innovative features to solve this problem, which include; Fine grained policies to decide which extensions to allow / block and parameters include extension permissions, creation date, last update, reviews, ratings, user count, author attributes etc SquareX blocks network requests sent by extensions at run time – based on policies, heuristics and machine learning insights SquareX is also experimenting with dynamic analysis of Chrome Extensions using a modified Chromium browser in its cloud server Fine grained policies to decide which extensions to allow / block and parameters include extension permissions, creation date, last update, reviews, ratings, user count, author attributes etc SquareX blocks network requests sent by extensions at run time – based on policies, heuristics and machine learning insights SquareX is also experimenting with dynamic analysis of Chrome Extensions using a modified Chromium browser in its cloud server These are part of SquareX’s Browser Detection and Response solution which is being deployed at medium-large enterprises and is effectively blocking these attacks. Browser Detection and Response Browser Detection and Response Vivek Ramachandran, Founder & CEO of SquareX, warned about the mounting risks: “Browser extensions are a blind spot for EDR/XDR and SWGs have no way to infer their presence. This has made browser extensions a very effective and potent technique to silently be installed and monitor enterprise users, and attackers are leveraging them to monitor communication over web calls, act on the victim’s behalf to give permissions to external parties, steal cookies and other site data and so on.” Vivek Ramachandran , Founder & CEO of SquareX , warned about the mounting risks: “Browser extensions are a blind spot for EDR/XDR and SWGs have no way to infer their presence. This has made browser extensions a very effective and potent technique to silently be installed and monitor enterprise users, and attackers are leveraging them to monitor communication over web calls, act on the victim’s behalf to give permissions to external parties, steal cookies and other site data and so on.” Vivek Ramachandran Vivek Ramachandran SquareX SquareX “Our research proves that without dynamic analysis and the ability for enterprises to apply stringent policies, it will not be possible to identify and block these attacks. Google MV3, though well intended, is still far away from enforcing security at both a design and implementation phase,” said Vivek Ramachandran. “Our research proves that without dynamic analysis and the ability for enterprises to apply stringent policies, it will not be possible to identify and block these attacks. Google MV3, though well intended, is still far away from enforcing security at both a design and implementation phase,” said Vivek Ramachandran. About SquareX SquareX helps organizations detect, mitigate and threat-hunt client-side web attacks happening against their users in real time. SquareX SquareX SquareX's industry-first Browser Detection and Response (BDR) solution, takes an attack-focused approach to browser security, ensuring enterprise users are protected against advanced threats like malicious QR Codes, Browser-in-the-Browser phishing, macro-based malware, malicious extensions and other web attacks encompassing malicious files, websites, scripts, and compromised networks. With SquareX, enterprises can also provide contractors and remote workers with secure access to internal applications, enterprise SaaS, and convert the browsers on BYOD / unmanaged devices into trusted browsing sessions. Contact Head of PR Junice Liew SquareX junice@sqrx.com This story was distributed as a release by CyberNewsWire under HackerNoon’s Business Blogging Program. Learn more about the program here This story was distributed as a release by CyberNewsWire under HackerNoon’s Business Blogging Program. Learn more about the program here here here

SquareX Shows How Malicious Extensions Bypass Google’s MV3 Restrictions

SquareX يوضح كيف تتجاوز الإضافات الضارة قيود MV3 الخاصة بشركة Google

SquareX показва как злонамерените разширения заобикалят MV3 ограниченията на Google

SquareX muestra cómo las extensiones maliciosas eluden las restricciones MV3 de Google

SquareX montre comment les extensions malveillantes contournent les restrictions MV3 de Google

A SquareX megmutatja, hogy a rosszindulatú bővítmények hogyan kerülik meg a Google MV3 korlátozásait

SquareX mostra come le estensioni dannose aggirano le restrizioni MV3 di Google

SquareX、悪質な拡張機能がGoogleのMV3制限を回避する方法を紹介

SquareX აჩვენებს, თუ როგორ აცილებენ მავნე გაფართოებები Google-ის MV3 შეზღუდვებს

SquareX покажува како злонамерните екстензии ги заобиколуваат ограничувањата за MV3 на Google

SquareX laat zien hoe kwaadaardige extensies de MV3-beperkingen van Google omzeilen

SquareX mostra como extensões maliciosas contornam as restrições MV3 do Google

SquareX waxay tusinaysaa sida Fidinta xaasidnimada leh ay uga gudubto Google's MV3

In this exclusive interview, Andrey Nayman, co-founder of Magic Square, shares his journey in the blockchain industry and the inspirations behind creating a community-driven App Store. Ishan Pandey: Hi Andrey, great to have you here. To start off, could you share a bit about your background and what inspired you to co-found Magic Square? Andrey Nayman: I've been in the blockchain industry since 2013. My experiences leading various tech companies inspired me to start Magic Square. We're building a community-driven App Store that makes using decentralized apps simpler and safer. I’m passionate about using my blockchain knowledge to create something valuable and innovative for users worldwide. Ishan Pandey: Magic Square offers a unique ecosystem for discovering Web3 projects. Can you explain what sets Magic Store apart from other platforms in the space? Andrey Nayman: Magic Square is all about putting users first. Imagine it like a big online store, but instead of selling physical items, we present Web3 Apps. Magic Store is a way for users to discover all trusted Dapps and engage with them in one place. What makes us different is that we check every app really carefully before putting it in our store. This makes our platform safe and easy to use, whether you're new to Web3 or already an expert. Ishan Pandey: Given the diverse offerings within the Magic Square ecosystem like Magic Karma and Hot Offers, how do these features enhance the user experience and engagement? Andrey Nayman: Magic Karma and Hot Offers are innovative features designed to enhance the user experience on Magic Square. Magic Karma serves as our reputation system – users receive daily rewards for contributing to the Magic Ecosystem, such as trying out new apps or participating in community activities, earning them points. These points can then be exchanged for exciting rewards. On the other hand, Hot Offers form our engagement layer, where users can complete tasks provided by the listed projects and receive real payments for their efforts. Together, these features keep our users happy and engaged while using our platform. Ishan Pandey: The $SQR token plays a central role in the Magic Square community. Could you discuss how it integrates into the ecosystem and what its benefits are for users? Andrey Nayman: The $SQR token is like the heart of Magic Square. Users earn $SQR as a reward for doing things on our platform every day. You can stake $SQR to unlock special benefits if you're a member. You'll get paid in $SQR when you earn rewards through Hot Offers. We also use $SQR to reach the right people in our marketing efforts. To join Launchpad sales, you must stake even more $SQR. So, $SQR is important for everyone using Magic Square. It's what keeps everything running smoothly and helps our community grow. Ishan Pandey: Community-driven platforms depend heavily on user participation. How does Magic Square encourage community involvement in project voting and validation? Andrey Nayman: Community is at the core of every crypto project. At Magic Square, we believe in bringing the crypto community together in one place. Here, users can decide which projects make it onto our platform, shaping their future. Our community moderators actively gather feedback, which our product team uses to create a roadmap aligned with their needs. To make our community feel valued and engaged, we offer daily rewards through Karma Points and exciting opportunities through Hot Offers. We also host regular surveys where users can share their ideas and feedback, ensuring our platform evolves based on their input. Ishan Pandey: Affiliate programs are common in Web3, but Magic Boost offers a unique twist. Can you elaborate on how this program works and what its advantages are for both users and developers? Andrey Nayman: Affiliate programs are common in Web3, but Magic Boost offers a unique twist. Here's how it works: Users can easily monetize their communities by sharing affiliate offers from listed projects. With one click, they share these offers on social media. When their friends engage and convert, users receive instant payouts. This benefits the user by allowing them to earn effortlessly while supporting their favorite projects. For developers, Magic Boost expands their reach and boosts user acquisition, enhancing app visibility and credibility. Ishan Pandey: Looking forward, what are some key developments or milestones that the Magic Square community can anticipate in the near future? Andrey Nayman: In the near future, our Launchpad will be a big focus. It's designed to help projects grow steadily and connect with our community. We'll customize our approach, using both public and private sales, to give projects the best chance of success. With our active community, we'll make sure these projects get plenty of attention. Stay tuned for more updates as we work together to shape the future of Web3! Don’t forget to like and share the story! Vested Interest Disclosure: This author is an independent contributor publishing via our business blogging program. HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYOR. In this exclusive interview, Andrey Nayman, co-founder of Magic Square, shares his journey in the blockchain industry and the inspirations behind creating a community-driven App Store. Ishan Pandey: Hi Andrey, great to have you here. To start off, could you share a bit about your background and what inspired you to co-found Magic Square? Ishan Pandey: Hi Andrey, great to have you here. To start off, could you share a bit about your background and what inspired you to co-found Magic Square? Ishan Pandey: Hi Andrey, great to have you here. To start off, could you share a bit about your background and what inspired you to co-found Magic Square? Andrey Nayman: I've been in the blockchain industry since 2013. My experiences leading various tech companies inspired me to start Magic Square. We're building a community-driven App Store that makes using decentralized apps simpler and safer. I’m passionate about using my blockchain knowledge to create something valuable and innovative for users worldwide. Andrey Nayman: Ishan Pandey: Magic Square offers a unique ecosystem for discovering Web3 projects. Can you explain what sets Magic Store apart from other platforms in the space? Ishan Pandey: Magic Square offers a unique ecosystem for discovering Web3 projects. Can you explain what sets Magic Store apart from other platforms in the space? Ishan Pandey: Magic Square offers a unique ecosystem for discovering Web3 projects. Can you explain what sets Magic Store apart from other platforms in the space? Andrey Nayman: Magic Square is all about putting users first. Imagine it like a big online store, but instead of selling physical items, we present Web3 Apps. Magic Store is a way for users to discover all trusted Dapps and engage with them in one place. What makes us different is that we check every app really carefully before putting it in our store. This makes our platform safe and easy to use, whether you're new to Web3 or already an expert. Andrey Nayman: Ishan Pandey: Given the diverse offerings within the Magic Square ecosystem like Magic Karma and Hot Offers, how do these features enhance the user experience and engagement? Ishan Pandey: Given the diverse offerings within the Magic Square ecosystem like Magic Karma and Hot Offers, how do these features enhance the user experience and engagement? Ishan Pandey: Given the diverse offerings within the Magic Square ecosystem like Magic Karma and Hot Offers, how do these features enhance the user experience and engagement? Andrey Nayman: Magic Karma and Hot Offers are innovative features designed to enhance the user experience on Magic Square. Magic Karma serves as our reputation system – users receive daily rewards for contributing to the Magic Ecosystem, such as trying out new apps or participating in community activities, earning them points. These points can then be exchanged for exciting rewards. On the other hand, Hot Offers form our engagement layer, where users can complete tasks provided by the listed projects and receive real payments for their efforts. Together, these features keep our users happy and engaged while using our platform. Andrey Nayman: Ishan Pandey: The $SQR token plays a central role in the Magic Square community. Could you discuss how it integrates into the ecosystem and what its benefits are for users? Ishan Pandey: The $SQR token plays a central role in the Magic Square community. Could you discuss how it integrates into the ecosystem and what its benefits are for users? Ishan Pandey: The $SQR token plays a central role in the Magic Square community. Could you discuss how it integrates into the ecosystem and what its benefits are for users? Andrey Nayman: The $SQR token is like the heart of Magic Square. Users earn $SQR as a reward for doing things on our platform every day. You can stake $SQR to unlock special benefits if you're a member. You'll get paid in $SQR when you earn rewards through Hot Offers. We also use $SQR to reach the right people in our marketing efforts. Andrey Nayman: To join Launchpad sales, you must stake even more $SQR. So, $SQR is important for everyone using Magic Square. It's what keeps everything running smoothly and helps our community grow. Ishan Pandey: Community-driven platforms depend heavily on user participation. How does Magic Square encourage community involvement in project voting and validation? Ishan Pandey: Community-driven platforms depend heavily on user participation. How does Magic Square encourage community involvement in project voting and validation? Ishan Pandey: Community-driven platforms depend heavily on user participation. How does Magic Square encourage community involvement in project voting and validation? Andrey Nayman: Community is at the core of every crypto project. At Magic Square, we believe in bringing the crypto community together in one place. Here, users can decide which projects make it onto our platform, shaping their future. Our community moderators actively gather feedback, which our product team uses to create a roadmap aligned with their needs. To make our community feel valued and engaged, we offer daily rewards through Karma Points and exciting opportunities through Hot Offers. We also host regular surveys where users can share their ideas and feedback, ensuring our platform evolves based on their input. Andrey Nayman: Ishan Pandey: Affiliate programs are common in Web3, but Magic Boost offers a unique twist. Can you elaborate on how this program works and what its advantages are for both users and developers? Ishan Pandey: Affiliate programs are common in Web3, but Magic Boost offers a unique twist. Can you elaborate on how this program works and what its advantages are for both users and developers? Ishan Pandey: Affiliate programs are common in Web3, but Magic Boost offers a unique twist. Can you elaborate on how this program works and what its advantages are for both users and developers? Andrey Nayman: Affiliate programs are common in Web3, but Magic Boost offers a unique twist. Here's how it works: Users can easily monetize their communities by sharing affiliate offers from listed projects. With one click, they share these offers on social media. When their friends engage and convert, users receive instant payouts. This benefits the user by allowing them to earn effortlessly while supporting their favorite projects. For developers, Magic Boost expands their reach and boosts user acquisition, enhancing app visibility and credibility. Andrey Nayman: Ishan Pandey: Looking forward, what are some key developments or milestones that the Magic Square community can anticipate in the near future? Ishan Pandey: Looking forward, what are some key developments or milestones that the Magic Square community can anticipate in the near future? Ishan Pandey: Looking forward, what are some key developments or milestones that the Magic Square community can anticipate in the near future? Andrey Nayman: In the near future, our Launchpad will be a big focus. It's designed to help projects grow steadily and connect with our community. We'll customize our approach, using both public and private sales, to give projects the best chance of success. With our active community, we'll make sure these projects get plenty of attention. Stay tuned for more updates as we work together to shape the future of Web3! Andrey Nayman: Don’t forget to like and share the story! Vested Interest Disclosure: This author is an independent contributor publishing via our business blogging program. HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYOR. Vested Interest Disclosure: This author is an independent contributor publishing via our business blogging program . HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYOR. Vested Interest Disclosure: business blogging program business blogging program

Andrey Nayman Reveals the Vision Behind Magic Square's Unique Web3 Ecosystem

By utilizing simple mathematics and parallel computing, we can efficiently overcome performance constraints when manipulating large images. This article will explore how the geometric distortion features of can be used to transform color spaces, as demonstrated in the following example: ImageMetalling To comprehend the code discussed in this article, it would be beneficial to have knowledge of the Color Lookup Table, geometric distortion, and how to work with pixels using the GPU. Main idea Up until now, to alter the color ratio in an image's pixels, we've employed a variety of intuitive and familiar tools such as curves, levels, and , among others. Essentially, these tools manipulate the image's color space. Hue/Saturation What if we could replace all these tools with a single, definitive method? This method is unique because it can replace all other tools simultaneously. A method that's accurate due to its perceptual impact, aligning with the nature of human perception, rather than a literal or simplistic approach, like how Photoshop operates. This is the method we will explore in this article. Let's see how we can manipulate not just the image itself but also its color representation. This will be done in a simple, mechanical way that is accessible to all. Deformation of objects using the MLS method There are various methods of image deformation, such as or . These are indeed fundamental and excellent tools. However, they often suffer from limitations such as the restricted number of control points that can be manipulated, or the excessively smooth extrapolation of the target distortions. affine transformation nonlinear transformation on the B-Spline When dealing with color spaces, we may want to tightly associate a control point and its surroundings with their deformation. Furthermore, the set of points can be arbitrary. This issue has been effectively addressed for animation tasks in computer graphics, but it is still not widely used as a color correction tool in major photo editing software. is a method for reconstructing (approximating) an objective function on a set of known values using the values of the weighted mean square sum. This remarkable technique has many applications in regression analysis, statistics, Big Data, and deep learning. MLS Solving the problem So, imagine we have N control points on a plane. To visualize this, think of a rubber surface scattered with balls that lay flat on a hard table. Each of these balls can be moved and fixed to a target point. Now, envision this surface colored with a slice of color space, for instance, RGB, HSV, or, for the sake of clarity and smoothness, Lab. L=80 Then, with respect to affine transformation, we need to find a solution for the equation: - root mean square weight, - desired transformation function of point , X - initial position of control points, - target position of control points Let's minimize this: in our case, we need to find such a function: in which the sum of all differences between points (where we arrived) and the value of the function from the starting point (where we started) — would be minimal, or in mathematical terms, tend to zero. On the other hand, any affine transformation can be written through the equation: where is the transformation matrix (like ), and is the . M CCM T translation matrix We can assume that for each variable in equation , we can find the minimum value for equation . (1) (2) Because the derivatives with respect to each of the variables in are zero for equation , we can directly solve for through the matrix in equation . The partial derivatives with respect to the free variables in form a linear system of equations. (2) T M (2) T Now, as you can see, we are able to express as a function of : T M Where and — are : weighted centroids This enables us to express equation as a function of : (2) M As you can see, it makes no difference to us which points we look for in the minimum of equation . Whether to search for points (1) or distances between weighted centroids and control points. Therefore, we will not be solving equation for the original control points but instead for new ones: and Considering equation , the solution to the problem is simplified to finding M from: (1) (3) Next, the problem is simplified by differentiating equation and determining as a solution to a system of linear equations using inverse matrices. To find the inverse matrix, we must reduce the system to a solution that includes square matrices on both sides of the equation. And so from we first obtain: (4) M (4), Boom! That’s it, the required function has been found, use to substitute into : (5) (3) The only remaining task is to code the solution , as exemplified in the . Given that there is no necessity for a system of equations solver, and that the matrices involved are merely 2x2, the entire mechanism should, in theory, operate fairly rapidly. (6) ImageMetalling-16 example Expansion of optimization conditions Upon closer examination of , we can observe certain characteristics associated with it, as we have solved the problem by considering it as an affine transformation task, i.e., the task of scaling and shifting. It's clear that in reality, deformations are not confined to just these two parameters. Therefore, can be expressed in a more complex form. For instance, consider transformation which permits uneven translations of the original field. Therefore, let's envision M as: At the same time M M similarity where and are connected through the similarity coefficient M’ M’’ with other vectors and Since all are similar to each other as plane vectors, we can write the similarity equality as an equation for the two-dimensional case: Or rewrite it as: From which it follows that we can now write as: (1) Since we expressed M through and and both of them were found, then we can rewrite by solving , as we did in the previous step: (5) (7) Where In conclusion, the presents a method to link similarity transformation with the covariance matrix and to identify for what's termed the rigid transformation. This approach allows for maximum deformation of spatial properties. At first glance, this method appears quite intuitive, and we anticipate it will be effective in deforming certain color spaces for our intended purposes. original publication M However, this hypothesis still needs to be empirically tested, which you can do yourself. This can be accomplished by downloading and experimenting with the deformations in various spaces, and with different settings for finding . the example M Implementation We will write the main code of the solver in , taking into account the fact that the code will be executed both on the in Metal cores and on the ; for this, it will have to be wrapped in an extension. Obviously, running the solver code on the GPU will be beneficial if the size of the array of control points and the space that we will be modifying are large. C++ GPU CPU Objective-C++ Code of kernels using the solver
kernel void kernel_mlsSolver(
                             constant float2  *input_points  [[buffer(0)]],
                             device float2    *output_points [[buffer(1)]],
                             constant float2  *p      [[buffer(2)]],
                             constant float2  *q      [[buffer(3)]],
                             constant int    &count   [[buffer(4)]],
                             constant MLSSolverKind  &kind [[buffer(5)]],
                             constant float    &alpha [[buffer(6)]],
                             uint gid [[thread_position_in_grid]]
                             ){
    ///
    /// The core of the solver calculates new coordinates for each point 
    ///
    float2 point = input_points[gid];
    IMPMLSSolver solver = IMPMLSSolver(point,p,q,count,kind,alpha);
    output_points[gid] = solver.value(point);
} General MLS solver code

class IMPMLSSolver{
     
public:
         
    /**
     Create a solver for MNC
      
     @param point the current point relative to which we calculate the deformation 
     @param p array of initial control points
     @param q array of control target points
     @param count number of control points
     @param kind solver type: affine transformation, similarity transformation, transformation with maximum possible rigidity
     @param alpha - degree of deformation
     */
    IMPMLSSolver(const float2 point, 
#ifndef __METAL_VERSION__              
              const float2 *p, 
              const float2 *q,
#else
//
// It's important to understand that this is not just a different data type, but a different layout mechanism
// data in memory. In a specific solution for Metal Java, we pass arrays to
// GPU constant memory.                 
//                 
              constant float2 *p, 
              constant float2 *q,
#endif
              const int count, 
              const MLSSolverKind kind = MLSSolverKindAffine, 
              const float alpha = 1.0): 
    point_(point),
    kind_(kind), 
    alpha_(alpha), 
    count_(count),
    weight_(0),
    p_(p), q_(q)
    {
         
#ifndef __METAL_VERSION__
        //
        // For the assembly CPU, allocate temporary buffers for calculating intermediate values
        // solver
        //
        w_ = new float[count_];   
        pHat_ = new float2[count_];   
        qHat_ = new float2[count_];   
#endif
         
        // Finding the weights  
        solveW();
         
        // Finding a vector with a "header"
        solveHat();
         
        // Solving the main optimization equation
        solveM();  
         
#ifndef __METAL_VERSION__
        delete pHat_;
        delete qHat_;
        pHat_=qHat_=0;
#endif
         
    }    
         
    /**
     Returns the deformation point by coordinate
 
     @param point starting position
     @return new position 
     */
    float2 value(float2 point) {
        if (count_ <= 0) return point;   
        return (point - pStar_) * M + qStar_;
    }
     
    ~IMPMLSSolver() {
#ifndef __METAL_VERSION__
        delete w_;
        if (pHat_) delete pHat_;
        if (qHat_) delete qHat_;
#endif
    }
     
private:
     
    MLSSolverKind kind_;
    float   alpha_;
    int     count_;
    float2  point_;
#ifndef __METAL_VERSION__              
    const float2 *p_;
    const float2 *q_;
    float *w_;
    float2 *pHat_;
    float2 *qHat_;    
#else
    constant float2 *p_;
    constant float2 *q_;
    thread float w_[MLS_MAXIMUM_POINTS];
    thread float2 pHat_[MLS_MAXIMUM_POINTS];
    thread float2 qHat_[MLS_MAXIMUM_POINTS];
#endif
     
    float weight_;
    float2 pStar_;
    float2 qStar_;
    float mu_;
     
    float2x2 M;
     
    void solveW() {
         
        weight_ = 0;
        pStar_ = float2(0);
        qStar_ = float2(0);
 
        for (int i=0; i<count_ && i<MLS_MAXIMUM_POINTS; i++) {
             
            float d =  pow(distance(p_[i], point_), 2*alpha_);
             
            if (d < FLT_EPSILON)  d = FLT_EPSILON; 
             
            w_[i] = 1.0 / d;
            weight_ = weight_ + w_[i];
             
            pStar_ += w_[i] * p_[i];
            qStar_ += w_[i] * q_[i];
 
        }
         
        if (weight_ < FLT_EPSILON)  weight_ = FLT_EPSILON;
         
        pStar_ = pStar_ / weight_;                
        qStar_ = qStar_ / weight_;
    }  
     
    void solveHat(){        
         
        mu_ = 0;
         
        float _rmu1 = 0;
        float _rmu2 = 0;
         
        for (int i=0; i<count_ && i<MLS_MAXIMUM_POINTS; i++) {
             
            pHat_[i] = p_[i] - pStar_;                        
            qHat_[i] = q_[i] - qStar_;
             
            switch (kind_) {            
                case MLSSolverKindSimilarity:
                    mu_ += similarityMu(i);
                    break;                
                case MLSSolverKindRigid:
                    _rmu1 += rigidMu1(i); 
                    _rmu2 += rigidMu2(i);
                    break;
                default:
                    break;
            }
        }
         
        switch (kind_) {            
            case MLSSolverKindRigid:
                mu_ = sqrt(_rmu1*_rmu1 + _rmu2*_rmu2);
                break;
            default:
                break;
        }
         
        if (mu_ < FLT_EPSILON)  mu_ = FLT_EPSILON; 
         
        mu_ = 1/mu_;
    }
     
    void solveM() {
        switch (kind_) {
            case MLSSolverKindAffine:
                M = affineM();
                break;
            case MLSSolverKindSimilarity:
            case MLSSolverKindRigid:
                M = similarityM(point_);
                break;
        }
    }  
                    
    float2x2 affineM() {
        float2x2 mi = (float2x2){(float2){0,0},(float2){0,0}};
        float2x2 mj = (float2x2){(float2){0,0},(float2){0,0}};
 
        for (int i=0; i<count_ && i<MLS_MAXIMUM_POINTS; i++) {
             
            float2x2 pti({
                pHat_[i], 
                float2(0)                
            });
             
            float2x2 ppi({
                (float2){w_[i] * pHat_[i].x, 0.0}, 
                (float2){w_[i] * pHat_[i].y, 0.0}                
            });
             
            mi = mi + (float2x2)(pti * ppi);
             
             
            float2x2 ptj({
                w_[i] * pHat_[i], 
                float2(0)                
            });
             
            float2x2 qpj({
                (float2){qHat_[i].x, 0.0}, 
                (float2){qHat_[i].y, 0.0}                
            });
             
            mj = mj + (float2x2)(ptj * qpj);
        }        
        return inverse(mi) * mj;
    }
     
     
    float similarityMu(int i)  {
        return w_[i]*dot(pHat_[i], pHat_[i]);
    }
     
    float2x2 similarityM(float2 value) {
         
        float2x2 m = (float2x2){(float2){0,0},(float2){0,0}};
         
        for (int i=0; i<count_ && i<MLS_MAXIMUM_POINTS; i++) {
             
            float2 sp = -1 * __slashReflect(pHat_[i]);
            float2 sq = -1 * __slashReflect(qHat_[i]);
             
            float2x2 _p({
                (float2){w_[i] * pHat_[i].x, w_[i] * sp.x}, 
                (float2){w_[i] * pHat_[i].y, w_[i] * sp.y}                
            });
             
            float2x2 _q({qHat_[i], sq});
                         
            m = m + (float2x2)(_p * _q);
        }
        return  mu_ * m; 
    }
     
    float rigidMu1(int i) {
        return w_[i]*dot(qHat_[i], pHat_[i]);        
    }
     
    float rigidMu2(int i) {
        return w_[i]*dot(qHat_[i],  __slashReflect(pHat_[i]));        
    }
}; Examples In the application, you can interact with the colors of the picture; the colors will be marked on the Lab color space plane: a-b (in the code, you can replace it with any other). Colors that you want to fix and not change are also worth noting. Then, pull the color dot that you want to change to another. In this example, I highlighted the color of the bottle and chaotically marked a few more. As you can see, by significantly changing one color, the rest are practically unaffected, which can be very convenient for, for example, commercial photographers engaged in digital subject photography or for enthusiasts who dream of matching the color of a digital image to a film one. And this is what the RGB cube looks like after our Lab space settings. For reference, the application also draws a hald-LUT and an HV plane from HSV space. In this example, light color correction has been made: the cold tint of the original slide has been removed. This is how the LUT of the original RGB cube is approximately transformed. It is obvious that the presented framework for working with color spaces can become a universal color correction tool for future generations of digital photographers, when film finally leaves the scene, and color and sound design become a common practice. Also published . here

Understanding Color Space Transform Using The Moving Least Squares Method

The Way We Live Now by Anthony Trollope, is part of the HackerNoon Books Series. You can jump to any chapter in this book . BREAKFAST IN BERKELEY SQUARE. here BREAKFAST IN BERKELEY SQUARE. Lord Nidderdale was greatly disgusted with his own part of the performance when he left the House of Commons, and was, we may say, disgusted with his own position generally, when he considered all its circumstances. That had been at the commencement of the evening, and Melmotte had not then been tipsy; but he had behaved with unsurpassable arrogance and vulgarity, and had made the young lord drink the cup of his own disgrace to the very dregs. Everybody now knew it as a positive fact that the charges made against the man were to become matter of investigation before the chief magistrate for the City, everybody knew that he had committed forgery upon forgery, everybody knew that he could not pay for the property which he had pretended to buy, and that he was actually a ruined man;—and yet he had seized Nidderdale by the hand, and called the young lord "his dear boy" before the whole House. And then he had made himself conspicuous as this man's advocate. If he had not himself spoken openly of his coming marriage with the girl, he had allowed other men to speak to him about it. He had quarrelled with one man for saying that Melmotte was a rogue, and had confidentially told his most intimate friends that in spite of a little vulgarity of manner, Melmotte at bottom was a very good fellow. How was he now to back out of his intimacy with the Melmottes generally? He was engaged to marry the girl, and there was nothing of which he could accuse her. He acknowledged to himself that she deserved well at his hands. Though at this moment he hated the father most bitterly, as those odious words, and the tone in which they had been pronounced, rang in his ears, nevertheless he had some kindly feeling for the girl. Of course he could not marry her now. That was manifestly out of the question. She herself, as well as all others, had known that she was to be married for her money, and now that bubble had been burst. But he felt that he owed it to her, as to a comrade who had on the whole been loyal to him, to have some personal explanation with herself. He arranged in his own mind the sort of speech that he would make to her. "Of course you know it can't be. It was all arranged because you were to have a lot of money, and now it turns out that you haven't got any. And I haven't got any, and we should have nothing to live upon. It's out of the question. But, upon my word, I'm very sorry, for I like you very much, and I really think we should have got on uncommon well together." That was the kind of speech that he suggested to himself, but he did not know how to find for himself the opportunity of making it. He thought that he must put it all into a letter. But then that would be tantamount to a written confession that he had made her an offer of marriage, and he feared that Melmotte,—or Madame Melmotte on his behalf, if the great man himself were absent, in prison,—might make an ungenerous use of such an admission. Between seven and eight he went into the Beargarden, and there he saw Dolly Longestaffe and others. Everybody was talking about Melmotte, the prevailing belief being that he was at this moment in custody. Dolly was full of his own griefs; but consoled amidst them by a sense of his own importance. "I wonder whether it's true," he was saying to Lord Grasslough. "He has an appointment to meet me and my governor at twelve o'clock to-morrow, and to pay us what he owes us. He swore yesterday that he would have the money to-morrow. But he can't keep his appointment, you know, if he's in prison." "You won't see the money, Dolly, you may swear to that," said Grasslough. "I don't suppose I shall. By George, what an ass my governor has been. He had no more right than you have to give up the property. Here's Nidderdale. He could tell us where he is; but I'm afraid to speak to him since he cut up so rough the other night." In a moment the conversation was stopped; but when Lord Grasslough asked Nidderdale in a whisper whether he knew anything about Melmotte, the latter answered out loud, "Yes;—I left him in the House half an hour ago." "People are saying that he has been arrested." "I heard that also; but he certainly had not been arrested when I left the House." Then he went up and put his hand on Dolly Longestaffe's shoulder, and spoke to him. "I suppose you were about right the other night and I was about wrong; but you could understand what it was that I meant. I'm afraid this is a bad look out for both of us." "Yes;—I understand. It's deuced bad for me," said Dolly. "I think you're very well out of it. But I'm glad there's not to be a quarrel. Suppose we have a rubber of whist." Later on in the night news was brought to the club that Melmotte had tried to make a speech in the House, that he had been very drunk, and that he had tumbled over, upsetting Beauchamp Beauclerk in his fall. "By George, I should like to have seen that!" said Dolly. "I am very glad I was not there," said Nidderdale. It was three o'clock before they left the card table, at which time Melmotte was lying dead upon the floor in Mr. Longestaffe's house. On the following morning, at ten o'clock, Lord Nidderdale sat at breakfast with his father in the old lord's house in Berkeley Square. From thence the house which Melmotte had hired was not above a few hundred yards distant. At this time the young lord was living with his father, and the two had now met by appointment in order that something might be settled between them as to the proposed marriage. The Marquis was not a very pleasant companion when the affairs in which he was interested did not go exactly as he would have them. He could be very cross and say most disagreeable words,—so that the ladies of the family, and others connected with him, for the most part, found it impossible to live with him. But his eldest son had endured him;—partly perhaps because, being the eldest, he had been treated with a nearer approach to courtesy, but chiefly by means of his own extreme good humour. What did a few hard words matter? If his father was ungracious to him, of course he knew what all that meant. As long as his father would make fair allowance for his own peccadilloes,—he also would make allowances for his father's roughness. All this was based on his grand theory of live and let live. He expected his father to be a little cross on this occasion, and he acknowledged to himself that there was cause for it. He was a little late himself, and he found his father already buttering his toast. "I don't believe you'd get out of bed a moment sooner than you liked if you could save the whole property by it." "You show me how I can make a guinea by it, sir, and see if I don't earn the money." Then he sat down and poured himself out a cup of tea, and looked at the kidneys and looked at the fish. "I suppose you were drinking last night," said the old lord. "Not particular." The old man turned round and gnashed his teeth at him. "The fact is, sir, I don't drink. Everybody knows that." "I know when you're in the country you can't live without champagne. Well;—what have you got to say about all this?" "What have you got to say?" "You've made a pretty kettle of fish of it." "I've been guided by you in everything. Come, now; you ought to own that. I suppose the whole thing is over?" "I don't see why it should be over. I'm told she has got her own money." Then Nidderdale described to his father Melmotte's behaviour in the House on the preceding evening. "What the devil does that matter?" said the old man. "You're not going to marry the man himself." "I shouldn't wonder if he's in gaol now." "And what does that matter? She's not in gaol. And if the money is hers, she can't lose it because he goes to prison. Beggars mustn't be choosers. How do you mean to live if you don't marry this girl?" "I shall scrape on, I suppose. I must look for somebody else." The Marquis showed very plainly by his demeanour that he did not give his son much credit either for diligence or for ingenuity in making such a search. "At any rate, sir, I can't marry the daughter of a man who is to be put upon his trial for forgery." "I can't see what that has to do with you." "I couldn't do it, sir. I'd do anything else to oblige you, but I couldn't do that. And, moreover, I don't believe in the money." "Then you may just go to the devil," said the old Marquis turning himself round in his chair, and lighting a cigar as he took up the newspaper. Nidderdale went on with his breakfast with perfect equanimity, and when he had finished lighted his cigar. "They tell me," said the old man, "that one of those Goldsheiner girls will have a lot of money." "A Jewess," suggested Nidderdale. "What difference does that make?" "Oh no;—not in the least;—if the money's really there. Have you heard any sum named, sir?" The old man only grunted. "There are two sisters and two brothers. I don't suppose the girls would have a hundred thousand each." "They say the widow of that brewer who died the other day has about twenty thousand a year." "It's only for her life, sir." "She could insure her life. D——me, sir, we must do something. If you turn up your nose at one woman after another how do you mean to live?" "I don't think that a woman of forty with only a life interest would be a good speculation. Of course I'll think of it if you press it." The old man growled again. "You see, sir, I've been so much in earnest about this girl that I haven't thought of inquiring about any one else. There always is some one up with a lot of money. It's a pity there shouldn't be a regular statement published with the amount of money, and what is expected in return. It 'd save a deal of trouble." "If you can't talk more seriously than that you'd better go away," said the old Marquis. At that moment a footman came into the room and told Lord Nidderdale that a man particularly wished to see him in the hall. He was not always anxious to see those who called on him, and he asked the servant whether he knew who the man was. "I believe, my lord, he's one of the domestics from Mr. Melmotte's in Bruton Street," said the footman, who was no doubt fully acquainted with all the circumstances of Lord Nidderdale's engagement. The son, who was still smoking, looked at his father as though in doubt. "You'd better go and see," said the Marquis. But Nidderdale before he went asked a question as to what he had better do if Melmotte had sent for him. "Go and see Melmotte. Why should you be afraid to see him? Tell him that you are ready to marry the girl if you can see the money down, but that you won't stir a step till it has been actually paid over." "He knows that already," said Nidderdale as he left the room. In the hall he found a man whom he recognised as Melmotte's butler, a ponderous, elderly, heavy man who now had a letter in his hand. But the lord could tell by the man's face and manner that he himself had some story to tell. "Is there anything the matter?" "Yes, my lord,—yes. Oh, dear,—oh, dear! I think you'll be sorry to hear it. There was none who came there he seemed to take to so much as your lordship." "They've taken him to prison!" exclaimed Nidderdale. But the man shook his head. "What is it then? He can't be dead." Then the man nodded his head, and, putting his hand up to his face, burst into tears. "Mr. Melmotte dead! He was in the House of Commons last night. I saw him myself. How did he die?" But the fat, ponderous man was so affected by the tragedy he had witnessed, that he could not as yet give any account of the scene of his master's death, but simply handed the note which he had in his hand to Lord Nidderdale. It was from Marie, and had been written within half an hour of the time at which news had been brought to her of what had occurred. The note was as follows:— Dear Lord Nidderdale, The man will tell you what has happened. I feel as though I was mad. I do not know who to send to. Will you come to me, only for a few minutes? Marie. He read it standing up in the hall, and then again asked the man as to the manner of his master's death. And now the Marquis, gathering from a word or two that he heard and from his son's delay that something special had occurred, hobbled out into the hall. "Mr. Melmotte is—dead," said his son. The old man dropped his stick, and fell back against the wall. "This man says that he is dead, and here is a letter from Marie asking me to go there. How was it that he—died?" "It was—poison," said the butler solemnly. "There has been a doctor already, and there isn't no doubt of that. He took it all by himself last night. He came home, perhaps a little fresh, and he had in brandy and soda and cigars;—and sat himself down all to himself. Then in the morning, when the young woman went in,—there he was,—poisoned! I see him lay on the ground, and I helped to lift him up, and there was that smell of prussic acid that I knew what he had been and done just the same as when the doctor came and told us." Before the man could be allowed to go back, there was a consultation between the father and son as to a compliance with the request which Marie had made in her first misery. The Marquis thought that his son had better not go to Bruton Street. "What's the use? What good can you do? She'll only be falling into your arms, and that's what you've got to avoid,—at any rate, till you know how things are." But Nidderdale's better feelings would not allow him to submit to this advice. He had been engaged to marry the girl, and she in her abject misery had turned to him as the friend she knew best. At any rate for the time the heartlessness of his usual life deserted him, and he felt willing to devote himself to the girl not for what he could get,—but because she had so nearly been so near to him. "I couldn't refuse her," he said over and over again. "I couldn't bring myself to do it. Oh, no;—I shall certainly go." "You'll get into a mess if you do." "Then I must get into a mess. I shall certainly go. I will go at once. It is very disagreeable, but I cannot possibly refuse. It would be abominable." Then going back to the hall, he sent a message by the butler to Marie, saying that he would be with her in less than half an hour. "Don't you go and make a fool of yourself," his father said to him when he was alone. "This is just one of those times when a man may ruin himself by being soft-hearted." Nidderdale simply shook his head as he took his hat and gloves to go across to Bruton Street. About HackerNoon Book Series: We bring you the most important technical, scientific, and insightful public domain books. This book is part of the public domain. Anthony Trollope (2004). The Way We Live Now. Urbana, Illinois: Project Gutenberg. Retrieved October 2022 https://www.gutenberg.org/cache/epub/5231/pg5231-images.html This eBook is for the use of anyone anywhere at no cost and with almost no restrictions whatsoever. You may copy it, give it away or re-use it under the terms of the Project Gutenberg License included with this eBook or online at , located at . www.gutenberg.org https://www.gutenberg.org/policy/license.html

BREAKFAST IN BERKELEY SQUARE.

Phineas Redux by Anthony Trollope, is part of the HackerNoon Books Series. You can jump to any chapter in this book . ONCE AGAIN IN PORTMAN SQUARE. here ONCE AGAIN IN PORTMAN SQUARE. On the Wednesday in Easter week Lord Brentford and Lady Laura Kennedy reached Portman Square from Dresden, and Phineas, who had remained in town, was summoned thither by a note written at Dover. "We arrived here to-day, and shall be in town to-morrow afternoon, between four and five. Papa wants to see you especially. Can you manage to be with us in the Square at about eight? I know it will be inconvenient, but you will put up with inconvenience. I don't like to keep Papa up late; and if he is tired he won't speak to you as he would if you came early.—L. K." Phineas was engaged to dine with Lord Cantrip; but he wrote to excuse himself,—telling the simple truth. He had been asked to see Lord Brentford on business, and must obey the summons. He was shown into a sitting-room on the ground floor, which he had always known as the Earl's own room, and there he found Lord Brentford alone. The last time he had been there he had come to plead with the Earl on behalf of Lord Chiltern, and the Earl had then been a stern self-willed man, vigorous from a sense of power, and very able to maintain and to express his own feelings. Now he was a broken-down old man,—whose mind had been, as it were, unbooted and put into moral slippers for the remainder of its term of existence upon earth. He half shuffled up out of his chair as Phineas came up to him, and spoke as though every calamity in the world were oppressing him. "Such a passage! Oh, very bad, indeed! I thought it would have been the death of me. Laura thought it better to come on." The fact, however, had been that the Earl had so many objections to staying at Calais, that his daughter had felt herself obliged to yield to him. "You must be glad at any rate to have got home," said Phineas. "Home! I don't know what you call home. I don't suppose I shall ever feel any place to be home again." "You'll go to Saulsby;—will you not?" "How can I tell? If Chiltern would have kept the house up, of course I should have gone there. But he never would do anything like anybody else. Violet wants me to go to that place they've got there, but I shan't do that." "It's a comfortable house." "I hate horses and dogs, and I won't go." There was nothing more to be said on that point. "I hope Lady Laura is well." "No, she's not. How should she be well? She's anything but well. She'll be in directly, but she thought I ought to see you first. I suppose this wretched man is really mad." "I am told so." "He never was anything else since I knew him. What are we to do now? Forster says it won't look well to ask for a separation only because he's insane. He tried to shoot you?" "And very nearly succeeded." "Forster says that if we do anything, all that must come out." "There need not be the slightest hesitation as far as I am concerned, Lord Brentford." "You know he keeps all her money." "At present I suppose he couldn't give it up." "Why not? Why shouldn't he give it up? God bless my soul! Forty thousand pounds and all for nothing. When he married he declared that he didn't care about it! Money was nothing to him! So she lent it to Chiltern." "I remember." "But they hadn't been together a year before he asked for it. Now there it is;—and if she were to die to-morrow it would be lost to the family. Something must be done, you know. I can't let her money go in that way." "You'll do what Mr. Forster suggests, no doubt." "But he won't suggest anything. They never do. He doesn't care what becomes of the money. It never ought to have been given up as it was." "It was settled, I suppose." "Yes;—if there were children. And it will come back to her if he dies first. But mad people never do die. That's a well-known fact. They've nothing to trouble them, and they live for ever. It'll all go to some cousin of his that nobody ever saw." "Not as long as Lady Laura lives." "But she does not get a penny of the income;—not a penny. There never was anything so cruel. He has published all manner of accusations against her." "Nobody believes a word of that, my lord." "And then when she is dragged forward by the necessity of vindicating her character, he goes mad and keeps all her money! There never was anything so cruel since the world began." This continued for half-an-hour, and then Lady Laura came in. Nothing had come, or could have come, from the consultation with the Earl. Had it gone on for another hour, he would simply have continued to grumble, and have persevered in insisting upon the hardships he endured. Lady Laura was in black, and looked sad, and old, and careworn; but she did not seem to be ill. Phineas could not but think at the moment how entirely her youth had passed away from her. She came and sat close by him, and began at once to speak of the late debate. "Of course they'll go out," she said. "I presume they will." "And our party will come in." "Oh, yes;—Mr. Gresham, and the two dukes, and Lord Cantrip,—with Legge Wilson, Sir Harry Coldfoot, and the rest of them." "And you?" Phineas smiled, and tried to smile pleasantly, as he answered, "I don't know that they'll put themselves out by doing very much for me." "They'll do something." "I fancy not. Indeed, Lady Laura, to tell the truth at once, I know that they don't mean to offer me anything." "After making you give up your place in Ireland?" "They didn't make me give it up. I should never dream of using such an argument to any one. Of course I had to judge for myself. There is nothing to be said about it;—only it is so." As he told her this he strove to look light-hearted, and so to speak that she should not see the depth of his disappointment;—but he failed altogether. She knew him too well not to read his whole heart in the matter. "Who has said it?" she asked. "Nobody says things of that kind, and yet one knows." "And why is it?" "How can I say? There are various reasons,—and, perhaps, very good reasons. What I did before makes men think that they can't depend on me. At any rate it is so." "Shall you not speak to Mr. Gresham?" "Certainly not." "What do you say, Papa?" "How can I understand it, my dear? There used to be a kind of honour in these things, but that's all old-fashioned now. Ministers used to think of their political friends; but in these days they only regard their political enemies. If you can make a Minister afraid of you, then it becomes worth his while to buy you up. Most of the young men rise now by making themselves thoroughly disagreeable. Abuse a Minister every night for half a session, and you may be sure to be in office the other half,—if you care about it." "May I speak to Barrington Erle?" asked Lady Laura. "I had rather you did not. Of course I must take it as it comes." "But, my dear Mr. Finn, people do make efforts in such cases. I don't doubt but that at this moment there are a dozen men moving heaven and earth to secure something. No one has more friends than you have." Had not her father been present he would have told her what his friends were doing for him, and how unhappy such interferences made him; but he could not explain all this before the Earl. "I would so much rather hear about yourself," he said, again smiling. "There is but little to say about us. I suppose Papa has told you?" But the Earl had told him nothing, and indeed, there was nothing to tell. The lawyer had advised that Mr. Kennedy's friends should be informed that Lady Laura now intended to live in England, and that they should be invited to make to her some statement as to Mr. Kennedy's condition. If necessary he, on her behalf, would justify her departure from her husband's roof by a reference to the outrageous conduct of which Mr. Kennedy had since been guilty. In regard to Lady Laura's fortune, Mr. Forster said that she could no doubt apply for alimony, and that if the application were pressed at law she would probably obtain it;—but he could not recommend such a step at the present moment. As to the accusation which had been made against her character, and which had become public through the malice of the editor of The People's Banner, Mr. Forster thought that the best refutation would be found in her return to England. At any rate he would advise no further step at the present moment. Should any further libel appear in the columns of the newspaper, then the question might be again considered. Mr. Forster had already been in Portman Square, and this had been the result of the conference. "There is not much comfort in it all,—is there?" said Lady Laura. "There is no comfort in anything," said the Earl. When Phineas took his leave Lady Laura followed him out into the hall, and they went together into the large, gloomy dining-room,—gloomy and silent now, but which in former days he had known to be brilliant with many lights, and cheerful with eager voices. "I must have one word with you," she said, standing close to him against the table, and putting her hand upon his arm. "Amidst all my sorrow, I have been so thankful that he did not—kill you." "I almost wish he had." "Oh, Phineas!—how can you say words so wicked! Would you have had him a murderer?" "A madman is responsible for nothing." "Where should I have been? What should I have done? But of course you do not mean it. You have everything in life before you. Say some word to me more comfortable than that. You cannot think how I have looked forward to meeting you again. It has robbed the last month of half its sadness." He put his arm round her waist and pressed her to his side, but he said nothing. "It was so good of you to go to him as you did. How was he looking?" "Twenty years older than when you saw him last." "But how in health?" "He was thin and haggard." "Was he pale?" "No; flushed and red. He had not shaved himself for days; nor, as I believe, had he been out of his room since he came up to London. I fancy that he will not live long." "Poor fellow;—unhappy man! I was very wrong to marry him, Phineas." "I have never said so;—nor, indeed, thought so." "But I have thought so; and I say it also,—to you. I owe him any reparation that I can make him; but I could not have lived with him. I had no idea, before, that the nature of two human beings could be so unlike. I so often remember what you told me of him,—here; in this house, when I first brought you together. Alas, how sad it has been!" "Sad, indeed." "But can this be true that you tell me of yourself? "It is quite true. I could not say so before your father, but it is Mr. Bonteen's doing. There is no remedy. I am sure of that. I am only afraid that people are interfering for me in a manner that will be as disagreeable to me as it will be useless." "What friends?" she asked. He was still standing with his arm round her waist, and he did not like to mention the name of Madame Goesler. "The Duchess of Omnium,—whom you remember as Lady Glencora Palliser." "Is she a friend of yours?" "No;—not particularly. But she is an indiscreet woman, and hates Bonteen, and has taken it into her stupid head to interest herself in my concerns. It is no doing of mine, and yet I cannot help it." "She will succeed." "I don't want assistance from such a quarter; and I feel sure that she will not succeed." "What will you do, Phineas?" "What shall I do? Carry on the battle as long as I can without getting into debt, and then—vanish." "You vanished once before,—did you not,—with a wife?" "And now I shall vanish alone. My poor little wife! It seems all like a dream. She was so good, so pure, so pretty, so loving!" "Loving! A man's love is so easily transferred;—as easily as a woman's hand;—is it not, Phineas? Say the word, for it is what you are thinking." "I was thinking of no such thing." "You must think it—You need not be afraid to reproach me. I could bear it from you. What could I not bear from you? Oh, Phineas;—if I had only known myself then, as I do now!" "It is too late for regrets," he said. There was something in the words which grated on her feelings, and induced her at length to withdraw herself from his arm. Too late for regrets! She had never told herself that it was not too late. She was the wife of another man, and therefore, surely it was too late. But still the word coming from his mouth was painful to her. It seemed to signify that for him at least the game was all over. "Yes, indeed," she said,—"if our regrets and remorse were at our own disposal! You might as well say that it is too late for unhappiness, too late for weariness, too late for all the misery that comes from a life's disappointment." "I should have said that indulgence in regrets is vain." "That is a scrap of philosophy which I have heard so often before! But we will not quarrel, will we, on the first day of my return?" "I hope not." "And I may speak to Barrington?" "No; certainly not." "But I shall. How can I help it? He will be here to-morrow, and will be full of the coming changes. How should I not mention your name? He knows—not all that has passed, but too much not to be aware of my anxiety. Of course your name will come up?" "What I request,—what I demand is, that you ask no favour for me. Your father will miss you,—will he not? I had better go now." "Good night, Phineas." "Good night, dear friend." "Dearest, dearest friend," she said. Then he left her, and without assistance, let himself out into the square. In her intercourse with him there was a passion the expression of which caused him sorrow and almost dismay. He did not say so even to himself, but he felt that a time might come in which she would resent the coldness of demeanour which it would be imperative upon him to adopt in his intercourse with her. He knew how imprudent he had been to stand there with his arm round her waist. About HackerNoon Book Series: We bring you the most important technical, scientific, and insightful public domain books. This book is part of the public domain. Anthony Trollope (2006). Phineas Redux. Urbana, Illinois: Project Gutenberg. Retrieved https://www.gutenberg.org/cache/epub/18640/pg18640-images.html This eBook is for the use of anyone anywhere at no cost and with almost no restrictions whatsoever. You may copy it, give it away or re-use it under the terms of the Project Gutenberg License included with this eBook or online at , located at . www.gutenberg.org https://www.gutenberg.org/policy/license.html

ONCE AGAIN IN PORTMAN SQUARE.

The Way We Live Now by Anthony Trollope is part of the HackerNoon Books Series. You can jump to any chapter in this book . IN GROSVENOR SQUARE here IN GROSVENOR SQUARE. Marie Melmotte was hardly satisfied with the note which she received from Didon early on the Monday morning. With a volubility of French eloquence, Didon declared that she would be turned out of the house if either Monsieur or Madame were to know what she was doing. Marie told her that Madame would certainly never dismiss her. "Well, perhaps not Madame," said Didon, who knew too much about Madame to be dismissed; "but Monsieur!" Marie declared that by no possibility could Monsieur know anything about it. In that house nobody ever told anything to Monsieur. He was regarded as the general enemy, against whom the whole household was always making ambushes, always firing guns from behind rocks and trees. It is not a pleasant condition for a master of a house; but in this house the master at any rate knew how he was placed. It never occurred to him to trust any one. Of course his daughter might run away. But who would run away with her without money? And there could be no money except from him. He knew himself and his own strength. He was not the man to forgive a girl, and then bestow his wealth on the Lothario who had injured him. His daughter was valuable to him because she might make him the father-in-law of a Marquis or an Earl; but the higher that he rose without such assistance, the less need had he of his daughter's aid. Lord Alfred was certainly very useful to him. Lord Alfred had whispered into his ear that by certain conduct and by certain uses of his money, he himself might be made a baronet. "But if they should say that I'm not an Englishman?" suggested Melmotte. Lord Alfred had explained that it was not necessary that he should have been born in England, or even that he should have an English name. No questions would be asked. Let him first get into Parliament, and then spend a little money on the proper side,—by which Lord Alfred meant the Conservative side,—and be munificent in his entertainments, and the baronetcy would be almost a matter of course. Indeed, there was no knowing what honours might not be achieved in the present days by money scattered with a liberal hand. In these conversations, Melmotte would speak of his money and power of making money as though they were unlimited,—and Lord Alfred believed him. Marie was dissatisfied with her letter,—not because it described her father as "cutting up very rough." To her who had known her father all her life that was a matter of course. But there was no word of love in the note. An impassioned correspondence carried on through Didon would be delightful to her. She was quite capable of loving, and she did love the young man. She had, no doubt, consented to accept the addresses of others whom she did not love,—but this she had done at the moment almost of her first introduction to the marvellous world in which she was now living. As days went on she ceased to be a child, and her courage grew within her. She became conscious of an identity of her own, which feeling was produced in great part by the contempt which accompanied her increasing familiarity with grand people and grand names and grand things. She was no longer afraid of saying No to the Nidderdales on account of any awe of them personally. It might be that she should acknowledge herself to be obliged to obey her father, though she was drifting away even from the sense of that obligation. Had her mind been as it was now when Lord Nidderdale first came to her, she might indeed have loved him, who, as a man, was infinitely better than Sir Felix, and who, had he thought it to be necessary, would have put some grace into his love-making. But at that time she had been childish. He, finding her to be a child, had hardly spoken to her. And she, child though she was, had resented such usage. But a few months in London had changed all this, and now she was a child no longer. She was in love with Sir Felix, and had told her love. Whatever difficulties there might be, she intended to be true. If necessary, she would run away. Sir Felix was her idol, and she abandoned herself to its worship. But she desired that her idol should be of flesh and blood, and not of wood. She was at first half-inclined to be angry; but as she sat with his letter in her hand, she remembered that he did not know Didon as well as she did, and that he might be afraid to trust his raptures to such custody. She could write to him at his club, and having no such fear, she could write warmly. —, Grosvenor Square. Early Monday Morning. Dearest, Dearest Felix, I have just got your note;—such a scrap! Of course papa would talk about money because he never thinks of anything else. I don't know anything about money, and I don't care in the least how much you have got. Papa has got plenty, and I think he would give us some if we were once married. I have told mamma, but mamma is always afraid of everything. Papa is very cross to her sometimes;—more so than to me. I will try to tell him, though I can't always get at him. I very often hardly see him all day long. But I don't mean to be afraid of him, and will tell him that on my word and honour I will never marry any one except you. I don't think he will beat me, but if he does, I'll bear it,—for your sake. He does beat mamma sometimes, I know. You can write to me quite safely through Didon. I think if you would call some day and give her something, it would help, as she is very fond of money. Do write and tell me that you love me. I love you better than anything in the world, and I will never,—never give you up. I suppose you can come and call,—unless papa tells the man in the hall not to let you in. I'll find that out from Didon, but I can't do it before sending this letter. Papa dined out yesterday somewhere with that Lord Alfred, so I haven't seen him since you were here. I never see him before he goes into the city in the morning. Now I am going down-stairs to breakfast with mamma and that Miss Longestaffe. She is a stuck-up thing. Didn't you think so at Caversham? Good-bye. You are my own, own, own darling Felix, And I am your own, own affectionate ladylove, Marie. Sir Felix when he read this letter at his club in the afternoon of the Monday, turned up his nose and shook his head. He thought if there were much of that kind of thing to be done, he could not go on with it, even though the marriage were certain, and the money secure. "What an infernal little ass!" he said to himself as he crumpled the letter up. Marie having intrusted her letter to Didon, together with a little present of gloves and shoes, went down to breakfast. Her mother was the first there, and Miss Longestaffe soon followed. That lady, when she found that she was not expected to breakfast with the master of the house, abandoned the idea of having her meal sent to her in her own room. Madame Melmotte she must endure. With Madame Melmotte she had to go out in the carriage every day. Indeed she could only go to those parties to which Madame Melmotte accompanied her. If the London season was to be of any use at all, she must accustom herself to the companionship of Madame Melmotte. The man kept himself very much apart from her. She met him only at dinner, and that not often. Madame Melmotte was very bad; but she was silent, and seemed to understand that her guest was only her guest as a matter of business. But Miss Longestaffe already perceived that her old acquaintances were changed in their manner to her. She had written to her dear friend Lady Monogram, whom she had known intimately as Miss Triplex, and whose marriage with Sir Damask Monogram had been splendid preferment, telling how she had been kept down in Suffolk at the time of her friend's last party, and how she had been driven to consent to return to London as the guest of Madame Melmotte. She hoped her friend would not throw her off on that account. She had been very affectionate, with a poor attempt at fun, and rather humble. Georgiana Longestaffe had never been humble before; but the Monograms were people so much thought of and in such an excellent set! She would do anything rather than lose the Monograms. But it was of no use. She had been humble in vain, for Lady Monogram had not even answered her note. "She never really cared for anybody but herself," Georgiana said in her wretched solitude. Then, too, she had found that Lord Nidderdale's manner to her had been quite changed. She was not a fool, and could read these signs with sufficient accuracy. There had been little flirtations between her and Nidderdale,—meaning nothing, as every one knew that Nidderdale must marry money; but in none of them had he spoken to her as he spoke when he met her in Madame Melmotte's drawing-room. She could see it in the faces of people as they greeted her in the park,—especially in the faces of the men. She had always carried herself with a certain high demeanour, and had been able to maintain it. All that was now gone from her, and she knew it. Though the thing was as yet but a few days old she understood that others understood that she had degraded herself. "What's all this about?" Lord Grasslough had said to her, seeing her come into a room behind Madame Melmotte. She had simpered, had tried to laugh, and had then turned away her face. "Impudent scoundrel!" she said to herself, knowing that a fortnight ago he would not have dared to address her in such a tone. A day or two afterwards an occurrence took place worthy of commemoration. Dolly Longestaffe called on his sister! His mind must have been much stirred when he allowed himself to be moved to such uncommon action. He came too at a very early hour, not much after noon, when it was his custom to be eating his breakfast in bed. He declared at once to the servant that he did not wish to see Madame Melmotte or any of the family. He had called to see his sister. He was therefore shown into a separate room where Georgiana joined him. "What's all this about?" She tried to laugh as she tossed her head. "What brings you here, I wonder? This is quite an unexpected compliment." "My being here doesn't matter. I can go anywhere without doing much harm. Why are you staying with these people?" "Ask papa." "I don't suppose he sent you here?" "That's just what he did do." "You needn't have come, I suppose, unless you liked it. Is it because they are none of them coming up?" "Exactly that, Dolly. What a wonderful young man you are for guessing!" "Don't you feel ashamed of yourself?" "No;—not a bit." "Then I feel ashamed for you." "Everybody comes here." "No;—everybody does not come and stay here as you are doing. Everybody doesn't make themselves a part of the family. I have heard of nobody doing it except you. I thought you used to think so much of yourself." "I think as much of myself as ever I did," said Georgiana, hardly able to restrain her tears. "I can tell you nobody else will think much of you if you remain here. I could hardly believe it when Nidderdale told me." "What did he say, Dolly?" "He didn't say much to me, but I could see what he thought. And of course everybody thinks the same. How you can like the people yourself is what I can't understand!" "I don't like them,—I hate them." "Then why do you come and live with them?" "Oh, Dolly, it is impossible to make you understand. A man is so different. You can go just where you please, and do what you like. And if you're short of money, people will give you credit. And you can live by yourself, and all that sort of thing. How should you like to be shut up down at Caversham all the season?" "I shouldn't mind it,—only for the governor." "You have got a property of your own. Your fortune is made for you. What is to become of me?" "You mean about marrying?" "I mean altogether," said the poor girl, unable to be quite as explicit with her brother, as she had been with her father, and mother, and sister. "Of course I have to think of myself." "I don't see how the Melmottes are to help you. The long and the short of it is, you oughtn't to be here. It's not often I interfere, but when I heard it I thought I'd come and tell you. I shall write to the governor, and tell him too. He should have known better." "Don't write to papa, Dolly!" "Yes, I shall. I am not going to see everything going to the devil without saying a word. Good-bye." As soon as he had left he hurried down to some club that was open,—not the Beargarden, as it was long before the Beargarden hours,—and actually did write a letter to his father. My Dear Father, I have seen Georgiana at Mr. Melmotte's house. She ought not to be there. I suppose you don't know it, but everybody says he's a swindler. For the sake of the family I hope you will get her home again. It seems to me that Bruton Street is the proper place for the girls at this time of the year. Your affectionate son, Adolphus Longestaffe. This letter fell upon old Mr. Longestaffe at Caversham like a thunderbolt. It was marvellous to him that his son should have been instigated to write a letter. The Melmottes must be very bad indeed,—worse than he had thought,—or their iniquities would not have brought about such energy as this. But the passage which angered him most was that which told him that he ought to have taken his family back to town. This had come from his son, who had refused to do anything to help him in his difficulties. About HackerNoon Book Series: We bring you the most important technical, scientific, and insightful public domain books. This book is part of the public domain. Anthony Trollope (2004). The Way We Live Now. Urbana, Illinois: Project Gutenberg. Retrieved October 2022 https://www.gutenberg.org/cache/epub/5231/pg5231-images.html This eBook is for the use of anyone anywhere at no cost and with almost no restrictions whatsoever. You may copy it, give it away or re-use it under the terms of the Project Gutenberg License included with this eBook or online at , located at . www.gutenberg.org https://www.gutenberg.org/policy/license.html

IN GROSVENOR SQUARE

Regex 101: Practical Tips for Mastering Regular Expressions

Authors:
(1) Hoon Kim, Beeble AI, and contributed equally to this work;
(2) Minje Jang, Beeble AI, and contributed equally to this work;
(3) Wonjun Yoon, Beeble AI, and contributed equally to this work;
(4) Jisoo Lee, Beeble AI, and contributed equally to this work;
(5) Donghyun Na, Beeble AI, and contributed equally to this work;
(6) Sanghyun Woo, New York University, and contributed equally to this work. Editor's Note: This is Part 9 of 14 of a study introducing a method for improving how light and shadows can be applied to human portraits in digital images. Read the rest below. Table of Links Abstract and 1. Introduction
2. Related Work
3. SwitchLight and 3.1. Preliminaries
3.2. Problem Formulation
3.3. Architecture
3.4. Objectives
4. Multi-Masked Autoencoder Pre-training
5. Data
6. Experiments
7. Conclusion Appendix A. Implementation Details
B. User Study Interface
C. Video Demonstration
D. Additional Qualitative Results & References 6. Experiments This section details our experimental results. We begin with a comparative evaluation of our method against state-ofthe-art approaches using the OLAT dataset. We also employ images from the FFHQ–test [25] for user studies. For qualitative analysis, we utilize copyright-free portrait images from Pexels [1]. Additionally, we conduct ablation studies to validate the efficacy of our pre-training framework and architectural design choice. Subsequently, we detail the additional features and conclude by discussing its limitations. Our evaluation uses the OLAT test set, comprising 35 subjects and 11 lighting environments, ensuring no overlap with the train set. Evaluation metrics. We employ several key metrics for evaluating the prediction accuracy; Mean Absolute Error (MAE), Mean Squared Error (MSE), Structural Similarity Index Measure (SSIM) and Learned Perceptual Image Patch Similarity (LPIPS). While these metrics offer valuable quantitative insights, they do not fully capture the subtleties of visual quality enhancement. Therefore, we emphasize the importance of qualitative evaluations to gain a comprehensive understanding of model performance. Baselines. We compared our approach with three state-of-the-art baselines: Single Image Portrait Relighting (SIPR) [45], which uses a single neural network for relighting; Total Relight (TR) [34], employing multiple neural networks that incorporate the Phong reflectance model; and Lumos [52], a TR adaptation for synthetic datasets. Due to the lack of publicly available code or model from these methods, we either integrated their techniques into our framework or requested the original authors to process our inputs with their models and share the results. Quantitative Comparisons. The results in Table. 1 shows our method outperforming SIPR and TR baselines, demonstrating the significance of incorporating advanced rendering physics and reflectance models. The transition from SIPR to TR emphasizes the value of physics-based design, while the shift from TR to our approach underscores the importance of transitioning from the empirical Phong model to the more accurate Cook-Torrance model. Additionally, pretraining contributes to further enhancements, as evidenced by the improved image details, depicted in Fig 6. Qualitative Comparisons. Our relighting method exhibits several key advantages over previous approaches, as showcased in Fig. 7. It effectively harmonizes light direction and softness, avoiding harsh highlights and inaccurate lighting that are commonly observed in other methods. A notable strength of our approach lies in its ability to capture highfrequency details like specular highlights and hard shadows. Additionally, as shown in the second row, it preserves facial details and identity, ensuring high fidelity to the subject’s original features and mitigating common distortions seen in previous approaches. Moreover, our approach excels in handling skin tones, producing natural and accurate results under various lighting conditions. This is clearly demonstrated in the fourth row, where our method contrasts sharply with the over-saturated or pale tones from previous methods. Finally, the nuanced treatment of hair is highlighted in the sixth row, where our approach maintains luster and detail, unlike the flattened effect typical in other methods. More qualitative results are available in our supplementary video demonstration. User Study. We conducted a human subjective test to evaluate the visual quality of relighitng results, summarized in Table. 2. In each test case, workers were presented with an input image and an environment map. They were asked to compare the relighting results from three methods–Ours, Lumos, and TR–based on three criteria: 1) consistency of lighting with the environment map, 2) preservation of facial details, and 3) maintenance of the original identity. To ensure unbiased evaluations, the order of the methods presented was randomized. To aid in understanding the concept of consistent lighting, relit balls were displayed alongside the images. The study included a total of 256 images, consisting of 32 portraits each illuminated with 8 different HDRIs. Each worker was tasked with selecting the best image for each specific criterion, randomly assessing 30 samples. A total of 47 workers participated in the study. The results indicate a strong preference for our results over the baseline methods across all evaluated metrics. Ablation Studies. We analyze our two major design choices in Table. 3: the MMAE pre-training framework and DiffuseNet. The MMAE, which integrates dynamic masking with generative objectives, outperforms MAE. This superiority is mainly due to the incorporation of challenging masks and global coherence objectives, enabling the model to learn richer features during pre-training. Furthermore, our method of predicting diffuse render demonstrates superiority over direct albedo prediction. Firstly, we see it simplifies the learning process, as predicting diffuse render is more closely related to the original image. Secondly, our approach effectively distinguishes between the influences of illumination (diffuse shading) and surface properties (diffuse render). This distinction is crucial for accurately modeling the intrinsic color of surfaces, as it enables independent and precise evaluation of each element (see Eqn. 9). In contrast, methods that predict albedo directly often struggle to differentiate between these factors, leading to significant inaccuracies in color constancy, as shown in Fig. 8. Applications. We present two applications using predicted intrinsics in Fig. 9. First, real-time PBR via Cook-Torrance components in Three.js graphics library. Second, switching the lighting environment between the source and reference images. Further details are in the supplementary video. Limitations. We identified a few failure cases in Fig. 10. First, our model struggles with neutralizing strong shadows, which leads to inaccurate facial geometry and residual shadow artifacts in both albedo and relit images. Incorporating shadow augmentation techniques [16, 54] during training could mitigate this issue. Second, the model incorrectly interprets reflective surfaces, such as sunglasses, as opaque objects in the normal image. This error prevents the model from properly removing reflective highlights in the albedo and relit images. Lastly, the model inaccurately predicts the albedo for face paint. Implementing a semantic mask [52] to distinguish different semantic regions separately from the skin could help resolving these issues. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. Authors: (1) Hoon Kim, Beeble AI, and contributed equally to this work; (2) Minje Jang, Beeble AI, and contributed equally to this work; (3) Wonjun Yoon, Beeble AI, and contributed equally to this work; (4) Jisoo Lee, Beeble AI, and contributed equally to this work; (5) Donghyun Na, Beeble AI, and contributed equally to this work; (6) Sanghyun Woo, New York University, and contributed equally to this work. Authors: Authors: (1) Hoon Kim, Beeble AI, and contributed equally to this work; (2) Minje Jang, Beeble AI, and contributed equally to this work; (3) Wonjun Yoon, Beeble AI, and contributed equally to this work; (4) Jisoo Lee, Beeble AI, and contributed equally to this work; (5) Donghyun Na, Beeble AI, and contributed equally to this work; (6) Sanghyun Woo, New York University, and contributed equally to this work. Editor's Note: This is Part 9 of 14 of a study introducing a method for improving how light and shadows can be applied to human portraits in digital images. Read the rest below. Editor's Note: This is Part 9 of 14 of a study introducing a method for improving how light and shadows can be applied to human portraits in digital images. Read the rest below. Editor's Note: This is Part 9 of 14 of a study introducing a method for improving how light and shadows can be applied to human portraits in digital images. Read the rest below. Editor's Note: This is Part 9 of 14 of a study introducing a method for improving how light and shadows can be applied to human portraits in digital images. Read the rest below. Table of Links Abstract and 1. Introduction 2. Related Work 3. SwitchLight and 3.1. Preliminaries 3.2. Problem Formulation 3.3. Architecture 3.4. Objectives 4. Multi-Masked Autoencoder Pre-training 5. Data 6. Experiments 7. Conclusion Abstract and 1. Introduction Abstract and 1. Introduction 2. Related Work 2. Related Work 3. SwitchLight and 3.1. Preliminaries 3. SwitchLight and 3.1. Preliminaries 3.2. Problem Formulation 3.2. Problem Formulation 3.3. Architecture 3.3. Architecture 3.4. Objectives 3.4. Objectives 4. Multi-Masked Autoencoder Pre-training 4. Multi-Masked Autoencoder Pre-training 5. Data 5. Data 6. Experiments 6. Experiments 7. Conclusion 7. Conclusion Appendix A. Implementation Details B. User Study Interface C. Video Demonstration D. Additional Qualitative Results & References A. Implementation Details A. Implementation Details B. User Study Interface B. User Study Interface C. Video Demonstration C. Video Demonstration D. Additional Qualitative Results & References D. Additional Qualitative Results & References 6. Experiments This section details our experimental results. We begin with a comparative evaluation of our method against state-ofthe-art approaches using the OLAT dataset. We also employ images from the FFHQ–test [25] for user studies. For qualitative analysis, we utilize copyright-free portrait images from Pexels [1]. Additionally, we conduct ablation studies to validate the efficacy of our pre-training framework and architectural design choice. Subsequently, we detail the additional features and conclude by discussing its limitations. Our evaluation uses the OLAT test set, comprising 35 subjects and 11 lighting environments, ensuring no overlap with the train set. Evaluation metrics. We employ several key metrics for evaluating the prediction accuracy; Mean Absolute Error ( MAE ), Mean Squared Error ( MSE ), Structural Similarity Index Measure ( SSIM ) and Learned Perceptual Image Patch Similarity ( LPIPS ). While these metrics offer valuable quantitative insights, they do not fully capture the subtleties of visual quality enhancement. Therefore, we emphasize the importance of qualitative evaluations to gain a comprehensive understanding of model performance. Evaluation metrics. MAE MSE SSIM LPIPS Baselines . We compared our approach with three state-of-the-art baselines: Single Image Portrait Relighting ( SIPR ) [45], which uses a single neural network for relighting; Total Relight ( TR ) [34], employing multiple neural networks that incorporate the Phong reflectance model; and Lumos [52], a TR adaptation for synthetic datasets. Due to the lack of publicly available code or model from these methods, we either integrated their techniques into our framework or requested the original authors to process our inputs with their models and share the results. Baselines SIPR TR Lumos Quantitative Comparisons. The results in Table. 1 shows our method outperforming SIPR and TR baselines, demonstrating the significance of incorporating advanced rendering physics and reflectance models. The transition from SIPR to TR emphasizes the value of physics-based design, while the shift from TR to our approach underscores the importance of transitioning from the empirical Phong model to the more accurate Cook-Torrance model. Additionally, pretraining contributes to further enhancements, as evidenced by the improved image details, depicted in Fig 6. Quantitative Comparisons. Qualitative Comparisons. Our relighting method exhibits several key advantages over previous approaches, as showcased in Fig. 7. It effectively harmonizes light direction and softness, avoiding harsh highlights and inaccurate lighting that are commonly observed in other methods. A notable strength of our approach lies in its ability to capture highfrequency details like specular highlights and hard shadows. Additionally, as shown in the second row, it preserves facial details and identity, ensuring high fidelity to the subject’s original features and mitigating common distortions seen in previous approaches. Moreover, our approach excels in handling skin tones, producing natural and accurate results under various lighting conditions. This is clearly demonstrated in the fourth row, where our method contrasts sharply with the over-saturated or pale tones from previous methods. Finally, the nuanced treatment of hair is highlighted in the sixth row, where our approach maintains luster and detail, unlike the flattened effect typical in other methods. More qualitative results are available in our supplementary video demonstration. Qualitative Comparisons. User Study. We conducted a human subjective test to evaluate the visual quality of relighitng results, summarized in Table. 2. In each test case, workers were presented with an input image and an environment map. They were asked to compare the relighting results from three methods–Ours, Lumos, and TR–based on three criteria: 1) consistency of lighting with the environment map, 2) preservation of facial details, and 3) maintenance of the original identity. To ensure unbiased evaluations, the order of the methods presented was randomized. To aid in understanding the concept of consistent lighting, relit balls were displayed alongside the images. The study included a total of 256 images, consisting of 32 portraits each illuminated with 8 different HDRIs. Each worker was tasked with selecting the best image for each specific criterion, randomly assessing 30 samples. A total of 47 workers participated in the study. The results indicate a strong preference for our results over the baseline methods across all evaluated metrics. User Study. Ablation Studies. We analyze our two major design choices in Table. 3: the MMAE pre-training framework and DiffuseNet. The MMAE, which integrates dynamic masking with generative objectives, outperforms MAE. This superiority is mainly due to the incorporation of challenging masks and global coherence objectives, enabling the model to learn richer features during pre-training. Furthermore, our method of predicting diffuse render demonstrates superiority over direct albedo prediction. Firstly, we see it simplifies the learning process, as predicting diffuse render is more closely related to the original image. Secondly, our approach effectively distinguishes between the influences of illumination (diffuse shading) and surface properties (diffuse render). This distinction is crucial for accurately modeling the intrinsic color of surfaces, as it enables independent and precise evaluation of each element (see Eqn. 9). In contrast, methods that predict albedo directly often struggle to differentiate between these factors, leading to significant inaccuracies in color constancy, as shown in Fig. 8. Ablation Studies. Applications . We present two applications using predicted intrinsics in Fig. 9. First, real-time PBR via Cook-Torrance Applications components in Three.js graphics library. Second, switching the lighting environment between the source and reference images. Further details are in the supplementary video. Limitations . We identified a few failure cases in Fig. 10. First, our model struggles with neutralizing strong shadows, which leads to inaccurate facial geometry and residual shadow artifacts in both albedo and relit images. Incorporating shadow augmentation techniques [16, 54] during training could mitigate this issue. Second, the model incorrectly interprets reflective surfaces, such as sunglasses, as opaque objects in the normal image. This error prevents the model from properly removing reflective highlights in the albedo and relit images. Lastly, the model inaccurately predicts the albedo for face paint. Implementing a semantic mask [52] to distinguish different semantic regions separately from the skin could help resolving these issues. Limitations This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. available on arxiv

Researchers Asked 47 People to Judge AI-Enhanced Portraits—Here’s What They Chose

Authors:
(1) An Yan, UC San Diego, ayan@ucsd.edu;
(2) Zhengyuan Yang, Microsoft Corporation, zhengyang@microsoft.com with equal contributions;
(3) Wanrong Zhu, UC Santa Barbara, wanrongzhu@ucsb.edu;
(4) Kevin Lin, Microsoft Corporation, keli@microsoft.com;
(5) Linjie Li, Microsoft Corporation, lindsey.li@mocrosoft.com;
(6) Jianfeng Wang, Microsoft Corporation, jianfw@mocrosoft.com;
(7) Jianwei Yang, Microsoft Corporation, jianwei.yang@mocrosoft.com;
(8) Yiwu Zhong, University of Wisconsin-Madison, yzhong52@wisc.edu;
(9) Julian McAuley, UC San Diego, jmcauley@ucsd.edu;
(10) Jianfeng Gao, Microsoft Corporation, jfgao@mocrosoft.com;
(11) Zicheng Liu, Microsoft Corporation, zliu@mocrosoft.com;
(12) Lijuan Wang, Microsoft Corporation, lijuanw@mocrosoft.com. Editor’s note: This is the part 10 of 13 of a paper evaluating the use of a generative AI to navigate smartphones. You can read the rest of the paper via the table of links below. Table of Links Abstract and 1 Introduction
2 Related Work
3 MM-Navigator
3.1 Problem Formulation and 3.2 Screen Grounding and Navigation via Set of Mark
3.3 History Generation via Multimodal Self Summarization
4 iOS Screen Navigation Experiment
4.1 Experimental Setup
4.2 Intended Action Description
4.3 Localized Action Execution and 4.4 The Current State with GPT-4V
5 Android Screen Navigation Experiment
5.1 Experimental Setup
5.2 Performance Comparison
5.3 Ablation Studies
5.4 Error Analysis
6 Discussion
7 Conclusion and References 5.3 Ablation Studies For the ablation studies, we randomly sampled 50 episodes in total from 5 categories, which is a different subset used by the main results. Different tagging methods. We first perform an ablation study to compare the performance with different methods to add tags on screen, shown in Table 4. We consider three methods: (1) By side which adds tags with black squares (same style as (Rawles et al., 2023) by the left side of each detected icon; (2) Red which uses red circles for each tag; (3) Center which adds tags with black squares at the center of each detected box. First, adding tags by the left side of boxes may cause problems, for example, some icons may be too close to each other, hence leading to slightly worse results. For tagging styles, we didn’t find a significant difference between red cycles and black rectangles, though empirically black rectangles (Yang et al., 2023b) perform slightly better. Different prompts. We then perform robustness check with different prompting variants: (1) Baseline: Simply ask GPT-4V to take actions; (2) Think: Prompt GPT-4V to think step by step (Kojima et al., 2022); (3) Detail: Provide more context for this task. Overall, we did not observe improvements by “thinking step by step”, but adding more task descriptions helps GPT-4V to better execute actions. This paper is available on arxiv under CC BY 4.0 DEED license. Authors: (1) An Yan, UC San Diego, ayan@ucsd.edu; (2) Zhengyuan Yang, Microsoft Corporation, zhengyang@microsoft.com with equal contributions; (3) Wanrong Zhu, UC Santa Barbara, wanrongzhu@ucsb.edu; (4) Kevin Lin, Microsoft Corporation, keli@microsoft.com; (5) Linjie Li, Microsoft Corporation, lindsey.li@mocrosoft.com; (6) Jianfeng Wang, Microsoft Corporation, jianfw@mocrosoft.com; (7) Jianwei Yang, Microsoft Corporation, jianwei.yang@mocrosoft.com; (8) Yiwu Zhong, University of Wisconsin-Madison, yzhong52@wisc.edu; (9) Julian McAuley, UC San Diego, jmcauley@ucsd.edu; (10) Jianfeng Gao, Microsoft Corporation, jfgao@mocrosoft.com; (11) Zicheng Liu, Microsoft Corporation, zliu@mocrosoft.com; (12) Lijuan Wang, Microsoft Corporation, lijuanw@mocrosoft.com. Authors: Authors: (1) An Yan, UC San Diego, ayan@ucsd.edu; (2) Zhengyuan Yang, Microsoft Corporation, zhengyang@microsoft.com with equal contributions; (3) Wanrong Zhu, UC Santa Barbara, wanrongzhu@ucsb.edu; (4) Kevin Lin, Microsoft Corporation, keli@microsoft.com; (5) Linjie Li, Microsoft Corporation, lindsey.li@mocrosoft.com; (6) Jianfeng Wang, Microsoft Corporation, jianfw@mocrosoft.com; (7) Jianwei Yang, Microsoft Corporation, jianwei.yang@mocrosoft.com; (8) Yiwu Zhong, University of Wisconsin-Madison, yzhong52@wisc.edu; (9) Julian McAuley, UC San Diego, jmcauley@ucsd.edu; (10) Jianfeng Gao, Microsoft Corporation, jfgao@mocrosoft.com; (11) Zicheng Liu, Microsoft Corporation, zliu@mocrosoft.com; (12) Lijuan Wang, Microsoft Corporation, lijuanw@mocrosoft.com. Editor’s note: This is the part 10 of 13 of a paper evaluating the use of a generative AI to navigate smartphones. You can read the rest of the paper via the table of links below. Editor’s note: This is the part 10 of 13 of a paper evaluating the use of a generative AI to navigate smartphones. You can read the rest of the paper via the table of links below. Editor’s note: This is the part 10 of 13 of a paper evaluating the use of a generative AI to navigate smartphones. You can read the rest of the paper via the table of links below. Editor’s note: This is the part 10 of 13 of a paper evaluating the use of a generative AI to navigate smartphones. You can read the rest of the paper via the table of links below. Table of Links Abstract and 1 Introduction 2 Related Work 3 MM-Navigator 3.1 Problem Formulation and 3.2 Screen Grounding and Navigation via Set of Mark 3.3 History Generation via Multimodal Self Summarization 4 iOS Screen Navigation Experiment 4.1 Experimental Setup 4.2 Intended Action Description 4.3 Localized Action Execution and 4.4 The Current State with GPT-4V 5 Android Screen Navigation Experiment 5.1 Experimental Setup 5.2 Performance Comparison 5.3 Ablation Studies 5.4 Error Analysis 6 Discussion 7 Conclusion and References Abstract and 1 Introduction Abstract and 1 Introduction 2 Related Work 2 Related Work 3 MM-Navigator 3.1 Problem Formulation and 3.2 Screen Grounding and Navigation via Set of Mark 3.1 Problem Formulation and 3.2 Screen Grounding and Navigation via Set of Mark 3.3 History Generation via Multimodal Self Summarization 3.3 History Generation via Multimodal Self Summarization 4 iOS Screen Navigation Experiment 4.1 Experimental Setup 4.1 Experimental Setup 4.2 Intended Action Description 4.2 Intended Action Description 4.3 Localized Action Execution and 4.4 The Current State with GPT-4V 4.3 Localized Action Execution and 4.4 The Current State with GPT-4V 5 Android Screen Navigation Experiment 5.1 Experimental Setup 5.1 Experimental Setup 5.2 Performance Comparison 5.2 Performance Comparison 5.3 Ablation Studies 5.3 Ablation Studies 5.4 Error Analysis 5.4 Error Analysis 6 Discussion 6 Discussion 7 Conclusion and References 7 Conclusion and References 5.3 Ablation Studies For the ablation studies, we randomly sampled 50 episodes in total from 5 categories, which is a different subset used by the main results. Different tagging methods. We first perform an ablation study to compare the performance with different methods to add tags on screen, shown in Table 4. We consider three methods: (1) By side which adds tags with black squares (same style as (Rawles et al., 2023) by the left side of each detected icon; (2) Red which uses red circles for each tag; (3) Center which adds tags with black squares at the center of each detected box. First, adding tags by the left side of boxes may cause problems, for example, some icons may be too close to each other, hence leading to slightly worse results. For tagging styles, we didn’t find a significant difference between red cycles and black rectangles, though empirically black rectangles (Yang et al., 2023b) perform slightly better. Different tagging methods. Different prompts. We then perform robustness check with different prompting variants: (1) Baseline: Simply ask GPT-4V to take actions; (2) Think: Prompt GPT-4V to think step by step (Kojima et al., 2022); (3) Detail: Provide more context for this task. Overall, we did not observe improvements by “thinking step by step”, but adding more task descriptions helps GPT-4V to better execute actions. Different prompts. This paper is available on arxiv under CC BY 4.0 DEED license. This paper is available on arxiv under CC BY 4.0 DEED license. available on arxiv

Study Finds AI Shouldn't Be Made to Think Deeply to Navigate Your Phone

People are living in shopping malls. The once-beacons of capitalism and leisure, the American shopping mall, went from being almost totally vacant and abandoned to being restored as indoor residential communities. https://youtu.be/J1GIF6VNipE?si=XM8pnAnsgS69mBDn&embedable=true To give you a brief history: Arcades were shopping malls in 19th-century Paris. Some of these arcades still exist to this day. Arcades are a series of architectural arches that make up a pathway. Perhaps the most notable points of interest concerning arcades pertain to the very word itself. The term "arcade" contains the French root "arc," relating to an "arch" or "bow." This makes sense when we think of the first arcades as simply a series of arches within an architectural structure. You may have noticed these repeated arches or arcades in the Colosseum of ancient Rome or perhaps in your local mosque or church. Not only does the "arcade" relate to architecture, but it was later used to describe shopping centers or "shopping arcades." In his unfinished, posthumously released work "The Arcades Project," philosopher Walter Benjamin collected various writings about the glass-roofed, iron-covered centers built for leisure and commerce in Paris during the 1930s. This eclectic collection is seen as a framework for post-modernist thinking. It is partly because of this architectural relation that arcades can directly reference shopping malls. Of course, in the modern era, underneath these arches and within shopping centers, we also encounter the type of "arcade" we are more familiar with—the video arcade. Fast forward to today, those very arcades of antiquity, which became the American shopping malls of the modern era, were abandoned and turned into liminal spaces—haunting thresholds with no purpose. Now, they are being reported to have become residential areas, bringing people together as communities under one roof. Isn't it ironic that the concept of Arcadia as a utopia became one of commerce and merchants, only to eventually return to the symbolism related to the utopian society of myth? By 2022, 192 U.S. malls planned to add housing. Lafayette Square Mall in Indianapolis will open apartments, including affordable housing, in a former Sears by 2025. The Arcade Providence in Rhode Island, America’s oldest mall, was converted into micro-apartments and retail spaces years ago. This shift is reshaping Arcadism theory.  Malls—once utopian shopping hubs—declined with tech advances and are now being reborn as places to live. A fascinating full-circle moment. The benefits of these indoor shopping centers as living spaces include the accessibility and utilitarian architectural design that incredibly designed shopping centers facilitate. Through capitalism and innovation come peaks of utopian ideology. This story is still very much ongoing and will be covered in my future works on Arcadism. People are living in shopping malls. The once-beacons of capitalism and leisure, the American shopping mall, went from being almost totally vacant and abandoned to being restored as indoor residential communities. https://youtu.be/J1GIF6VNipE?si=XM8pnAnsgS69mBDn&embedable=true https://youtu.be/J1GIF6VNipE?si=XM8pnAnsgS69mBDn&embedable=true To give you a brief history: Arcades were shopping malls in 19th-century Paris. Some of these arcades still exist to this day. Arcades are a series of architectural arches that make up a pathway. Perhaps the most notable points of interest concerning arcades pertain to the very word itself. The term "arcade" contains the French root "arc," relating to an "arch" or "bow." This makes sense when we think of the first arcades as simply a series of arches within an architectural structure. You may have noticed these repeated arches or arcades in the Colosseum of ancient Rome or perhaps in your local mosque or church. Not only does the "arcade" relate to architecture, but it was later used to describe shopping centers or "shopping arcades." In his unfinished, posthumously released work "The Arcades Project," philosopher Walter Benjamin collected various writings about the glass-roofed, iron-covered centers built for leisure and commerce in Paris during the 1930s. This eclectic collection is seen as a framework for post-modernist thinking. It is partly because of this architectural relation that arcades can directly reference shopping malls. Of course, in the modern era, underneath these arches and within shopping centers, we also encounter the type of "arcade" we are more familiar with—the video arcade. Fast forward to today, those very arcades of antiquity, which became the American shopping malls of the modern era, were abandoned and turned into liminal spaces—haunting thresholds with no purpose. Now, they are being reported to have become residential areas, bringing people together as communities under one roof. Isn't it ironic that the concept of Arcadia as a utopia became one of commerce and merchants, only to eventually return to the symbolism related to the utopian society of myth? By 2022, 192 U.S. malls planned to add housing. Lafayette Square Mall in Indianapolis will open apartments, including affordable housing, in a former Sears by 2025. The Arcade Providence in Rhode Island, America’s oldest mall, was converted into micro-apartments and retail spaces years ago. This shift is reshaping Arcadism theory.  Malls—once utopian shopping hubs—declined with tech advances and are now being reborn as places to live. A fascinating full-circle moment. The benefits of these indoor shopping centers as living spaces include the accessibility and utilitarian architectural design that incredibly designed shopping centers facilitate. Through capitalism and innovation come peaks of utopian ideology. This story is still very much ongoing and will be covered in my future works on Arcadism.

Stories, News & Mentions about SQUARK

Squark’s Stories on HackerNoon

Squark's News & Mentions