At Yogi, we help some of the most recognizable brands in the world create better products and user experiences for their customers by providing deep insight into customer feedback via Yogi's reviews and ratings analysis SaaS platform

Yogi

Machine Coding and Low-Level Design (LLD) rounds are integral components of technical interviews for software engineering positions. These rounds test your ability to transform your coding skills into real-world solutions and require a deep understanding of data structures, algorithms, and software architecture. To help you prepare effectively, I've compiled a comprehensive guide with resources, practice questions, and informative videos to ace these rounds. Most of the resources are based on , but you can get an idea and code with your preferred language. Java programming language Below are some sample questions for the machine coding round: Sample Questions Design a Parking Lot: Another classic LLD question. You are asked to design a parking lot management system. Design an Online Shopping System: A more complex LLD problem that tests your ability to design a scalable e-commerce system. Design an Online Auction System Recently asked me from one of the top companies - more about this later. Recently, I was researching LLD, and I was confused about how and where to start, so I came up with a strategy of my own. I started with the theory first with the help of by Yogita Sharma. She breaks down complex concepts into simple analogies and examples that you will remember easily. I would also suggest you start here and learn what are low-level design basics first and a few design patterns. sudoCode Then, move on to questions and practice them through the different resources mentioned below. Resources: Theory Short Video Lessons - Low-Level Design Course (sudoCode) Everything you need for LLD Interviews in one place - Grokking the Low-Level Design Interview The OG Book - (go through this only if you have enough/extra time to explore, probably months cause it takes you deep into theory) Design Patterns Questions and More Industry - (do at least 2-3 from this resource to get an idea of how the questions are framed) Standard Questions workat.tech - This YouTube playlist consists of most of the questions/solutions, Use case diagram creation, different design patterns, and more. (Mode of Language - Hindi) Concept && Coding - OG Again, most of the standard questions explanation with solutions. Soumyajit Bhattacharyay Mock Interviews Keerti Purswani (OG, GOAT for HLD and System Design) Guarav Sen Tips and Tricks Practice 😀 Only writing code and practicing can make you comfortable with these kind of interviews, and time management is must. Mostly time limit to complete the problem statement is 90-120mins for most companies. I have also practiced some of the machine coding questions myself. You can check out my GitHub repo for the same; I will keep updating it from time to time. Github - System Design and follow me for the latest updates. Star the repository, This is all you need. In conclusion, the Machine Coding and LLD rounds can be challenging, but with the right resources, practice, and understanding, you can excel in these interviews. Remember, it's not just about finding the right answer but also about demonstrating your problem-solving and design skills. So, immerse yourself in practice and theory, and you'll be well-prepared to tackle these technical interviews successfully. Good luck on your coding journey! All the best, and keep grinding.

How to Master Machine Coding/LLD Round: Resources and Tools

The Man Who Could Work Miracles

VR/AR (virtual reality / augmented reality) technologies could help us in some ways. However, they also could pose a threat to humanity. is the environment that our brain was not designed for. Surrounding ourselves with VR/AR may lead to the degradation of what makes us human. It can turn us into irrational and immoral slaves incapable of setting and achieving long-term goals and moving our civilization into a bright future. Metaverse In this post, I'd like to focus on the potential dangers of VR/AR for humanity, if they are used inappropriately. Before introducing VR/AR technologies into our real life, it is important to conduct extensive research on the potential impact they could have on human values and behavior. “Technology and industry have distanced people from nature and magic and human values.” Laura Esquivel Let me start with my personal story of dealing with . To me, VR is not only about 3D world and VR headsets. As a teenager, I was addicted to Sony PlayStation games, specifically soccer simulator. I played for hours almost every day for several years. Eventually, I became a master. I could quite easily win almost every game and all the trophies. But it was not real. It mattered only to me and happened only in my mind and on the TV screen. You see, becoming a winner in a game is about pressing buttons at the right time. I identified myself with great players and big teams, but my body was still the same. I didn't really become much better at playing soccer myself. I realized that all I got is an illusion that I was sort of invincible, but in reality, I wasn't. Not even close. VR My addiction cost me a lot of time and caused some harm to my health. Plus, I kind of forgot who I was. That realization is what pushed me to spend as much time as I could with the real ball and players on the real field. Playing soccer in real life taught me a lot of lessons that I use in life. Now, 15 years later, I would like to take a look at my addiction and the potential impact that Metaverse could have on human behavior and our society from a scientific point of view. “Technology challenges us to assert our human values, which means that first of all, we have to figure out what they are.” Sherry Turkle VR/AR technologies and other can destroy your personality on different levels and make you act irrationally in the real world. ICT (information and communication technologies) Why is it important for us to be moral and rational in our world? Why is it important to preserve universal values? Well, basically, that is what makes us human. To me, being rational means being as close to our reality as possible. It is about acting according to the laws of our reality/universe. The more rational we are, the better choices we make, and the better results we get. Hence, the happier we are ) Universal values are moral principles (our behavioral «compass») shared by the majority of people. We need to act rationally and in accordance with our universal values to stay human. Losing your identity Forgetting who you really are It is really easy nowadays for people to create their online identities. You can modify your photos with the help of a graphics editor and create a user profile pretending to be something you are not. It is easy to change your appearance, but it is hard to change your values. People can forget who they really are this way and become something that other people force them to be. Borrowing identity/behavior or its elements People are not just trying to come up with their own fictitious appearance, but they can also borrow someone's identity/behavior or its elements. In 1961, it was shown by Albert Bandura that children mimic the behavior demonstrated to them. We can learn by imitating others’ behavior [1]. This phenomenon could be explained by the presence of mirror neurons in our brains. These neurons fire when we do certain actions, as well as when we see or hear someone else performing the same actions [2]. This way, a lot of people become full of behavioral patterns that are not always beneficial neither in terms of morality nor in terms of rationality. Think about all the patterns people could get from games, movies, anime etc. Plus, currently, our representations in Metaverse are quite odd (there are no legs etc.). These limitations may cause irrational behavior that people then demonstrate in real life. Also, we can get supernatural abilities in VR. Ultimately, we can get used to them so much that in the real world we forget that we don't have them and behave very irrationally in some situations. We are shaped by our environment. Is Metaverse the best environment for people with all its irrationality and absurdity? Not yet, in my opinion. Neurophysiological foundation. Brain Two brain operating modes Default/Passive mode The default system is the neuronal network that activates at rest, when we are self-absorbed and not engaged in solving the problems of the external world. This is when our brain organizes information and calculates various options to solve problems. This is when our brain is the most creative. Active mode. Responsive problem-solving network / central executive network This system activates when we consume information coming in from the outside. Both systems are antagonistic to each other. In other words, activation of the active mode leads to inhibition of the default mode. During the first 25 years of our life, neuronal networks responsible for the default mode are being formed. That's why it is of critical importance for us to not knock our default system out by keeping our central executive network active all day long (consuming tons of information via gadgets). The brain of teenagers stops forming if they spend 70% of their time in front of their gadgets [3, 4]. Three minds Theo Compernolle distinguishes 3 minds in his book called «BrainChains: Discover your brain, to unleash its full potential in a hyperconnected, multitasking world»: Reflex mind It is the oldest and fastest one. It is the feelings and impressions originating in our mind without effort. It can't act proactively or learn from the past. It is unconscious, autonomic, uncontrolled, and works based on reflexes. Reflecting mind It is slow, conscious, rational, controlled, and works based on logic. It requires a lot of energy and processes information sequentially dealing with one thought/task at a time. It allows us to think about things that don't exist. It allows us to fantasize and invent. It needs to rest and sleep. It can take over the reflex mind. Archiving mind This mind is sort of a «librarian» that receives millions of bites of information coming from the outside through sense organs and thoughts coming from our reflecting mind. It decides what to store in our long-term memory and what to reject. Archiving mind is task-negative, because it is active when there is no any cognitive task done by our mind. It is the most active when our reflecting mind is asleep or at rest [5]. I assume that the reflecting mind and reflex mind correspond to the aforementioned central executive network, and archiving mind corresponds to the default system. Cognitive biases Reflex mind and reflecting mind are what referred to as System 1 and System 2. System 1 is prone to systematic errors/deviations from rationality in judgments. Today we call them . However, in the past, a lot of them (if not all) served as an adaptive mechanism that enabled our ancestors to react instantly to threats and helped us to survive in the wild savanna. But in today's world, System 1 is what makes us irrational. And System 2 is what supposed to take control over System 1 and check its judgments [5, 6]. Daniel Khaneman cognitive biases Now, with the arrival of the Metaverse, we are entering the world of VR/AR. I suppose that in this new environment it is going to be harder for our System 2 to check the rationality of judgments generated by System 1, and I believe we will get a lot of new cognitive biases. This way, it should be even harder for us to understand who we are, what our values are, and where we are going as humanity. Degradation of society and human consciousness Our modern tech-driven world forces us to be hyperconnected and to do multiple tasks at once. This is something that the human brain was not designed for.  We need to rest for our archiving mind to organize information, and our reflecting mind can deal with one task/thought at a time to be efficient. In the XXI century, we constantly get bombarded with tons of information via technologies that we have surrounded ourselves with. Our obsession with digital technologies leads to our intelligence (as social, as ) degradation and results in us losing our ability to focus and even sleep (this is why you may think about using by the way) [5, 7]. emotional f.lux Excessive usage of digital technologies leads to the default system deactivation and the central executive network overactivation. As a result, our default system may not get enough time to even figure out who we are and form our value system. I believe that this is why yogis close their eyes and keep calm during meditations. This helps them to know themselves. Our default system is responsible for our ability to plan the future. Without it, we lose our motivation/desire to evolve and learn, and we can't set and achieve our long-term goals. Plus, our brain just tries to save energy by setting and achieving simpler goals. The default system also has social meaning. Throughout our evolution, it helped people to form social relationships. In today's tech-driven world, we spend less and less time with other people. This leads to the societal degradation of humanity. We start forgetting who we are. We get a sort of "digital autism" epidemic. It is characterized by poverty of speech and social isolation. People lose their ability and desire to connect and communicate with other people, they stop valuing humans and lose their ability to evolve and know themselves and others. At the same time, the overactivation of the central executive network results in our desire to get pleasure as easily and fast as possible. And if we can't achieve this, we may become aggressive and suicidal. Why would we try to achieve our long-term goals, if we can get the same satisfaction in VR without leaving our room? And setting long-term goals is what helps us to survive as a species because it helps to predict the future and act proactively [3]. may be another reason why we become unable to set and achieve long-term goals. We start answering simpler questions. Substitution bias Speaking of motivation, we need to remember the “Candle problem” experiment which showed that extrinsic motivators like money narrow our focus and inhibit our creativity and thinking. And in order to move our science and civilization forward into a bright future, we have to overcome a lot of problems, we need to be as creative as possible. We need to have intrinsic motivation based on our values [8]. We have already brought the world of money, crypto, and NFTs into VR. It would actually be interesting to do the opposite. We could try to build in VR the experimental world without any form of money (think about ). The Venus project VR and AR technologies were created by some people. How do we know what their real intentions are? We need to remember that VR/AR, just like any other technology, can serve as a weapon and a means of manipulation. It is far more dangerous than the real weapon, as it can impact our language, mentality, and culture. VR can be used to manipulate our consciousness and destroy whole nations. It can destroy what makes us human, our universal values [3]. Biochemical foundation. Addiction I believe you noticed that the Internet is quite addictive. With the introduction of cryptocurrency, Internet has become much more addictive. Being the next-generation Internet (Internet + VR/AR), Metaverse should be even more addictive. On a molecular level, our addiction is caused by molecules like . This molecule, produced in different parts of the brain, forces us to constantly seek/want new information. This is why social media news feeds are so addictive. Dopamine helps to keep us motivated in the learning process, which is critical for our survival. Other molecules called allow us to feel pleasure and be satisfied with results. These two systems complement each other. The first one is stronger than the second one. Usually, we are not satisfied with the results we get, so we keep searching anyway. dopamine opioids In today's tech-driven world, these systems get abused. It becomes harder and harder for people to resist the temptations (checking emails, twits, messages etc.). The same happens to people in the Metaverse which provides even more opportunities [9]. Ultimately, we can get exhausted and unable to set and achieve long-term goals . constantly seeking pleasure like rats Biomechanical catastrophe As far as I know, when it comes to movement, Metaverse solution is the omnidirectional treadmill that allows people to walk and run in any direction staying actually in the same place. To be fair, running and walking on them looks pretty unnatural. And natural movement is exactly what we need in order for our vertebral column to stay healthy. We need to let our column to move like a spring to feed the cells of . We need to walk at least 1 hour every day for our intervertebral discs to stay healthy. You can't walk this way in the room with VR headset, because the space is limited. You can try outside your house, but then there are some risks to get hit by a car or something. intervertebral discs No natural movement - no life. People use a lot of tools to transmit information: inkwell and feather pen, ball pen, keyboard, and stylus. VR requires us to use motion controllers now. There are different rhythms in our bodies. Superimposing rhythms (including the rhythm of the heart) with muscle efforts is beneficial in that case, if they are followed by periods of rest. In this case, muscle efforts strengthen the rhythm of the heart. This happens when we use a feather pen. Otherwise, a lot of our muscles get blocked, and our bodies get poisoned by metabolic waste. This happens when we use a ball pen. In this case, our diaphragm gets blocked, and other body rhythms superimpose with muscle efforts inappropriately as well. I believe the same happens when we use a stylus and motion controllers for VR. In terms of health, a feather pen is the most beneficial tool [10]. In search of harmony People are reflections of the environment they are in. And Metaverse can change our environment drastically. It is up to us to decide in which direction. We can either take into account the properties of our brain and use VR/AR to build an environment that will help the human mind to reach its full potential and develop our inner world and human values. For example, we could use VR/AR in education (think of , for example). Or, we could create such an environment that will turn humanity into unethical and immoral slaves of our technologies. virtual science simulations Some people use drugs and alcohol to escape reality. Why don't we bravely face our reality, acknowledge our problems, and do whatever we can to overcome them, rather than trying to escape them and hide in the Metaverse? When joining crypto or Metaverse projects, it is easy to ask questions like: How much money will we get? instead of Who do we become? and What kind of person this type of environment makes me? teaches people in his books that the most important is who we become. Jim Rohn It is time for people to ask the right questions: Do we become rats constantly searching for pleasure and repeating like a mantra «To the MOON!»? or Do we become determined people with beautiful dreams and a clear vision of our future? Hope that answering these questions will help us to do the right choice. Metaphorically speaking, VR/AR technologies should not become our «crutches». In this case, human values will «atrophy». VR/AR should become «ski poles» that will facilitate the growth of our morality and help us to move our civilization into a humane future. The headline image was composed by me with the help of the , , , , , and images. All other images sourced from Pixabay. choice blue ball red ball joystick VR headset soccer field The divider was created by me. Reference Bandura and Bobo Keysers C. Mirror neurons. Volume 19, ISSUE 21, PR971-R973, November 17, 2009 Nekrasov, A.S. & Nekrasova, N.A. & Nekrasov, S.I.. (2021). Impact of information technology on a person and his consciousness. Ekonomicheskie i sotsial’no-gumanitarnye issledovaniya. 130-135. 10.24151/2409-1073-2021-2-130-135. Raichle М.Е., MacLeod А.М., Snyder A.Z., Powers W.J., Gusnard D.A., Shulman G.L. A default mode of brain function. Proceedings of the National Academy of Sciences of the United States of America, 2001. 98(2). P. 676—682. Theo Compernolle. BrainChains: Discover your brain, to unleash its full potential in a hyperconnected, multitasking world, 2014. Daniel Kahneman. Thinking, Fast and Slow, 2011 Small GW, Lee J, Kaufman A, Jalil J, Siddarth P, Gaddipati H, Moody TD, Bookheimer SY. Brain health consequences of digital technology use. Dialogues Clin Neurosci. 2020 Jun;22(2):179-187. doi: 10.31887/DCNS.2020.22.2/gsmall. PMID: 32699518; PMCID: PMC7366948. GLUCKSBERG S. The influence of strength of drive on functional fixedness and perceptual recognition. J Exp Psychol. 1962 Jan;63:36-41. doi: 10.1037/h0044683. PMID: 13899303. Why We're All Addicted to Texts, Twitter and Google Vladimir Bazarny. The Human Child: Psychophysiology of Development and Regress, Jun 16, 2020

Metaverse + Realverse: In Search Of Harmony

Types of Influencers You Can Choose Based on Your Niche

Every year, over 500,000 new businesses start in the US, and over 200 billion dollars are spent on marketing, as reported by . Statista It can be difficult for any company to stand out in such a crowded marketplace. It's become more crucial than ever to have an innovative, adaptable, and persuasive marketing plan to stand out from the competition. a strategy that fosters virality, word-of-mouth, and organic growth in addition to aiding in customer acquisition. Growth marketing is the term of this innovative and effective strategy for creating a devoted user base. Let's examine what growth marketing comprises and what qualifications a growth marketer needs to succeed. Growth Marketing: What is it? Growth marketing is a data-driven strategy that employs experiments to figure out how to improve outcomes. Using an A/B test, for instance, to determine which push notification consumers prefer. When a system is effective, it's used constantly. When a system isn't, more tests are run. Marketing initiatives have typically been kept apart from product development, with the marketing division kept separate from other parts of the company. Growth marketing is now part of the product development process. The growth metrics that matter must be thoroughly studied by growth marketers. Although the term “growth marketing" suggests a strategy solely focused on luring in more people, there is more to it than that. Customer discovery is where growth marketing must put its roots. Customer discovery, also known as product/market fit, is exactly what it sounds like: identifying the people who are experiencing the issues that your solution can solve most effectively. Steve Blank, a businessman, and author asserts that: "No company plan survives initial interaction with clients." The growth marketing strategy can also be directed by a thorough grasp of the consumer to best serve the company's objectives. For instance, LinkedIn identified users' intense interest in their vanity metrics, which made it simple for the company to persuade users to provide more information, such as personal contacts, to benefit from the network effect. Growth marketing's objectives The broad objectives of growth marketing are objectives that any company should have: maintain current clients Obtain new clients The rise in revenue. Growth plans improve in effectiveness as the company grows, both in terms of cost and conversion rate. But how can one expect that growth marketing will help them achieve any of these broad objectives? Client Retention Retaining current consumers is arguably growth marketing's most crucial objective. To ensure loyalty to your company's product or service, the satisfaction of current customers should be given priority. It is far simpler to sell to an established customer than a new one since there is less friction when they make a repeat transaction. In light of this, it is paradoxical to attract new clients if existing ones aren't sticking around. A business concept that has trouble keeping customers can suffer from high customer acquisition expenses. Prioritize client retention and happiness before scaling consumer acquisition expenditures. Brand loyalty increases as people return and continue to be pleased with their experiences. Word-of-mouth marketing can be fueled by brand advocates to increase consumer acquisition. Regions to check Incentive schemes Onboarding new users Email triggers Getting new clients The chase for new consumers becomes commercially viable if customer retention has reached a stable level. The capacity to successfully approach new users and persuade them to buy is the acquisition of new consumers. Being visible in the marketplace is one of the foundational principles of client acquisition. Where are clients looking for a solution to their issue, and how does your offering fit into this? It might be an indication of sustainability to know how your customer acquisition strategy fits with customer retention. How do these two growth marketing objectives advance the next objective—profit? Regions to check the internet Paid searches Risk-free test periods Increased earnings If your growth marketing strategy isn't sustainable, your revenue may increase while profitability decreases. A growth marketing strategy that is effective will open up new revenue streams without unduly raising costs. If done correctly, client lifetime value will rise as retention rates outperform attrition. Regions to check Annual payment Customer feedback List of package deals Traditional marketing versus growth marketing Traditional marketing and growth marketing are highly dissimilar, just as digital marketing and traditional marketing are very dissimilar. Growth hacking is less about using strategies that are often utilized in marketing and more about having a mentality for growth. An open mind is necessary for growth marketing Think about Frank Zappa's comparison of the mind to a parachute. If it's closed, it won't function. Although leaving the status quo can be challenging, perhaps we should all concentrate on the "status Woah." Consider what tiny adjustments can amaze your users. Businesses with a growth mindset don't squander money on marketing initiatives that haven't worked. Before committing fully to a certain strategy, growth marketers ease into their marketing budgets by taking an experimental approach to these investments. It can be challenging to demonstrate the return on investment for many traditional marketing techniques. When compared to the click-through rate of an online advertisement, it is harder to determine and quantify the value of a physical location like a billboard or magazine ad. Because of this, a growth marketing approach prefers the flexibility to any strict norm. Process of Growth Marketing Growth marketing is based on the iterative learning principle, which involves systematic optimization and gradual increases in client acquisition, retention, and lifetime value. Growth marketers can learn from this process which approaches are effective and, more importantly, which ones are not. What steps must you take to perfect the growth marketing process? Charts measuring hypothetical performance for growth marketing Actionable Metrics Definition It's crucial to incorporate a control group in the testing procedure as a baseline to assess the real performance of variables. A control group is isolated to prevent results from being incorrectly interpreted or given the wrong credit. Growth marketers need to be aware of the metrics that are important to their company and assess the results of experiments. Every marketer must be aware of vanity metrics, even though they might change greatly depending on the business. Vanity metrics are fake measurements that mimic sustainable progress. Similar to the number of app installs, these don't truly show whether the app is seeing a strong retention rate, a crucially important indicator. Test A to Z Possessing the courage to try new things is perhaps the most crucial element of growth marketing. Experimentation can produce significant effects even with small changes. A/B testing is a well-liked application of the kind of experimentation typical in growth marketing due to its relative ease. However, we quip lightly when we question, "Why stop at B?" Testing the moving elements and their effects on performance indicators is a key component of growth marketing. The effect that seemingly random aspects of marketing content, such as font weight, background color, and button corner radius, can have on conversion rates may surprise you. Because of this, some of the most crucial work done by marketers involves micro testing. Adopt Improvements Incrementally It's time to grow what worked now that you have observed a considerable improvement in key indicators. Finding techniques that performed well in testing, if used in a population sample, are probably effective when applied on a broad scale. Growth is scalable, but growth is frequently not scalable. For this reason, it's crucial to assess whether solutions are effective before increasing these techniques. Fail quickly, absolutely, but progress as slowly as is necessary to succeed over the long term. Growth Marketing Guidelines Growth marketing strategies might not be as appealing as quick growth hacks, but they are nonetheless crucial for any business pursuing long-term expansion. Growth marketing is based on best practices that cover planning, carrying out, reviewing, and revising. Purpose of growth marketing Establish objectives and goals. Setting goals and performance targets can help choose which techniques to use, even though plans and forecasts may never match up with reality. “You’ve got to be very careful if you don’t know where you are going because you might not get there, '' Yogi Berra memorably said. Be flexible. It can be tempting to use every marketing tool in your toolbox if you have the responsibility of expanding a company. However, when crossing new waters, it's crucial to use caution and prudence. Sometimes doing nothing is the hardest thing to do. By starting small, validating a technique, and only moving forward if it is successful, we must learn to develop patience and discipline. Try Out Different Approaches Since the best practices are so similar, the essay might also be named "How Growth Marketing Is Like Investing." Portfolio diversity, which is a best practice in growth marketing, is a frequent theme in the investment world. Don't rely solely on a single approach by putting all your eggs in one basket. Choose the growth marketing techniques that will work best together by distributing a smaller amount of the overall budget among them. Recognize Your Failures It's critical to be aware of failing tactics and ready to abandon them. The sunk cost fallacy is the propensity to stick with the same course of action even when the results are unfavorable rather than changing it. Recognize when something isn't working out and make the required changes to your approach. Remember Where You Came From Keep in mind how crucial customer retention is. Growth marketers frequently undervalue the significance of customer retention in favor of obtaining new clients. For instance, mobile app marketers will occasionally attempt to conceal detrimental retention numbers like daily active users. According to infamous Silicon Valley investor and founder of Ycombinator Paul Graham: "It’s better to have 100 people that love you than a million people that sort of like your product" Focus on acquiring 100 devoted fans rather than attempting to obtain millions of readers, downloads, or consumers. After that, scale the model you have shown to be effective with the first 100. Yes, it is feasible for a fundamental shift in strategy to have a direct, beneficial effect on a company. It is more likely, though, that growth will occur iteratively and that results will follow the conventional hockey stick pattern, becoming more prominent over time. Conclusion Growth marketers must have an analytical mindset and be prepared to act on their discoveries. The growth marketing toolbox is essential to laying a solid foundation for growth, much like the relationship between a carpenter and their toolbox. Knowing what tactics to use, what metrics to monitor, and which tools to monitor them can be challenging.

The Complete Guide to Growth Marketing

Part of the broader and computer vision realms, human pose estimation (HPE) technology has been gradually making its presence seen in all kinds of software apps and hardware solutions. Still, human pose estimation seemed to be stuck at the edge, failing to cross into mainstream adoption. artificial intelligence But the COVID-19 outbreak pushed entire industries over the technology tipping point. that the share of digital products in companies’ portfolios has accelerated by seven years. And that affected the computer vision market, too. It reached the in 2020 and is projected to grow further at a CAGR of 7.3% by 2028. McKinsey reports value of $11.32 billion Such rapid growth in demand has marked the time for human pose estimation technology to step into widespread commercial use. In this article, we will look under the hood of human pose estimation technology, explore its use case, and share tips from on how to implement HPE so that it drives real value. the ITRex CTO Kirill Stashevsky What is human pose estimation? It is easy for a human to tell whether a person is squatting down or waving a hand. For a machine? Not so simple. When shown a photo or a video frame, a computer “sees” a collection of pixels. Human pose estimation technology enables a computer to model human bodies and recognize body postures in images and videos, including real-time footage. How does human pose estimation work? Traditionally, human pose estimation relied on machine learning algorithms, specifically random forest. The boiled down to presenting a human body as a collection of parts arranged in a deformable structure, such as the one below. essence of the ML-based approach But the classical, ML-based method struggled to identify hidden body parts and failed at multi-person pose estimation. That has sparked the transition to deep learning that is used for human pose estimation today. There are many ways to interpret a human posture with deep learning (hence, the many approaches and libraries, which we explore in the subsequent section). Here’s how a typical deep learning model for human pose estimation runs. A standard deep learning model for human pose estimation has a convolutional neural network (CNN) at its base. And the CNN’s architecture comprises two key components — an encoder (in some approaches referred to as an estimator) and a decoder (or a detector). An encoder extracts specific features, called keypoints, from an input image. The number of the keypoints depends on the approach and usually ranges from 17 to 33. Examples of keypoints might include a left elbow, a right knee, or a neck base. The encoder runs the input through a set of narrowing convolution blocks and extracts the features of all people in the image. A decoder then creates a keypoint probability heatmap. It helps refine the encoder’s predictions and estimates how likely the extracted features are to be located in the marked areas of the image. So, the CNN maps the pre-estimated keypoints to the heatmap and outputs those with the highest probability rate. It then groups the associated keypoints into a skeleton-like body structure. For multi-person estimation, when there are several high-probability areas for each keypoint, say, two left elbows, an additional post-processing layer may be added to ensure each keypoint belongs to the right person. The method described above follows a bottom-up pipeline. It means that a CNN locates every instance of a particular keypoint first and then groups the associated ones into a body structure. A model may follow a top-down approach, too. In this case, a CNN first locates all the bodies in an image, putting them into bounding boxes, and then detects the keypoints within each bounding box. A rundown of essential human pose estimation approaches and libraries There are many modifications of the aforementioned method. Let’s look at the most popular ones and explore common open-source and proprietary libraries that could ease the implementation of human pose estimation. Common approaches to human pose estimation OpenPose A famous model, OpenPose relies on a convolutional neural network of two branches. One creates confidence maps for each keypoint found in an image. The other estimates the degree of association between the keypoints, predicting the so-called affinity fields. Combining the output from two branches, OpenPose builds a model of the human skeleton. And the subsequent layers of the network are used to refine the prediction. DeepCut In contrast to the previous model, wherein the CNN detects individual keypoints and uses them to “assemble” the skeleton, DeepCut handles both tasks simultaneously. It infers all the body parts in an image, classifies them, and assigns those lying close to each other to a specific person. RMPE RMPE helps improve the accuracy of top-down human pose estimation, where a posture is inferred within a bounding box. So, the overall performance of the model depends much on how accurately the body is localized. To reduce slips in bounding box localization, the authors of RMPE introduce an additional Symmetrical Spatial Transformer Network (SSTN) that cuts out a highly accurate single-person area from a potentially misplaced bounding box. So, the keypoints are detected within this area. Finally, a Spatial De-transformer Network (SDTN) is used to remap the output of the two components to the original image. Mask R-CNN With Mask R-CNN, the localization of a human body and keypoint detection run independently. A standard convolutional neural network extracts the keypoints, while a Region Proposal Network (RPN) locates the body. The extracted features are then passed into the parallel branches of the network that refine the predictions and generate body segmentation masks. Based on these masks, the network builds skeleton-like models for each person in the image. Libraries for speedier implementation of human pose estimation BlazePoze BlazePoze is a pose detection library developed and supported by Google. The library allows crafting robust human pose estimation engines that run in real time. In contrast to similar approaches that typically rely on COCO topology and detect 17 keypoints, BlazePose’s skeleton structure features 33 keypoints, including hands and feet, which is particularly beneficial for fitness and sports applications. Image source: Google AI Blog, On-device, Real-time Body Pose Tracking with MediaPipe BlazePose The keypoints are characterized by x and y coordinates, visibility, and vertical alignment. Unlike similar libraries that only rely on heatmaps for keypoint refinement, BlazePose uses a regression approach that combines heatmaps and offset prediction. Image source: Google AI Blog, On-device, Real-time Body Pose Tracking with MediaPipe BlazePose MoveNet MoveNet is an open-source library for 3D pose estimation that detects 17 keypoints. The library comes in two variants: Lightning and Thunder. The former is intended for latency-critical applications, while the latter is appropriate for apps that require high accuracy. Both Lightning and Thunder run faster than in real time on desktop, laptop, and mobile devices. The MoveNet architecture comprises a feature extractor and a set of four prediction heads — a person center heatmap, a keypoint regression field, a person keypoint heatmap, and a 2D keypoint offset field. Image source: TensorFlow Blog, Next-Generation Pose Detection with MoveNet and TensorFlow.js CenterNet CenterNet is a point-based object detection framework that can be extended to human pose estimation. CenterNet models people as a single point corresponding to the center point of a body’s bounding box. The model then determines other object characteristics, such as size, 3D location, orientation, and pose. Image source: Objects as Points, Zhou, Xingyi; Wang, Dequan; Kraehenbuehl, Philipp Where is human pose estimation applied? Recognizing human posture and movements has long been in focus for major industries, including sports, retail, and entertainment. Here’s a run-through of sectors where human posture estimation is in use. Fitness The pandemic has pushed more people to practice physical activities at home, making the demand for fitness apps with human pose estimation grow rapidly. HPE-powered fitness applications provide detailed live feedback on whether a user performs an exercise correctly. For that, a human pose estimation component compares a model extracted from camera footage with a benchmark, thus, providing for a safer home workout routine. Fitness apps featuring human pose estimation cater to various activities, from yoga to dancing to weight lifting. Professional sports Human pose estimation technology can help athletes improve their performance and assist judges in rating athletes unbiasedly. HPE-powered applications are applied for various tasks — from assessing the quality of figure skating elements to helping soccer players strike perfect kicks to allowing high jumpers to polish up their techniques. Gaming and filmmaking Character animation has long been an exhaustive and complex process. Today, it is facilitated with the help of human pose estimation. Graphics, textures, and other enhancements can now be easily applied to a tracked body, so the graphics render naturally even if the body actively moves. In interactive video gaming, human pose estimation is used to capture players’ motions and render them into the actions of virtual characters as well. Retail Whether trying to curb the effect of the pandemic or realize their vision of a , retailers have started turning to AR and real-time virtual effects. Human pose estimation backs up those aspirations, enabling such experiences as virtual try-on and real-time marketing. An HPE-powered app, whether running on a customer’s mobile phone or integrated into a fitting room’s mirror, allows scanning a person’s body and imposing 3D virtual elements on the estimated posture. And that works for trying out everything from clothes to shoes to jewelry. supermarket of the future Robot training Traditionally, industrial robots were trained with the help of 2D vision systems that called for time- and effort-intensive calibration. Today, human pose estimation provides for faster, more responsive, and more accurate robot training. Instead of programming robots to follow the set trajectories, one may teach a robot to recognize the pose and the motions of a human. Having estimated the posture of the demonstrator, a robot then devises the way it should move its articulators to perform the same motion. Security and surveillance Human pose estimation may be applied to analyze the footage from security cameras to prevent potentially alarming situations. Identifying a human posture and estimating its anomaly score, HPE-powered security software may predict suspicious actions or identify people who have fallen down or, say, potentially feel sick. Implementing human pose recognition: the peculiarities to keep in mind ITRex Group has recently helped a fitness tech startup create a . We sat down to talk to Kirill Stashevsky, the ITRex CTO, to discuss the specifics of implementing human pose estimation technology that contribute a lot to the project’s success but are often overlooked. fitness mirror powered by artificial intelligence and human pose estimation — How does one embark on the HPE implementation journey to ensure they produce a top-notch solution? What should one beware of during project planning to secure that further development efforts are headed in the right direction? Whether you craft a winning human pose estimation product depends largely on the decisions you make at the very beginning of the project. One such decision is selecting the optimum implementation strategy — one may choose to develop a solution from scratch or rely on one of the many human pose estimation libraries. Kirill: To choose the best-fitting approach, you need to clearly understand, among other issues, what exactly you aim to achieve with your future product, which platforms it will run on, and how much time you have until releasing your product to the market. Once you’ve clarified the vision, weigh it against the available strategies. Consider going the custom route if the task you are solving is narrow and non-trivial and requires the ultimate accuracy of human pose estimation. Keep in mind, however, that the development process is likely to be time- and effort-intensive. In turn, if you are developing a product with a mass-market appeal or a product that caters to a typical use case, going for library-based development would help build a quality prototype faster and with lower effort. Still, in many cases, you would have to adjust the given model to your specific use case by further training it on the data that best reflects the reality. — Suppose I decide to go for library-based development; what factors should I consider to choose the fitting one? You may go for a proprietary or an open-source library. Proprietary libraries could provide for more accurate pose estimation and require less customization. But you have to prepare a backup plan in case an owner, say, discontinues the library support. Kirill: Open-source libraries, in turn, often require more effort to configure. But with an experienced team, it may be an optimum option balancing the quality of recognition, moderate development costs, and fair time-to-market. Pay attention to the number of keypoints a library is able to recognize, too. A solution for dancers or yogis, for instance, may require identifying additional keypoints for hands and feet so that BlazePose might look like a more reasonable option. If latency is critical, select the library that runs at an FPS rate of 30 and higher, for example, MoveNet. — Why aren’t the standard datasets lying at the base of most models enough for an accurately performing solution? What data should I then use to further train and test the model? A well-performing human pose estimation model should be trained on the data that is authentic and representative of reality. The truth is that even the most expansive datasets often lack diversity and are not enough to yield reliable outcomes in real-life settings. That’s what we faced when developing the human pose estimation component for our client’s fitness mirror. Kirill: To mitigate the issue, we had to retrain the model on additional video footage filmed specifically to reflect the surroundings the mirror will be used in. So, we compiled a custom dataset of videos featuring people of different heights, body types, and skin colors exercising in various settings — from poorly lit rooms to spacious fitness studios — filmed at a particular angle. That helped us significantly increase the accuracy of the model. — Are there any other easy-to-overlook issues that still influence the accuracy of human pose estimation? Focusing on the innards of deep learning, development teams may fail to pay due attention to the cameras. So, make sure that the features of a camera (including its positioning, frame size, a frame rate, a shooting angle, as well as whether the camera is shooting statically or dynamically) used for filming training data match those of cameras real users might employ. Kirill: Let’s recap So, you are considering implementing a solution with human pose estimation features. Here are vital things to keep in mind to develop a winning application: Make sure that the strategy you choose to go by supports your vision. Go custom if you’re developing a solution for a specific, non-trivial task. Alternatively, consider implementing your future solution based on a readily available library — either open-source or proprietary When choosing an optimum human pose estimation library, pay particular attention to the number of keypoints the offered model can recognize and the processing speed, especially if your future application is latency-critical Whether you decide to develop a human pose estimation application from scratch or opt for library-based development, you will have to train a model further to guarantee it performs accurately When training the model, use data that is as diverse and representative of reality as possible At the very start of the project, make sure you have enough data to test the model Looking to develop top-notch software that recognizes body postures as accurately as a human eye? , and they will help you turn human pose estimation technology to your advantage. Tell ITRex about your vision

YOGI

meetyogi.com

-- employees

Since n.d.

COMPANY RANKING

YOGI

EVERGREEN INDEX #12047

Representing this company? Click here to claim and customize this page!

Yogi WIKI

Apply to get your Tech Company News Page live on HackerNoon.com today!

YOGI

meetyogi.com

-- employees

Since n.d.

COMPANY RANKING

YOGI

EVERGREEN INDEX #12047

RELATED COMPANIESRANK

STORIES ABOUT YOGI

Representing this company? Click here to claim and customize this page!

Yogi WIKI

Apply to get your Tech Company News Page live on HackerNoon.com today!