Big Data on Gesture Recognition and Machine Learning
Since our technology is more and more advanced as time goes by, traditional human-computer interaction has become increasingly difficult to meet people’s demands. In this digital era, people need faster and more efficient methods to obtain information and data. Traditional and single input and output devices are not fast and convenient enough, it also requires users to learn their own methods of use, which is extremely inefficient and completely a waste of time. Therefore, artificial intelligence comes out, and its rise has followed the changeover times, and it satisfied people’s needs. At the same time, gesture is one of the most important way for human to deliver information. It is simple, efficient, convenient, and universally acceptable. Therefore, gesture recognition has become an emerging field in intelligent human-computer interaction field, with great potential and future.
Keywords: gesture recognition, human, technology, big data, artificial intelligence, body language, facial expression
Technology is probably one of the most attracting things for people nowadays. Whether it is the new iPhone coming out or some random new technology that is bring into our life. It is a matter of fact that technology has become one of the essential parts of our life and our society. Simply, our life will change a lot without technology. As of today, since technology is improving so fast, there are many things that can be related to AI and machine learning. A lot of the ordinary things around our life becomes data. And the reason why they become data is because there is a need for them in having better technology to improve our life. For example, language was stored into data to produce technology like translator to provide convenience for people that does not speak the language. Another example is that roads were stored into data to produce GPS to guide direction for people. Nowadays, people values communication and interaction between others. Since gesture recognition is one of the most important ways to understand people and know their emotion, it becomes a popular field of study for many scientists. There are multiply field of study in gesture recognition and each require a lot of amount of time to know them well. For the report, we do research about hand gesture, body gesture and facial expression. Of course, there will be a lot of other fields related to gesture recognition, for example, like animal gestures. They all can be stored into data and get study in the research by scientists. Many people might have question about how gesture recognition are has anything to do with technology. They simply do not think that they can be related, but in fact, they are related. Companies like Intel and Microsoft have already created so many studies for new technology in that field. For example, Intel proposed combining facial recognition with device recognition to authenticate users. Studying gestures recognition will often reveal what the think. For example, when someone is lying, their eye will tend to look around and they tend to touch their nose with their hand, etc. So, studying gesture recognition will not only help people understand much more about human beings and it can also help our technology grow. For example, in AI and machine learning, studying gestures recognition will make or improve AI and machine learning to better understand humans and be more human-like.
Nowadays, people are doing more and more high-tech research, which also makes various high-tech products appear in society. For people, electricity is as important as water and air. Can’t imagine life without electricity. We can realize that technology is changing everything about people from all aspects. People living in the high-tech era are also forced to learn and understand the usage of various high-tech products. As a representative of high technology, artificial intelligence has also attracted widespread attention from society. Due to the emergence of artificial intelligence, people have also begun to realize that maintaining human characteristics is also an important aspect of high technology.
People’s living environment is inseparable from high technology. As for the use of human body information, almost every high-tech has different usage 1. For example, face recognition is used in many places to check-in. This kind of technology enables the machine to store the information of the human face and determine whether it is indeed the right person by judging the five senses. We are most familiar with using this technology in airports, customs, and companies to check in at work. Not only that, but the smartphones we use every day are also unlocked through this face recognition technology. Another example is the game console that we are very familiar with. Game consoles such as Xbox and PS already have methods for identifying people’s bodies. They can identify several key points of people through the images received by their own cameras, thus inputting this line of action into the world of the game.
Many researchers are now studying other applications of human movements, gestures, and facial expressions. One of the most influential ones is that Google’s scientists have developed a new computer vision method for hand perception. Google researchers identified the movement of a hand through twenty-one 3D points on the hand. Research Engineers Valentin Bazarevsky and Fan Zhang stated that “The ability to perceive the shape and motion of hands can be a vital component in improving the user experience across a variety of technological domains and platforms 2.” This model can currently identify many common cultural features. gesture. They have done experiments. When people play a game of “rock, paper, scissors” in front of the camera, this model can also judge everyone’s win or loss by recognizing gestures.
More than that, many artificial intelligences can now understand people’s feelings and intentions by identifying people’s facial expressions. This also allows us to know how big a database is behind this to support the operation of these studies. But collecting these data about gesture recognition is not easy. Many times we need to worry about not only whether the data we input is correct, but also whether the target identified by artificial intelligence is clear and the output information is accurate.
3. Gesture Recognition
Gesture recognition is mainly divided into two categories, one is based on external device recognition, the specific application is data gloves, wearing it on user’s hand, to obtain and analysis information through sensors. This method has obvious shortcomings, though it is accurate and has excellent response speed, but it is costly and is not good for large-scale promotion. The other one is the use of computer vision. People do not need to wear gloves. As its name implies, this method collects and analyzes information through a computer. It is convenient, comfortable, and not so limited based on external device identification. In contrast, it has greater potential and is more in line with the trend of the times. Of course, this method needs more effective and accurate algorithms to support, because the gestures made by different people at different times, in different environments and at different angles also represent different meanings. So, if we want more accurate information feedback. Then the advancement of algorithms and technology is inevitable. The development of gesture recognition is also the development of artificial intelligence, a process of the development of various algorithms from data gloves to the development of computer vision-based optical technology plays a role in promoting it.
3.1 Hand Gesture
3.1.1 Hand Gesture Recognition and Big Data
Hand gesture recognition is commonly used in gesture recognition because fingers are the most flexible and it is able to create different angles that will represent different meanings. The hand gesture itself is also an easy but efficient way for us human beings to communicate and send messages to each other. The existence of hand gestures can be considered easy but powerful. However, if we are using the application of hand gesture recognition, it is a much more complicated process. In real life, we can just ben our finger or simply make a fist so that other people will understand our message. But when using hand gesture recognition there are many processes that are being involved. Hand gesture is commonly used in geesture recoginitaion As we did our research and based on our life experiences, hand gesture recognition is a very hot topic and has all the potential to be the next wave. Hand gesture recognition has recently achieved big success in many fields. The advancement and development of hand gesture recognition is also the development of other technology such as the advancement of computer chips, the advancement of algorithms, the advancement of machine learning even advancement of deep learning, and the advancement of cameras from 2D to 3D. The most important part of hand gesture recognition is big data and machine learning. Because of the development of big data and machine learning, data scientists are able to have better datasets, build a more accurate and successful model and be able to process the information and predict the most accurate results. Hand gesture recognition is a significant link in Gesture recognition.However gesture recognition is also not only about hand gesture recognition, it also includes other body parts such as facial expression recognition and body gesture recognition. With the help of the whole system of different gesture recognitions, the data can be recorded and processed by AIs. The results or predictions can be used currently or later on for different purposes in different areas 3.
3.1.2 Principles of Hand Gesture Recognition
Hand gesture recognition is a complicated process involving many steps. And in order to get the most accurate result, it will need a large amount of quality data and a scientific model with high performance. Hand gesture recognition is also at a developing stage simply because there are so many possible factors that can influence the result. Possible factors include skin color, background color, hand gesture angle, and Bending angle, etc. To simplify the process of gesture recognition, AIs will use 3-D cameras to capture images. After that, the data of the image will be collected and processed by programs and built models. And lastly, AIs will be able to use that model to get an accurate result in order to have a corresponding response or predict future actions. To explain all processes of hand gesture recognition in detail, it includes graphic gathering, retreatment, skin color segmentation, hand gesture segmentation, and finally hand gesture recognition. Hand Gesture Recognition can not achieve the best accuracy without all any of these steps. Within all these steps, skin color segmentation is the most crucial step in order to increase accuracy and this process will be explained in the next session 4.
3.1.3 Gesture Segmentation and Algorithm, The Biggest Difficulty of Gesture Recognition.
If someone actually asks us a question which is what kind of recognition is going to have the maximum potential in the future? We will have Hand Gesture recognition as my answer without a doubt. Because in my opinion, Hand Gesture Recognition is really the next wave, as our technology is getting better and better, it will be a much easier and more efficient type of recognition that could possibly change our lives. If you compare Hand Gesture Recognition with Voice Recognition, you will see the biggest difference because everyone is using Hand Gesture all over the world in different ways while Voice is limited to people that are unable to make a sound and sound is a much more complicated type of data that in my opinion is not efficient enough to deliver a message, at least with lots of evidence indicating it is not easier than Hand Gesture Recognition. However, it doesn’t mean hand gesture doesn’t have any limit. Instead, Hand Gesture Recognition is influenced by the color in many ways including skin colors and the colors of the background. But skin color is also a great characteristic of recognition at the same time. So if we could overcome this shortcoming or obstacle, the biggest disadvantage of Hand Gesture Recognition could also become its biggest advantage since skin color has so many amazing characteristics that could be used as a huge benefit for Hand Gesture Recognition. Firstly, skin color is a unique attribute which means it has a similar meaning all over the world. For example, Asian people mostly have yellow skins, Western people mostly have white skins while African American people mostly have black skins. People might form different regions from all over the world but since their skins are similar in many ways, they are most likely to have at least similar hand gesture meanings according to different scientific studies. However, you might ask another realistic question which is what about many people who have similar skin colors but are coming from different groups of people who have a completely different cultural background which results in different Hand Gestures and people who have similar Hands Gesture but have much different skin colors. These are all realistic Hand Gesture Recognition problems and these are the problems that Hand Gesture Recognition already solved or is going to solve in the future. Firstly, for people who have similar skin colors but are coming from different groups of people who have a completely different cultural background, this is when skin color comes to play its role. Even though those people have similar skin color, their skin color can’t be exactly the same. Most of the time, it will be either darker or lighter and we might say it’s all yellow or white, but the machine will see it as its data format so even if it is all white, the type of white is still completely different. And this is when gesture segmentation or more accurately skin color segmentation makes a difference. Overall, us human read skin colors as the simple color we have learned from different textbooks but the computer or machine see the different color in the different color spaces and the data they receive and going to process will be much more accurate. In addition to that, scientists will need to do more in-depth research and studies in order to get the most accurate result. And for people who have similar Hands Gesture but have many different skin colors, scientists will need to collect more accurate data not only about the color and about the size, angles, etc. This more detailed information will help the machine read Hand Gesture more accurately in order to get the most beneficial feedback. The background color will undoubtedly provide lots of useless information and potentially negatively influence the result. In this way, Hand Gesture Recognition has developed its own color and gesture recognition algorithm and method to remove the most useless data or color and leave the valid ones. Lighting in different background settings will have a huge influence to and in most ways, it will negatively influence the result too. There are five most important steps in Hand Gesture Recognition which are Graphic Gathering, Pretreatment, Skin Color Segmentation, Gesture Segmentation, and lastly Gesture Recognition. All these different steps are all very crucial in order to get the most accurate feedback or result. It is pretty similar to the most data treatment process especially the first two steps where you first build a model, gather different types of data, clean the data after that, and use skin color segmentation and gesture segmentation before the last Gesture Recognition process 5.
3.2 Body Gesture
3.2.1 Introduction to Body Gesture
Body gestures, which can also be called body language, refer to humans expressing their ideas through the coordinated activities of various body parts. In our lives, body language is ubiquitous. It is like a bridge for our human communication. Through body language expression, it is often easier for us to understand what the other person wants to express. At the same time, we can express ourselves better. The profession of an actor is a good example of body language. This is a compulsory course for every actor because actors can only use their performances to let us know what they are expressing 6. At this time, body language becomes extremely important. Different characters have different body movements in different situations, and actors need to make the right body language at a specific time to let the audience know their inner feelings. Yes, the most important point of body language is to convey mood through movement.
In many cases, certain actions will make people feel emotions. For us who communicate with all kinds of people every day, there are also many body languages that we are more familiar with. For example, when a person hangs his head, it means that he is unhappy, walking back and forth is a sign of a person’s anxiety, and body shaking is caused by nervousness, etc.
3.2.2 Body Gesture and Big Data
As a piece of big data, body language requires data collected by studying human movements. Scientists found that when a person wants to convey a complete message, body language accounts for half. And because body language belongs to a person’s actions subconsciously, it is rarely deceptive. All of your nonverbal behaviors—the gestures you make, your posture, your tone of voice, how many eyes contact you make—send strong messages 7. In many cases, these unconscious messages from our bodies allow the people who communicate with us to feel our intentions. Even when we stop talking, these messages will not stop. This also explains why scientists want to collect data to let artificial intelligence understand human behavior. In order for artificial intelligence to understand human mood or intention from people’s body postures and actions, scientists have collected a lot of human body actions that show intentions in different situations through research. The music gesture artificial intelligence developed by MIT-IBM Watson AI Lab is a good example 8. The music gesture artificial intelligence developed by MIT-IBM Watson AI Lab can enable artificial intelligence to judge and isolate the sounds of individual instruments through body and gesture movements. This success is undoubtedly created by the big data of the entire body and gestures. The research room collects a large number of human structure actions to provide artificial intelligence with a large amount of information so that the artificial intelligence can judge what melody the musician is playing through body gestures and key points of the face. This can improve its ability to distinguish and separate sounds when artificial intelligence listens to the entire piece of music.
Most of the artificial intelligence’s analysis of the human body requires facial expressions and body movements. This recognition cannot be achieved only by calculation. What is needed is the collection of the meaning of different body movements of the human body by a large database. The more situations are collected, the more accurate the analysis of human emotions and intentions by artificial intelligence will be. The easiest way is to include more. Just like humans, broadening your horizons is a way to better understand the world. The way of recording actions is not complicated. Just set several key movable joints of the human body to several points, and then connect the red dots with lines to get the approximate shape of the human body. At this time, the actions made by the human body will be included in the artificial intelligence. In the recording process, the upper body and lower body can be recorded separately. In order to avoid in some cases, the existence of obstructions will cause artificial intelligence to fail to recognize correctly.
3.2.3 Random Forest Algorithm in Body Gesture Recognition
Body gesture recognition is pretty useful but pretty hard to achieve because of its limitations and harsh requirements. Without the development of all kinds of 3D cameras, body gesture recognition is just an unrealistic dream. In order to get important and precise data for the body gesture recognition to process, different angles, light, background all needs to be captured 9. For body gestures, the biggest difficulty is that if you only capture data in the front, it will not give you the correct information and result in most of the time. In this way, you will need precise data from different angles. A Korean team has done an experiment using three 3D cameras and three stereo cameras to capture images and record data from different angles. The data were recorded in a large database that includes captured data both from outside and inside. One of the most popular algorithms used in body gesture recognition is the random forest algorithm. It is very famous and useful in all types of machine learning projects. It is a type of supervised learning algorithm. Because there are all types of data are needed to be a record and process. The random forest algorithm is perfect for that, the biggest advantage of this algorithm is that it can let each individual tree mainly focus on one part or one characteristic of body gesture data because of this algorithm’s ability to combine all weak classifiers into a strong one 10. It is simple but so powerful and efficient. In addition to that, it works really well with body gesture recognition. With the algorithm and advanced cameras, precise data could be collected and AIs will be able to get useful information at different levels.
3.3 Face Gesture
3.3.1 Introduction to Face Gesture (Facial Expression)
Body language is one of the ways that we can express ourselves without saying any words. It has been suggested that body language may account for between 60 to 65% of all communication 6. According to expert, body language is used every day for us to communicate with each other. During our communication, we not only use words but also use body gestures, hand gestures and most importantly, we use facial expression most. During communication with different people, our face communicate different thoughts, idea, and emotion and the reason why we use facial expression more than any other body gestures is that when we have certain emotion, it is express in our face automatically. Facial expression is often not under our control. That is why people often say that the word that come out of mouth cannot always be true, but their facial expression will reveal what those people are thinking about. So, what is facial expression exactly? According to Jason Matthew Harley, Facial expressions are configurations of different micromotor movements in the face that are used to infer a person’s discrete emotional state 9. Some example of common facial expression will be: Happiness, Sadness, Anger, Surprise, Disgust, Fear, etc 6. Each facial expression will have some deep meaning behind it. For example, A simple facial expression like smiling can be translated into a sign of approval, or it can be translated into a sign of friendly. If we put all those emotion into big data, it will help us to understand ourselves much better.
3.3.2 Sense Organs on The Face
The facial expression expresses our emotion during the communication by micro movement of our sense organs. The most used organs are the eyes and mouth and sometimes, the eyebrows.
The eyes are one of the most important communication tools in our ways of communication with each other. When we communicate with each other, the eye contact will be inevitable. The signal in your eye will tell people what you are think. Eye gaze is a sample of paying attention when communicating with others. When you are talking to a person and if his eye is directly on you and both of you keep having eye contact. In this situation, this mean that he is interested in what you say and is paying attention to what you say. On the other hand, if the action of breaking eye contact happens very frequently, it means that he is not interested, distracted, or not paying attention to you and what you are saying.
Blinking is another eye signal that is very often and will happen in communicating with other people. When talking to other people, blinking is very usual and will happen every time when you are going to communicate with different people. But the frequency of blanking can give away what are you feeling right now. People often blink more rapidly when they are feeling distressed or uncomfortable. Infrequent blinking may indicate that a person is intentionally trying to control his or her eye movements 6. For example, when A person is lying, he might try to control his blinking frequency to make other people feel like he is calm and saying the truth. In order to persuade other people that he is calm and telling the truth, he will need to blink less frequently.
Pupil size is a very important facial expression. Pupil size can be a very subtle nonverbal communication signal. While light levels in the environment control pupil dilation, sometimes emotions can also cause small changes in pupil size 6. For example, when you are surprised by something, your pupil size will become noticeably larger than before. When having a communication, dilated eyes can also mean that the person is interesting in the communication.
Mouth expression and movement will also be a huge part in communicating with other and reading body language. The easiest example will be smiling. A micro movement of your mouth and lip will give signal to others about what do you think or how are you feeling. When you tighten your lips, it means that you either distaste, disapprove or distrust other people when having a conversation. When you bite your lips, it means that you are worried, anxious, or stressed. When someone tries to hide certain emotional reaction, they tend to cover their mouth in order not to display any facial expression through lip movement. For example, when you are laughing. The simple movement of turning up or down of the lip will also indicate what a person is feeling. When the mouth is slightly turn up, it might mean that the person is either feeling happy or optimistic. On the other hand, a slightly down-turned mouth can be an indicator of sadness, disapproval, or even an outright grimace 6.
3.3.3 Facial Expression and Big Data
Nowadays, since technology is so advance, everything around us can be turn into data and everything can be related to data. Facial expression is often study by different scientist in research because it allows us to understand more about human and communication between different people. One of the relatively new and promising trends in using facial expressions to classify learners' emotions is the development and use of software programs the automate the process of coding using advanced machine learning technologies. For example, FaceReader is a commercially available facial recognition program that uses an active appearance model to model participant faces and identifies their facial expression. The program further utilizes an artificial neural network, with seven outputs to classify learner’s emotions 11. Also, facial expression can be analyzed in other software programs like the Computer Expression Recognition Toolbox. Emotion is a huge study field in the technology field, and facial expression is one of the best ways to study and analyze people’s emotion. Emotion technology is becoming huge right now and will be even more popular in the future according to MIT Technology Review, Emotion recognition – or using technology to analyze facial expressions and infer feelings-is, by one estimate, set to be a $25 billion business by 2023 12. So back to the topic about big data and facial expression. Why are those things related? It is because, first everything is data around us. Your facial expression can be stored into data for other to learn and detect too. One of the examples is that, in 2003, The US Transportation Security Administration started training humans to spot potential terrorists by reading their facial expression. And by that, scientist believe that if human can do that, with data and AI technology, robot can detect facial expression more accurate than human.
3.3.4 The Problem with Detecting Emotion for Technology Nowadays
Even though facial expression can reveal people’s emotion and what they think, but there has been “growing pushback” against the statement. A group of scientists brought together a research after reviewing more than 1,000 paper on emotion detection. After the research, the conclusion of it is hard to use facial expressions alone to accurately tell how someone is feeling is made. Human’s mind is very hard to predict. People do not always cry when they feel down and smile when they feel happy. The facial expression can not always reveal the true feeling the person is feeling. Not only that, because there is not enough data for facial expression, people will often mistakenly categorize other’s facial expression. For example, Kairos, which is a facial biometrics company, promise retailers that it can use a emotion recognition technology to figure out how their customers are feeling. But when they are labeling the data to feed the algorithm, one big problem reveals. An observer might read a facial expression as “surprised,” but without asking the original person, it is hard to know what the real emotion was 12. So the problems with technology that involves around facial expression are first, there is not enough data. Second is that facial expression sometimes can not be always true.
3.3.5 Classification Algorithms
Nowadays, since technology is growing so fast, there are a lot of interaction between humans and computer. Facial expression plays an essential role in social interaction with other people. It is not arguably one of the best ways to understand human. “It is reported that facial expression constitutes 55% of the effect of a communicated message while language and voice constitute 7% and 38% respectively. With the rapid development of computer vision and artificial intelligence, facial expression recognition becomes the key technology of advanced human computer interaction 13.” This quote from the research shows that facial expression is one of the main tools that we are using to communicate with other people and interact with computer. So being able to recognize and identify the facial expression becomes relatively important. The main objective for facial expression recognitions is to use its conveying information automatically and correctly. As a result, feature extraction is very important to the facial expression recognition process. The process needs to be smooth and without any mistakes. So, algorithms are needed in the processes. Classification analysis is an important component of facial recognition, it is mainly used to find data distribution that is valuable and at the same time, find data models in the potential data. At present it has further study of the database, data mining, statistics, and other fields 13. In addition to that, one of the major obstacles and limitation of facial expression recognition is face detection. To detect the face, you will need to locate the faces in an image or a photograph. This is where scientists applicate classification algorithm, machine learning and deep learning. Recently, convolutional neural network model has become so successful that facial recognition is the next top wave 14.
With the development of artificial intelligence, human-computer interaction, big data, and machine learning even deep learning is getting more mature. Gesture Recognition including Hand Gesture Recognition, Body Gesture Recognition, and Face Gesture Recognition has finally come true into a real-life application and already achieved huge success in many areas. But it still has much more potential in all possible areas that could change people’s lives drastically in a good way. Gestures are the simplest and the most natural language of all human beings. It sends the clearest message for communicating between people, and even human and computers. Because of the more powerful cameras, better big data technology, and more efficient and effective algorithms from deep learning, Scientists are able to use color and the Gesture Segmentation method to remove useless color data in order to maximize the accuracy of the result. As we are doing our research, we also find out Hand Gesture Recognition is not the only Recognition in this area, Body Gesture Recognition and Face Gesture Recognition or facial expression are also very important, they can also deliver messages in the simplest way. They are also very effective when building relationships between humans and machines. Face Gesture or facial expression could not only deliver messages but even deliver emotions. Micromovements of facial expressions studied by different scientists could be very useful in predicting the emotions of humans. Body Gesture Recognition is also helpful as we did our research with the body gesture data scientists collected from different musicians with different instruments. They are able to predict the melodies or even the songs played by that musician. This is mind-blowing because with this type of technology and applications we are able to achieve more and use it in many possible fields. With all these combined, scientists could build a very successful and mature Gesture Recognition model to get the most accurate result or prediction. According to the research and our own analysis, we come up with a conclusion that Gesture Recognition will be the next hot trendy topic and are applicable in many possible areas including Security, AI, economics, manufacture, the game industry, and even medical services. With Gesture Recognition being applied, scientists are able to develop much smarter AIs and machines that can interact with humans more efficiently and more fluently. AIs will be able to receive and understand messages from humans more easily and will able to function better. This is also a great message for many handicapped people. With Hand Gesture Recognition being used, their life will also be easier and happier and that’s definitely something we are want to see because the overall goal of all the technologies is to make people’s life easier and bring the greatest amount of happiness to the greatest amount of people. However, the technology we have right now is not advanced enough yet, in order to get a more accurate result, we still need to develop better cameras, better algorithms, and better models. But we all believe that this era is the big data era, and everything could happen as big data and deep learning technology get more and more advanced and mature. We believe in and look forward to the beautiful future of Gesture Recognition. And we also think people should really pay more and closer attention to this field since Gesture Recognition is the next wave.
Srilatha, Poluka, and Tiruveedhula Saranya. “Advancements in Gesture Recognition Technology.” IOSR Journal of VLSI and Signal Processing, vol. 4, no. 4, 2014, pp. 01–07, iosrjournals.org/iosr-jvlsi/papers/vol4-issue4/Version-1/A04410107.pdf, 10.9790/4200-04410107. Accessed 25 Oct. 2020. ↩︎
Bazarevdsky, V., & Zhang, F. (2019, August 19). On-device, real-time hand tracking with MediaPipe. Google AI Blog. https://ai.googleblog.com/2019/08/on-device-real-time-hand-tracking-with.html ↩︎
F. Zhan, “Hand Gesture Recognition with Convolution Neural Networks,” 2019 IEEE 20th International Conference on Information Reuse and Integration for Data Science (IRI), Los Angeles, CA, USA, 2019, pp. 295-298, doi: 10.1109/IRI.2019.00054. ↩︎
Di Zhang, DZ.(2019) Research on Hand Gesture Recognition Technology Based on Machine Learning, Nanjing University of Posts and Telecommunications. ↩︎
A. Choudhury, A. K. Talukdar and K. K. Sarma, “A novel hand segmentation method for multiple-hand gesture recognition system under complex background,” 2014 International Conference on Signal Processing and Integrated Networks (SPIN), Noida, 2014, pp. 136-140, doi: 10.1109/SPIN.2014.6776936. ↩︎
Cherry, K. (2019, September 28). How to Read Body Language and Facial Expressions. Verywell Mind. Retrieved November 8, 2020, from https://www.verywellmind.com/understand-body-language-and-facial-expressions-4147228 ↩︎
Segal, J., Smith, M., Robinson, L., & Boose, G. (2020, October). Nonverbal Communication and Body Language. HelpGuide.org. https://www.helpguide.org/articles/relationships-communication/nonverbal-communication.htm ↩︎
Martineau, K. (2020, June 25). Identifying a melody by studying a musician’s body language. MIT News | Massachusetts Institute of Technology. https://news.mit.edu/2020/music-gesture-artificial-intelligence-identifies-melody-by-musician-body-language-0625 ↩︎
N. Normani et al., “A machine learning approach for gesture recognition with a lensless smart sensor system,” 2018 IEEE 15th International Conference on Wearable and Implantable Body Sensor Networks (BSN), Las Vegas, NV, 2018, pp. 136-139, doi: 10.1109/BSN.2018.8329677. ↩︎
Bon-Woo Hwang, Sungmin Kim and Seong-Whan Lee, “A full-body gesture database for automatic gesture recognition,” 7th International Conference on Automatic Face and Gesture Recognition (FGR06), Southampton, 2006, pp. 243-248, doi: 10.1109/FGR.2006.8. ↩︎
Harley, J. M. (2016). Facial Expression. ScienceDirect. https://www.sciencedirect.com/topics/computer-science/facial-expression ↩︎
Chen, A. (2019, July 26). Computers can’t tell if you’re happy when you smile. MIT Technology Review. https://www.technologyreview.com/2019/07/26/238782/emotion-recognition-technology-artifical-intelligence-inaccurate-psychology/ ↩︎
Ou, J. (2012). Classification algorithms research on facial expression recognition. Retrieved from https://www.sciencedirect.com/science/article/pii/S1875389212006438 ↩︎
Brownlee, J. (2020, August 24). How to perform face detection with deep learning. Retrieved from https://machinelearningmastery.com/how-to-perform-face-detection-with-classical-and-deep-learning-methods-in-python-with-keras/ ↩︎