面向人机交互的人脸表情识别技术(英文版)

Chapter 1 Introduction 1.1 Research Background and Significance In social life， emotion plays an important role in coordinating the relation- ship between people.The research of emotion is more and more concerned by the society. Facial expression is a unique form for human beings to convey their intentions and emotional states. In interpersonal communica-tion， facial expression is a very important way of communication besides voice. Psychological studies indicate that whether a listener likes your speech in human communication， 7% depends on your vocabulary，38% depends on your voice， and the remaining 55% depends on your facial expressions[1]. Facial expressions， as the carrier of information， reflect much human behavior and convey the hidden information that the voice cannot do[2-3]. Facial expressions are not only a natural way for people to express emotions， but also the main symbol for people to identify emotions， which plays a very important role in affective computing. For human-computer interaction to achieve natural and harmonious communication， it is neces- sary for computers to acquire and analyze facial expression information. Therefore， the research on facial expression recognition method needs to be further developed. The computer vision technology is noninvasive， passive， cheap， and natural， so it naturally becomes the key technology in the process of facial expression recognition. While the in-depth research on such technologies as computer vision， image analysis and pattern recognition can contribute to the advancement of facial expression recognition methods. At the same time， the study of facial expression recognition incorporates many other disciplines such as pattern recognition，image processing，analysis and understanding， computer graphics， cognitive science， neural computing， physiology， psychology， and so on. Therefore， the research on this issue could greatly enhance the development of the above-mentioned disciplines. What’s more， the technology of facial expression recognition is also the core subject of the research of the intelligent human-computer interface that essentially is to train the machine to capture the changes of human emotions. With the advances in science， the computer after being trained can to a certain extent mimic the human behavior， but still does not have the human feelings. However， with the development of the facial expression recogni- tion technology， the interface can be more natural to a certain extent so that the computer can respond to humans in accordance with the changes of human emotion. From this point， it can change people’s way of life to some extent. In other words， the observation of the facial expression can be used to understand psychological states. This is， as the saying goes， to take the cue from a person’s words and facial expressions. Nevertheless， the naked eyes cannot find some very fine changes of facial expressions， or it is easily to draw an opposite conclusion from that fine changes. After the introduc- tion of machine recognition of facial expressions， the results can be obtained objectively and in real time. Therefore， we can expect facial expression recognition to replace lie detector in the security sector in the future. In terms of medical care， for some mental patients， the expression analysis can be used as an auxiliary treatment. The analysis of patient’s expression can help doctors analyze the patient’s mental state， predict what would happen when the patient is in the face of diverse outside stimuli and thus take some responding precautionary measures， so that many incidences of accidents can be avoided. Facial expressions can also be used to supple-ment the lip language to help those having hearing problem communicate with the world they live in. On the commercial side， the devices that can keep track of user’s facial expressions will certainly come into being in the near future. The applica- tion of such technology in videophone or teleconferencing is undoubtedly of great significance for strengthening domestic and international business cooperation. If the technology is applied to the automotive， the machine vision equipment can detect the driver’s fatigued state and then alert the drive. This equipment can analyze the real-time mental state of the driver and judge whether the driver’s fatigue will exert negative effect on the driver or the fatigue will make the driver sleep while driving， it then will take the necessary measures such as alarming or playing exciting music to prevent potential traffic accidents. In addition， if the existing recreational machines that have no emotions can recognize the human emotions， and adjust the strategy in accordance with the users’emotional changes， it can provide individualized services and greatly enhance the user’s entertainment interest. Moreover， an intelligent toy with personality and the ability to understand and express emotions (such as an electr