| Emotion is an important component of human mind, which is a response to the objective world, including organization function, adaptation function, signal function and motivation function. Facial emotion, which is an external manifestation of human emotion, plays a very important role in daily communication. Quickly and accurately on face emotion recognition will help people receive information and communicate effectively in social life, which is an important source of information.According to many domestic and foreign researches on facial emotion, relative to low trait-anxiety, high trait-anxiety individuals have a degree of distortion on facial emotion recognition, especially on fear, anger, and other emotions. Therefore, there is a close relationship between facial emotion recognition and personal trait-anxiety. What about its relevance? How about the relationship? Is the accuracy of facial emotion recognition able to correctly reflect the level of trait-anxiety? Therefore, in this research, the author discusses the facial emotion recognition features of low and high trait-anxiety groups on7species facial emotions:anger, sad, happy, fear, surprise, disgust, and neutral. Through support vector regression (SVR) method, the research illustrates the mapping relation between facial emotion recognition and trait-anxiety, explores facial emotion recognition on forecast effect of trait-anxiety, and makes comparison with artificial neural network (ANN). The results in this research are as follows:(1) There are significant differences on facial emotion recognitions of fear, anger, neutral, happy, surprise between low and high trait-anxiety groups, especially extremely significant differences on fear, anger and neutral.(2) Through support vector regression (SVR), the mapping relationship is established between facial emotion recognition and trait-anxiety, and it verifies the prediction of facial emotion recognition on trait-anxiety.(3) Comparing with artificial neural network (ANN), the prediction of support vector regression (SVR) is much more effective.According to the results of the research, the author finds that different levels of trait-anxiety individuals have different features on facial emotion recognition, and support vector regression (SVR) machine can play an important role in psychological testing. |
