面部表情识别若干关键问题研究-计算机应用专业毕业论文.docxVIP

面部表情识别若干关键问题研究国关键词：面部表情识别，面部动作单元，交互式遗传算法，支持向量机，绝对评 面部表情识别若干关键问题研究国 关键词：面部表情识别，面部动作单元，交互式遗传算法，支持向量机，绝对评 价值，基于意识模块的神经网络，情感 本论文受国家自然科学基金项目(No资助。 AbstractWith Abstract With the increasing development and prevalence of information technology,the application of computers has spread to all aspects of human life In order to utilize computer resources more deeply and widely,we should endue computers with human wisdom．Human intelligence includes not only the ability of rational thinking and logical reasoning，but also that of emotion．In multimodal human—computer interaction，the theory and technique of understanding and recognizing human’8 emotion key points to establish natural and harmonious environment between humml and computer Faces play the primary role in human communication，as they can present varions affective information．Therefore，facial expression recognition is importaat in the research of multimodal human]一computcr interaction． The study in this thesis includes the following aspects： 1．In this thesis，we study the methods of facial action units recognition．We develop a recognition approach in which the hrteraetive Genetic Algorithm(IGA)is adopted．This approach adds user’S ability in comparison into retrieval process，and therefore OVCECOn*Cs mauy strict requests to facial images needed by classic ways It can retrieve the most matched sample image for testing image with satisfactory recognition rate 2．In order to reduce the user’S fatigue，which is one of the main problems of IGA 面部表情识别若干关键问题研究two 面部表情识别若干关键问题研究 two improved algorithms are developed IGA with Support Vector Machine，and IGA with absolute rating．The aim of these two algorithms is to accelerate convergence There sinmlative and subjective experiments to prove the efficiency of these methods 3．In this thesis，we apply Taylor’S emotion recognition neural network based on attention to the facial expression recognition in order to make the expression recognition algorithm have more[eatures of brain st