基于人脑情感注意回路的表情识别  

Expression Recognition Based on Human Emotion-Attention Circuit

作　　者：王上飞[1] 薛佳[1] 王煦法[1] 

机构地区：[1]中国科学技术大学计算机科学与技术系,合肥230027

出　　处：《模式识别与人工智能》2008年第5期598-602,共5页Pattern Recognition and Artificial Intelligence

基　　金：安徽省自然科学基金资助项目(No.070412056)

摘　　要：表情识别是自然人机交互中的重要环节之一.目前表情识别中的分类器主要采用机器学习的相关技术和方法,取得了一定的研究成果.而人脸表情识别被认为是人类视觉中独特的过程,因此本文提出采用模拟人脑认知情感结构的扩展神经网络方法进行表情识别,在隐层节点中体现基于注意的反馈控制.在 Cohn Kanade 表情数据库上进行实验验证,实验结果证明该算法的有效性.Expression recognition is a key issue in natural human-computer interaction. The machine leaning method is most commonly used to recognize expression from face images or image sequences, and good resuhs are acquired. However, expression recognition is considered as a special process of human vision and it may provide a new way to realize emotion recognition in computer by simulating human emotional information process. The expression recognition using extended neural networks is presented. It simulates the structure of emotion-attention circuit in human brain by adding feedback control for attention in hidden layer. The effectiveness of the proposed approach is evaluated on Cohn Kanade expression database, and good results are obtained.

关 键 词：表情识别 情感注意回路 扩展神经网络 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]