基于智能视觉动态人脸跟踪.docVIP

基于智能视觉动态人脸跟踪 　　摘 要： 传统方法中对动态人脸识别采用的是单演局部主方向编码识别，通过分块子模式的加权融合进行人脸特征提取，因为人脸表情和姿态变化会导致识别结果出现误差。在智能视觉模式下，提出一种基于信息熵子模式主成分分析的动态人脸跟踪识别方法。基于特征状态空间重构方法，将人脸图像分成大小相等的子模块，对子模块进行信息熵特征提取，采用主成分分析方法进行人脸特征分类。仿真结果表明，采用该算法进行动态人脸跟踪识别，能有效实现人脸表情动态跟踪，人脸识别性能较好、精度较高，性能优于传统算法。 　　关键词： 人脸识别； 智能视觉； 主成分分析； 信息熵 　　中图分类号： TN911?34； TP391 文献标识码： A 文章编号： 1004?373X（2015）24?0012?04 　　Dynamic face tracking based on intelligent vision 　　HAO Junshou1， DING Yanhui2 　　（1. Office of Teaching Affairs， Inner Mongolia Electronic Information Vocational Technical College， Hohhot 010070， China； 　　2. Department of Digital Media and Art， Inner Mongolia Electronic Information Vocational Technical College， Hohhot 010070， China） 　　Abstract： The face recognition and dynamic monitoring are realized by dynamic face tracking. The coding recognition method of monogenic local principal direction is adopted to recognize the dynamic face in traditional method. Because the facial expression and pose variation can cause error existed in recognition result in the process of facial feature extraction by means of weight fusion to the block sub?mode. a recognition method of dynamic face tracking based on principal component analysis of comentropy sub?mode is proposed in the mode of intelligent vision. Based on the method of feature state space reconstruction， the face image is divided into sub?modules with equal size， and then the comentropy feature is extracted from sub?modules， finally the principal component analysis method is used to classify the face features. The simulation results show that using this algorithm to track and recognize the dynamic face can effectively realize dynamic tracking of facial expression， and this algorithm has exellent face recognition performance and high precision. The performance of the algorithm is better than that of traditional algorithm. 　　Keywords： face recognition； intelligent vision； principal component analysis； comentropy 　　0 引 言 　　随着计算机智能视觉及图像处理技术的发展，基于智能视觉的图像处理技术广泛应用于模式识别、故障诊