基于特征加权模块双方向二维主成分分析的人脸识别  

Face recognition based on eigen weighted modular two-directional two-dimensional PCA

作　　者：顾播宇[1] 孙俊喜[1] 李洪祚[1] 刘红喜[1] 刘广文[1] 

机构地区：[1]长春理工大学电子信息工程学院,长春130022

出　　处：《吉林大学学报（工学版）》2014年第3期828-833,共6页Journal of Jilin University:Engineering and Technology Edition

基　　金：国家自然科学基金项目(60977052);吉林省科技发展计划项目(20100312)

摘　　要：对模块双方向二维主成分分析(Two-directional two-dimensional principal component analysis,(2D)2PCA)人脸识别算法进行了改进,提出了基于特征加权模块(2D)2PCA的人脸识别算法。首先对图像进行分块及(2D)2PCA特征提取。然后依据每一子图像块的特征对识别的贡献程度分配特征权重。最后对测试样本采用加权距离进行最近邻分类。该算法无需任何先验知识,依据子图像块在特征空间中的信息比重确定其贡献程度,从而实现自适应权重分配。试验结果表明:本文算法能够有效地提高人脸识别的正确率。An improved modular Two-directional Two-dimensional Principle Component Analysis （2D （PCA）） with eigen weight for face recognition is proposed. First, the image is divided into sub-blocks and the features are extracted by modular 2D（PCA）. Then, the feature weight is assigned to each sub-block of every image according to the contribution for recognition of the sub-block. Finally, the testing samples are classified by nearest neighborhood classification of weighted distance. The contribution of each sub-block is determined self-adaptively according to the proportion of local feature information in eigen space. This algorithm does not need any prior knowledge. Experimental results show that the recognition rate of the proposed algorithm is effectively improved.

关 键 词：信息处理技术 模式识别 人脸识别 主成分分析 特征加权 特征提取 

分 类 号：TN911.73[电子电信—通信与信息系统]