基于完全二维对称主成分分析的人脸识别

Face Recognition Based on Complete 2D Symmetrical PCA

镜像对称性是人脸的一个直观明显的自然特性,结合该特性在完全二维主成分分析的基础上提出完全二维对称主成分分析的人脸识别方法。该方法通过镜像变换得到奇对称样本和偶对称样本,分别对奇偶对称样本进行完全二维主成分分析,通过奇偶加权因子对奇偶对称样本的特征矩阵进行组合,并采用最近邻距离分类器分类。在ORL人脸数据库上的实验表明,该方法有较好的识别效果。

As facial symmetry is a natural characteristic of face images, this paper proposes a face recognition method based on complete two-dimensional PCA, and the Complete 2D Symmetrical PCA（C2DSPCA）. The odd symmetry samples and the even symmetry samples are received by mirror transforming. Odd/even symmetrical samples＇ eigen matrixes are separately extracted through the Completely 2D PCA（C2DPCA） and used to form the features by an odd-even weighted factor. A nearest neighbor classifier is employed to classify the extracted features. The method is evaluated on the ORL face image database. Experimental results show the proposed method achieves better performance.