基于稀疏差分和典型相关分析融合的人脸图像识别

Face Image Recognition Based on the Integration of Sparse Difference and Canonical Correlation Analysis

为了进一步提高人脸图像的识别率,提出了一种相关分析融合的人脸图像识别算法。首先通过采用压缩测量数据得到人脸图像,然后通过划分子模式的方法去除图像中的小样本,提取局部特征,采用成分分析提取人脸图像的整体特征,通过算法对人脸图像特征进行融合,从而消除人脸部冗余信息,最后通过3个人脸数据集对算法进行测试。仿真实验表明,本文的算法相对于参比算法,提高了人脸图像识别精度,具有很好的鲁棒性。

In order to further improve the accuracy of face image recognition, in this paper a face image recognition algorithm based on the integration of correlation analysis is presented. First, a face image is obtained by using compression measurement data, and then the small samples in the image can be eliminated by dividing subschema, to extract local features. By using component analysis, the overall characteristics of the face image are extracted, and through the algorithm, the characteristics of the face image are integrated, to eliminate the redundant information of the face image. Finally, the algorithm is tested by using three face data sets. The simulation experiments show that compared to the reference algorithms, the algorithm presented in this paper has improved the precision of face image recognition and has excellent robustness.