一种融合模块2DPCA与PCA的人脸识别方法
A method for face recognition by fusing modular 2DPCA with PCA
-
摘要: 针对主成分分析(PCA)求解高阶矩阵计算量很大和模块二维主成分分析(M2DPCA)特征数量仍然较大且有一定的相关性的问题,提出了融合模块2DPCA与PCA的方法进行人脸识别.该方法先通过M2DPCA对子图像进行特征提取,然后把每个图像中的子图像按分块的顺序重新组成新的矩阵,再对新的矩阵进行PCA.在ORL人脸库中实验,结果表明,该算法在一定程度上去除了特征参数间的相关性并大大减少了特征维数.Abstract: Aiming at the problem that principal component analysis(PCA) leads to a large amount of calculation in solving high rank-matrix and the modular two-dimensional principle component analysis(2DPCA) is still large in feature calculation and a certain correlation still exists in feature extraction, a method fusing the Modular 2DPCA with PCA was put forward. The method extracted feature from sub-image using M2DPCA, and re-formed a new matrix according to the order of sub-images of each image, then PCA was carried out on the new matrix. The experimental results in ORL human face database showed that the correlation among feature parameters was removed to a certain extent and it also greatly reduced the dimension of features.
-
点击查看大图
计量
- PDF下载量: 28
- 文章访问数: 693
- 引证文献数: 0
下载:
