基于加权平均脸的二维PCA人脸识别算法研究
首发时间:2008-05-14
摘要:提出了一种基于加权平均脸和二维主成分分析(PCA)特征空间相结合的人脸识别算法。通过引入样本类间和类内加权平均脸,二维PCA既使得类间散布特征矩阵最大化又减小了类内样本的差异,该方法较传统PCA和二维PCA方法具有更高的识别效果。最后通过ORL人脸图像库的试验结果,分析了加权因子对样本识别效果的影响并验证了本文方法的有效性。
For information in English, please click here
Research on Face Recognition Method Based on Two-dimensional PCA with Weighted Average Face Images
Abstract:A method based on weighted average face images combined with Two-Dimensional Principal Component Analysis (PCA) feature space is proposed for face recognition. By introducing samples among-class and within-class weighted average face images, Two-Dimensional Principal Component can not only maximize the scattering matrix of among-class but also reduce the differences of with-in class samples which have better recognition effect than the traditional PCA and Two-Dimensional PCA methods. At last, the effect weighted parameter make is analyzed on the recognition ratio and the extensive experiments based on ORL face database verify the effectiveness of the proposed method.
Keywords: Face recognition Principal Component Analysis feature extraction image processing
基金:
论文图表:
引用
No.2139623278412107****
同行评议
共计0人参与
勘误表
基于加权平均脸的二维PCA人脸识别算法研究
评论
全部评论0/1000