【摘要】 提出一种离散余弦变换和改进的分块二维主元分析相结合的人脸识别方法。该算法利用DCT压缩人脸图像以去掉人眼不敏感的中频分量与高频分量,这样有效降低所需特征的维数,减少计算量。通过IM2DPCA进行特征提取得到人脸识别特征,运用最近邻分类器完成人脸的识别。在基于ORL、YaleB、CAS-PEAL及Feret人脸数据库的实验结果证明该算法的有效性与稳健性。更多还原

【Abstract】 Puts forward a new face recognition method based on the combination of Discrete Cosine Transform (DCT) and improved modular Two Dimensional Principal Component Analysis(2DP-CA). Processes a face image by using DCT to take out the sensitive intermediate and high frequency components and reduce the character dimensions and the computation effectively,and the image by using IM2DPCA to obtain the face recognition features.The Nearest Neighbor(NN) classifier is selected to perform face recognition.The experimental results on ORL,YaleB,CAS-PEAL and Feret face database show that this method is robust and efficient in the face recogni-更多还原