摘要
在人脸识别中,基于PCA和粗糙集的联合自寻优特征选择算法首先利用PCA对人脸图像进行人脸特征的提取和约简,在此基础上利用粗糙集的自寻优约简方法进一步得到能充分体现人脸信息的最小人脸特征集合。实验证明此算法应用于人脸识别中不仅大大减少了特征的数量以及分类过程中的运算量,还有效的提高了人脸识别的正确率和减少了识别时间,并且对于一定范围内的不同取样的训练具有一定的稳定性。
In face recognition, firstly, the union adaptive algorithm based on PCA and rough set for feature selection extracts and selects the face features from the face images by PCA. Then we use the adaptive searching optimal reduct method based on rough sets to obtain the minimum set of face features which can describe the information of face completely. The experiment indicates that such algorithm not only can decrease the number of features and the computation of classification greatly, but also can increase the accuracy and decrease the time for human face recognition. Meanwhile, it is stable for training different samples in some area.
出处
《佛山科学技术学院学报（自然科学版）》
CAS
2007年第1期27-31,共5页
Journal of Foshan University(Natural Science Edition)
