基于非线性降维算法的人脸识别

he technique of human face recognition is a challengeable subject in the field of pattern recognition and machine vision. This paper mainly focused on its core part—feature the extraction and human face recognition algorithms. The paper collected and analyzed mass of recent internal and foreign papers and literatures about human face recognition. According to the understanding to the relevant papers and literatures, I implemented a complete linear human face recognition algorithm base on PCA and a non-linear human face recognition algorithm base on ISOMAP. We tested our algorithms on the UMIST dataset. Finally, the necessity and validity are confirmed via data visualization technique.