1.最流行的方法：主成分分析(Principal Component Analisis,PCA)和線性判別分析(Linear Discriminant Analysis,LDA)
相關(guān)論文：PCA ,Face Recognition using Eigenfaces LDA,Based on an optimized LDA algorithm for face recognition
2.流形學(xué)習(xí)算法: 等距離映射(Isometric mapping,Isomap),局部線性嵌入(locally linear embedding, LLE),拉普拉斯特征映射(laplacian eigenmap)和局部保持投影(Locality Preserving Projections,LPP)等
相關(guān)論文: Isomap, global geometric framework for nonlinear dimensionality reduction LLE, Nonlinear dimentionality reduction by locally linear embedding. laplacian eigenmap, Laplacian eigenmaps for dimensionality reduction and data representation . LPP, Learning a locality discriminanting projection for classification.
