关于二维主成分分析方法的研究

The principal component analysis (PCA), or the eigenfaces method, is a de facto standard in human face recognition. Numerous algorithms tried to generalize PCA in different aspects. More recently, a technique called two-dimensional PCA (2DPCA) was proposed to cut the computational cost of the standard PCA. Unlike PCA that treats images as vectors, 2DPCA views an image as a matrix. With a properly defined criterion, 2DPCA results in an eigenvalue problem which has a much lower dimensionality than that of PCA. In this paper, we show that 2DPCA is equivalent to a special case of an existing feature extraction method, i.e., the block-based PCA. Using the FERET database, extensive experimental results demonstrate that block-based PCA outperforms PCA on datasets that consist of relatively simple images for recognition, while PCA is more robust than 2DPCA in harder situations.

Is Two-dimensional PCA a New Technique?

The principal component analysis(PCA),or the eigenfaces method,is a de facto standard in human face recognition.Numerous algorithms tried to generalize PCA in different aspects.More recently,a technique called two-dimensional PCA(2DPCA)was proposed to cut the computational cost of the standard PCA.Unlike PCA that treats images as vectors,2DPCA views an image as a matrix.With a properly defined criterion,2DPCA results in an eigenvalue problem which has a much lower dimensionality than that of PCA.In this paper,we show that 2DPCA is equivalent to a special case of an existing feature extraction method,i.e.,the block-based PCA.Using the FERET database,extensive experimental results demonstrate that block-based PCA outperforms PCA on datasets that consist of relatively simple images for recognition,while PCA is more robust than 2DPCA in harder situations.