Abstract: Aimed at image feature extraction for field application of Principal Component Analysis（PCA） and Two-Dimensional Principal Component Analysis（2DPCA）, this paper uses the eigenvalue to solve the associated two-dimensional sample membership, and applies the relevant eigenvalue classification methods. This method combines two-dimensional eigenvalue, in feature extraction for face reconstruction image, and has advantages of fast stability and local features clear. By introducing the matrix inner product features and classification of 2DPCA to compare the results, the experiments show that, in the ORL and Yale database using the new classification method to identify the method with excellent effects.