A multi-source data face recognition algorithm

The recognition rate decrease rapidly when expression changes or an angle exits in face recognition.In order to solve this problem, we proposed a multi-source data recognition algorithm based on two-dimensional principal component analysis (2DPCA). By extracting the feature of the front, left side and right side face, we get three principal component matrices, and then select some principal component vectors from them to compose a new matrix. At last we use the Nearest Neighbor Classifier to do the recognition process based on the matrix.The experimental results on ORL and CAS-PEAL Face Database indicate that this method can achieve a better recognition rate.