The Study of Face Recognition Based on the Large Samples

A improved face recognition method based on the large samples was proposed in the paper. The face and the eyes were detected using the Adaboost algorithm; and according to the distance and angle of the eyes, the accurate face could be gotten. For improving the robustness of face recognition against illumination, the face image was processed by the Retinex theory. The face feature was extracted using the Gabor wavelet. The similarity of the faces was detected by the cosine and distance of face feature vectors. The validity of the algorithm is testified by the experiments.