A video-based face verification method is proposed.Stereo vision is used to roughly segment face area from its background

and then multiple-related template matching method is used to locate the face area finely.The face organs are extracted from the located face area and accordingly the face sample is cut out.Face samples are collected from video sequence

and a support vector machine classifier is trained on this sample set against other face samples as his verifier.Experimental results demonstrate the feasibility and robustness of this algorithm in complex real environment.