Research Of Face Recognition Algorithm Based On Active Shape Model

Face recognition technology which is to implement the personal identification or target pursuit through the characteristic information extraction by analysis of face image. There are three steps for this technology:Face Detection, Feature Extraction and Face Recognition and the robustness, efficiency and accuracy is the judging standard of this technology.The Active Shape Mode (ASM) proposed by Cooter which apply the global shape model to match the face initial shape and to match the contour edge in high precision based on the local gray model. The week of this model is that the accuracy of the result could be affected by the position of the initial shape and the interference factor in the contour edge.Above on. this paper is focus on the feature matching algorithm based on the Scale-invariant Feature Transform (SIFT). As the SIFT algorithm has the less affected by the illumination, revolve interference factor and could extract the SIFT feature in multi scale spatial more accurate, we extract the SIFT feature in different conditions and get the encouraging result. The SIFT feature vector for the ASM key feature point are proposed to realize the face contour precise match. The characteristics of the Gabor feature are also studied and apply it to replace the local gray information to reduce the interference factor and increase the accuracy of the face match.