摘要：In recent years the task of ecognizing human face with the help of a machine has acquired a significant attention among researchers. A wide range of commercial and law enforcement applications largely accounts for this developing trend. Although current face ecognition systems have reached a certain maturity level they are still far away from the capability of human perceptual system. In a 2D image the unique facial appearance normally consists of intensity curvatures which may be incorporated for ecognizing human face. This work inestigates the use of weighted distance transform to bring improvements in face ecognition. The weighted distance transform takes into account both the spatial distance among pixels and the local intensity variations. A standard face dataset is used to alidate the proposed method. The obtained results are comparable to the state-of-the-art face recognition techniques.
