Improved Chin Fitting Algorithm Based on An Adaptive Snake

Fitting of the chin is an essential part of facial model fitting systems. Many previous approaches to chin fitting have used active contour models or snakes. However, achieving successful chin fitting for a wide range of faces under varying lighting conditions has not proved straightforward. This paper describes an approach to chin fitting based on active contour models which is capable of operating successfully under unbalanced lighting conditions and for a wide range of faces. The system uses a fast adaptive algorithm which adjusts the snake parameters according to image characteristics. Results are presented for chin fitting on a large database of faces.