Face boundary extraction by statistical constraint active contour model

Facial feature detection is very important in the face recognition literature. We propose a facial boundary extraction method, which includes two steps. Firstly, the facial geometry feature, such as eye, mouth, nose is detected by the eigenfeature method. Secondly, under the help of the geometry feature, the face boundary can be extracted by the statistical constraint active contour model (SCACM) and repeating initialization. The experiment shows the proposed method is effective.