Research On Algorithm Of Face Detection Based On Skin-Color Detection Model

It is important that face detection in procession of automatic face recognition system. Face detection finds face in the image and it’s segmented from background. This paper presents that research on the face detection algorithm, which used students’information and applied in the identity authentication, network game, image search and other popular areas.Now the international on face detection algorithm is mainly including the following aspects:based on the characteristics of the method, based on the template matchingmethod, the detection method based on knowledge and the method based on facial appearance characteristics.When the image of a scene or face itself is more complex, in addition detect different rules using set parameters will be different, the practical difficulties of above method will be more independent when used in. According to the characteristics of facial key features, this article proposes a segmentation method of color that combines the characteristics of the mouth and skin, then combined it with Adaboost algorithm to research for human face detection algorithm.This article mainly the following circumstances:(1) Building Color space by collecting data information.Introduce the principle of Adaboost algorithm,shin segmentation is used to detecting candidate face region.(2) Improving the algorithm, according to the mouth gray between different whit skins, the ideal method of mouth segmentation is the segmentation algorithm based on gray scale. (3) Using constrained proportion about the mouth segmentation to location neonatal face.Experiments show that,this method shows more performance in detection,and has the advantages of low false detection rate of the Adaboost algorithm, it can be effectively used in multipose and complex background.The robustness of the importance is self-evident.