Multi-pose Face Detection Based on Adaptive Skin Color and Structure Model

Abstract

In this paper, propose a method of face detection based on adaptive skin color and structure model for multi-pose color images in a complex background. First, a face skin color model is built, and an adaptive Gamma correction is used in the skin color segmentation, which improve the skin color model to adapt to different illumination and multi-poses of face. Then a triangular structure model is constructed specific to the face features. And finally use the structure model to locate the face precisely on the basis of face images of skin color. The experimental results show the approach raised on this paper can adapt to the face detection under the conditions of different illumination environment, various poses and complex background, and the algorithm is quick and effective.