A Skin Color Model Based on Modified GLHS Space for Face Detection

Face detection is one of the most useful research topics in computer vision. In the past several decades, there has been noticeable and successful achievement in face detection. However, it is still a challenge to obtain both low computational complexity and high detection efficiency. The main problem in existing skin color based face detection algorithms is that different illumination conditions will lead to different clustering results of the skin colored pixels. In this paper, we propose a new skin color model based on the modified GLHS (generalized LHS) color space to solve this problem. Experimental results show that our model can detect skin pixels with high accuracy under various illumination conditions and backgrounds.