An Efficient Algorithm for Lip Segmentation in Color Face Images Based on Local Information

Lip detection is used in many applications such as face detection and lips reading. In previous works, researchers have considered whole of face image for lip detection. In this paper we propose a new algorithm. In our algorithm for reducing required calculation and increase accuracy of correct detection, we do not consider whole of the face image. We first remove the upper half part of the face image. Then, for estimate lip area, we divide remained lower half face image to equal parts. For each part we calculate statistical information such as standard deviation, and based on them we detect lip area in face image. For separate lip pixels from skin pixels, we use YCbCr and HSI color spaces at this work. We evaluate our work on CVL face database. Our experiments show that new algorithm gives better results than previous works on this database.