A Method for Detecting Face Pigment Regions Based on Improved LBF

Abstract

The traditional method which is prone to complex background interference cannot detect and identify various pigment regions. Beside that the detection effect is inaccurate and the efficiency is low. To solve these problems, the LBF segmentation model with image gradient information is proposed to integrate local information of image and global information of image gradient, and to construct an energy function containing image gradient information. The proposal makes the level set function converge to diversified pigment region boundaries in the evolutionary process. In the experiment of this thesis, firstly, the face image is segmented to acquire the local face data set. Then, the local face image set is enhanced through the Single-scale Retinex algorithm. Finally, the LBF segmentation model incorporating image gradient information is applied to detect the pigment regions.