基于积分图和粒子群优化的肤色分割  

Skin segmentation based on integral image and particle swarm optimization

作　　者：朱志亮[1] 刘富国[1] 陶向阳[1,2] 刘晓山[1,2] 

机构地区：[1]江西师范大学物理与通信电子学院,南昌330022 [2]江西省光电子与通信重点实验室,南昌330022

出　　处：《计算机工程与应用》2014年第21期171-174,237,共5页Computer Engineering and Applications

基　　金：江西省光电子与通信重点实验室开放基金(No.2011001);江西师范大学博士启动基金(No.4029)

摘　　要：为了更快地将肤色区域从图像中分割出来,在二维Ostu方法的基础上,提出了一种将积分图和粒子群优化相结合用于寻找最优分割阈值点的新方法。该方法通过对经过颜色补偿后的图片在YCb Cr颜色空间中用二维高斯肤色模型计算肤色概率图,根据二维Ostu算法确定粒子群优化的目标函数,将积分图的方法用于粒子群优化中目标函数的计算,在有效地获取最佳阈值点的同时减少了计算量。实验表明,该方法能有效并且快速地分割出肤色区域。To segment the skin color area faster, this paper puts forward a method to look for the best segmentation threshold point with integral image and particle swarm optimization based on the basic 2D Ostu method. The method uses two-dimen-sional gaussian skin model in the YCbCr color space with a color compensation picture to gain the skin probability graph, and then confirms the particle swarm optimization’s objective function by the 2D Ostu algorithm, and calculates the func-tion’s value with the integral image. The method can get the best threshold point effectively and can shorten the calcula-tion at the same time. Experiments show that the method can segment the skin color area effectively and quickly.

关 键 词：肤色分割 高斯肤色模型 OSTU算法 积分图 粒子群优化 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]