基于纹理与灰度协同进化的图像分割算法  

Image segmentation algorithm based on coevolution with texture and gray scale

作　　者：袁宝峰[1] 吴乐华[1] 曾伟[1] 

机构地区：[1]重庆通信学院信号与信息处理实验室,重庆400035

出　　处：《计算机应用》2009年第1期54-56,共3页journal of Computer Applications

基　　金：重庆市自然科学基金资助项目(CSTC2007BB2105)

摘　　要：为了获得更好的分割效果,成功地将局部二值模式(LBP)纹理模型和灰度特征纳入到合作型协同进化算法(Co-CEA)框架中,并实现了图像分割。该方法首先分别对LBP纹理模型和灰度特征进行编码,然后运用Co-CEA进行进化操作,最后通过本文提出的联合适应度函数确定分割区域。实验结果表明该方法在分割质量上效果明显,并有效地降低了时间复杂度。To obtain better segmentation effect, Local Binary Patterns （LBP） and gray-scale characteristics were brought into the framework of Cooperative Coevolutionary Algorithm （Co-CEA） successfully in this paper. LBP and gray-scale characteristics were firstly encoded, and then evolution operation was carried out using Co-CEA, Finally, the division of region was fixed by the united fitness function. Result shows that this method has very good effect in improving the quality of segmentation and lowering the time complexity.

关 键 词：局部二值模式(LBP) 合作型协同进化算法(Co—CEA) 联合适应度函数 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]