首页 | 本学科首页   官方微博 | 高级检索  
     

基于类标号的快速RSST图像分割方法
引用本文:何兵兵,程义民,潘浩,李杰.基于类标号的快速RSST图像分割方法[J].电路与系统学报,2009,14(5).
作者姓名:何兵兵  程义民  潘浩  李杰
作者单位:中国科学技术大学,电子科学与技术系,安徽,合肥,230027
摘    要:给出了一种基于类标号的快速回归最短生成树(RSST)图像分割方法.在传统RSST方法基础上,将区域相似性度量由像素域距离改进为类标号距离;通过定义类标号距离,可增加类标号图对应的加权图中权重为0的边所占比例,减少了传统方法中的区域合并次数,因此能较多的减少分割用时.方法已在微机上进行了模拟,结果表明,在相同分割数目和相当的分割效果情况下,分割用时明显减少.改进方法更适用于图像的实时分析和处理.

关 键 词:颜色量化  RSST方法  图像分割  类标号

A RSST image segmentation method based on cluster label
HE Bing-bing,CHENG Yi-min,PAN Hao,LI Jie.A RSST image segmentation method based on cluster label[J].Journal of Circuits and Systems,2009,14(5).
Authors:HE Bing-bing  CHENG Yi-min  PAN Hao  LI Jie
Affiliation:HE Bing-bing,CHENG Yi-min,PAN Hao,LI Jie(Dept.of EST,University of Science , Technology of China,Hefei 230027,China)
Abstract:The paper presents a recursive shortest spanning tree image segmentation method based on cluster label.Different from the traditional RSST method,the cluster label distance is used to measure the similarity of two regions instead of the pixel intensity distance in the new method,then the rate of zero weight edge rises in the graph mapped from the cluster label map,and the count of area combination decreases,finally the segmentation time reduces too.The result of experiments shows given the same area count t...
Keywords:Color Quantization  RSST method  image segmentation  cluster label  
本文献已被 CNKI 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号