图像分割方法综述  

A SURVEY ON IMAGE SEGMENTATION METHODS

作　　者：罗希平[1] 田捷[1] 诸葛婴[1] 王靖[1] 戴汝为[1] 

机构地区：[1]中国科学院自动化研究所人工智能实验室,北京100080

出　　处：《模式识别与人工智能》1999年第3期300-312,共13页Pattern Recognition and Artificial Intelligence

基　　金：国家自然科学基金;863高技术资助项目

摘　　要：图像分割是一个经典难题,从七十年代起图像分割问题就吸引了很多研究人员为之付出了巨大的努力,但到目前为止还不存在一个通用的方法,也不存在一个判断分割是否成功的客观标准。本文对图像分割方法,特别是近几年来图像分割领域中出现的新思路、新方法,或对原有方法的新的改进给出了一个比较全面的综述。Image segmentation is one of the classical difficult problem for researcher, much effort had been put into this problem ever since 1970s. But no method which had good result for general images had been proposed and no impersonal criterion for deciding whether the segmentation is success had been generally accepted up to now. In this article, we give a rather complete survey to the image segmentation methods, especially those new ideas, new methods and new improvement to the classical methods appeared in the literature of the last few years.

关 键 词：图像分割 图像处理 统计学 计算机 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]