摘要
在图像处理过程中,快速准确地提取感兴趣的目标是图像处理的一个重要任务.提出了一种基于Grabcut准确提取图像目标的算法,该方法根据用户指定前景物的基本形态获取初始的前景、未知的背景区域,并用mincut/maxflow对分割能量函数进行了优化,利用大量图像对本文方法进行了实验分析,结果表明了算法能有效提取图像中感兴趣的目标.
In image processing , tastly and accurately extracteu the russet t,, ,,~=. to v age processing tasks. An algorithm based Grabcut accurately extract the image target of the method is pro- posed, according to the basic form of user--specified foreground object to the prospect of access to the ini- tial, unknown background area, and the mincut / maxflow segmentation energy function is optimized, u- sing a large number of images in this paper to take the experimental analysis , the results show that the al- gorithm can effectively extract the image the target of interest .
出处
《青海师范大学学报（自然科学版）》
2012年第3期30-33,38,共5页
Journal of Qinghai Normal University(Natural Science Edition)
关键词
图割
GRABCUT
图像分割
GMM
graph cut
grabcut
image segmation
gaussian mixture model.
