基于Grabcut的图像目标提取

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.

Image target detection based on the Grabcut

In image processing,fastly and accurately extracted the target of interest is an important image processing tasks.An algorithm based Grabcut accurately extract the image target of the method is proposed,according to the basic form of user-specified foreground object to the prospect of access to the initial,unknown background area,and the mincut / maxflow segmentation energy function is optimized,using a large number of images in this paper to take the experimental analysis,the results show that the algorithm can effectively extract the image the target of interest.