基于超像素的Grabcut彩色图像分割

摘要

To overcome the disadvantage of time load for the image segmentation that set up the graph model in pixels,a Grabcut color image segmentation method which is based on the super pixels is proposed in this paper.Firstly,users can calibrate a rectangular box in the target zone manually,then split the'image into small areas of the similar color (super pixels) with the watershed algorithm two times.Set up the graph model using the super pixels as the graph nodes.In order to estimate the value of GMM,use the mean of the super pixels' color value to represent the all pixels in the same area.Finally,get the minimum value of the Gibbs energy with the minimum cut algorithm to achieve the optimal segmentation.Experimental results demonstrate that the new algorithm uses the little super pixels instead of the huge number of pixels.The algorithm achieves the excellent segmentation result in short run time,speeds up the pace of segmentation,enhances the efficiency of the algorithm.%针对以像素为节点建立图模型进行图像分割耗时的特点,文中提出了一种基于超像素的Grabcut彩色图像分割方法.首先用户在目标所在区域手动标定一个矩形框;然后用两次分水岭算法将图像过分割成区域内颜色相似的小区域(超像素),用分割得到的超像素作为图的结点构建图模型;以每个超像素的颜色均值代表所在分块的全部像素点估计GMM(高斯混合模型)参数;最后用最小割算法求得吉布斯能量的最小值达到最优分割.实验结果表明,该算法以极少数超像素代替海量像素,在得到较好分割结果的同时,极大地缩短了运行时间,加快了分割速度,提高了效率.