自适应形状先验的图割分割方法

Graph Cuts Segmentation Method of Adaptive Shape Priors

传统图割在交互式图像分割方面是比较成功的,但当图像含有噪声、部分被遮挡及背景较复杂的情况下,传统的图割方法并不能得到正确的分割结果.针对此问题,提出了一种将自适应形状先验合并到图割的方法,其思想是在图割框架中,除了通常的边界项和区域项外,将水平集距离函数的一个模板作为形状先验包含在图的边权重中,用图的边权重来传递关于图像和先验形状的信息;通过自适应调整参数来调整形状先验在图像中所起的作用;用加速稳健特征和随机抽样一致算法实现形状模板和目标的配准,使形状的变换具有仿射不变性.将此方法用于含有阴影、噪声污染和遮挡情况的自然图像进行处理,相比于不含形状先验的情况,该方法通过自适应形状先验信息约束分割目标的边缘,可以有效地应对阴影、遮挡和噪声问题,取得了较好的分割结果.

：The traditional graph cut is successful in interactive image segmentation, however, such method cannot get the correct seg- mentation results when the image is corrupted by noise, occluded and existed in complex background. To address this problem, a new image segmentation method is proposed by combining adaptive shape priors to graph cuts, besides the traditional boundary item and region item, the basic idea of the proposed graph cut framework is that a template of level set distance functions representing shape priors is contained in the edge weights, and the weight edges of the graph convey information about the image and the prior shape ;by adaptive adjustment of parameters to adjust the shape prior in the different images; Speeded up robust features and Random Sample Consensus algorithm is used to implement shape template and object registration, the shape of affine transformation invariance. The method used to contain corrupted by noises and occluded of natural image,compared to the case without shape prior,the object by con- straining the adaptive shape prior information of the edge, can effectively respond to shadows,occlusion and noise problems,and ob- tained the ideal segmentation results.