融合区域合并和Graph Cuts的彩色图像分割方法

针对使用Graph Cuts方法对图像进行分割极大影响分割精度这一问题，提出了一种新的融合区域分级合并和Graph Cuts的彩色图像分割算法。该算法首先使用均值漂移算法对图像进行初始分割，将原图像分割为具有较好边界的同质区域；然后，通过计算区域相似度对区域进行分级合并，之后构建精简的加权图，并使用Graph Cuts进行分割。多幅彩色图像的分割实验结果证明，所提算法具有较好的分割效果。

Color image segmentation based on integration of regional combination and graph cuts

Image segmentation using graph cuts method greatly affects the accuracy of segmentation result. To address this problem, a new method is proposed, integrating region combination and graph cuts. Firstly, it is obtained that the over segmentation of images with mean shift. The original image is divided into many homogeneous regions with good boundary. Secondly, regions with high similarity are merged. Then a weighted graph is built, and segmentation result is obtained with graph cuts method. Experimental results show that the proposed algorithm has better segmentation effect with multiple color image segmentation.