基于Mean Shift随机游走图像分割算法  

Mean Shift Based Random Walker Interactive Image Segmentation Algorithm

作　　者：依玉峰[1] 高立群[1] 郭丽[2] 

机构地区：[1]东北大学信息科学与工程学院,沈阳110819 [2]天津医科大学医学影像系,天津300203

出　　处：《计算机辅助设计与图形学学报》2011年第11期1875-1881,1878-1881,共7页Journal of Computer-Aided Design & Computer Graphics

基　　金：国家自然科学基金(81000639);中国博士后科学基金(20100470791);天津医科大学校级基金(2009ky08)

摘　　要：针对传统随机游走算法分割目标轮廓易受自然纹理背景干扰,并且算法运行效率低的问题,提出一种基于Mean Shift随机游走图像分割算法.首先应用Mean Shift算法对图像进行预分割,将图像分成许多同质区域,再将其代替经典随机游走算法中节点来建立对应的无向图;将彩色直方图作为区域描述算子,采用欧氏距离与高斯权函数相结合来建立区域间相似性权函数;最后应用离散电势理论计算图中节点间电势值,并根据节点电势值的大小对预分割得到的同质区域进行分类,以实现图像分割.实验结果表明,与传统方法相比,该算法在分割精度和运行效率上都有很大提高.A Mean Shift based random walker interactive image segmentation algorithm is proposed to solve the problems that the objective contour is prone to the influence of the natural texture background and computation speed is low. Firstly, image is segmented into many small homogeneous regions by Mean Shift pre-segmentation algorithm, and the homogeneous regions are used to build an undirected graph, instead of pixels. Color histogram is used as a descriptor to represent the region color feature statistics, and Euclidean distance and Gaussian weighting function are used to describe the similarity of adjacent regions. Finally, the discrete potential theory is used to calculate the potential of each node in the graph, and the final image segmentation can be achieved according to the greatest potential of each node in the graph. The results of experiments demonstrate that the segmentation accuracy and efficiency of our proposed method is improved significantly comparing with traditional random walker algorithms.

关 键 词：随机游走 图像分割 Mean SHIFT算法 彩色直方图 欧氏距离 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]