一种随机游走图像自动分割算法

An automatic random walker algorithm for image segmentation

为了解决传统随机游走算法需要过多的人工干预因素,限制了算法的通用性,提出一种自动随机游走图像分割方法—SWRW算法。首先应用二次分水岭进行预分割,将预分割形成的若干同质区域取代传统算法中的节点来建立无向图;利用色调均值差制定准则自动选取种子区域并对它们进行标记,将彩色直方图作为区域描述算子,采用巴氏系数和高斯权函数建立区域间相似性权函数;最后应用狄利克雷边界条件,实现图像分割。该方法运算速度快,避免了用户的繁琐操作,实现了完全自动分割。结果表明,与其他相关方法比较具有更好的鲁棒性和分割精度。

In order to solve the problem that requires some factors by manual in the traditional random walker algorithm, limit the generality of the algorithm, an automatic random walker （SWRW） image segmentation method is proposed. Firstly, image is segmented into many small homogeneous regions by secondary watershed pre-segmentation algorithm, then the homogeneous regions are used to build an undirected graph, instead of pixels. Standards have been formulated for the mean hue difference, which is used to select and make signs of seed regions, color histogram is used as a descriptor to represent the region color feature statistics, then Bhattacharyya coefficient and Gaussian weighting function are used to describe the similarity of adjacent regions. Finally, high-quality segmentation results can be achieved by solving Dirichlet boundary condition. The designed method reduces the runtime significantly which is very feasible for real-time image processing, it simplifies the user interactive operations as well as improves the segmentation accuracy and robustness of random walker algorithm. The experimental results show SWRW method has superior performance compared to the other related methods.