Abstract: This paper proposes a watershed segmentation algorithm based on adaptive marker extraction and energy function in order to overcome over-segmentation problem. The initial step gains the shortest path via information of edge and method of graph to automatically extract region markers. The next step is watershed transformation. And the last step is region merging with proposed energy function. Experimental results show that the proposed method can efficiently reduce over-segmentation and has better performance as well.

Key words: adaptive marker extraction, energy function, watershed algorithm

摘要： 针对传统分水岭算法的过分割问题，提出一种基于自适应标记提取和能量方程的改进算法。根据图像中的边缘信息和图论方法，得到最短边缘路径，从而自适应地提取出区域标记，进行分水岭变换，用提出的能量方程实现区域合并。利用医学细菌图像对提出的算法进行验证，实验结果表明该算法能有效解决分水岭算法的过分割问题，得到很好的分割效果。

关键词: 自适应标记提取, 能量方程, 分水岭算法