Abstract: For the special characters of human’s tongue, when directly segmenting the tongue’s image, over-segmentation and over-merging always exist. Based on the characters of the tongue’s image, this paper combines one of the Graph theory image segmentation methods and multi-resolution image segmentation together to segment the image on two different resolutions. Based on the two results’ situation, “AND” or “OR” operation is used to deal with them and the tongue body is exacted from the image. In practice, this method can effectively avoid the above problems.

Key words: image segmentation, Graph theory, Graph-Based algorithm, multi-resolution image segmentation

摘要： 由于人舌体的特殊性质，从舌像中直接分割舌体时常存在过分割和过合并现象。针对舌像特点提出了一种结合图论分割和多分辨率分割的图像分割算法，用一种图论分割算法让舌像在两种分辨率下分别进行分割，根据两种分割的结果把它们进行交或者并处理，从而有效地分割出舌体。实验结果表明这种方法能够有效避免直接使用图论分割时出现分割过度或者欠分割的情况。

关键词: 图像分割, 图论算法, Graph-Based算法, 多分辨率分割