Abstract: Traditional segmentation methods can not realize true medical image segmentation，therefore a new approach of medical image segmentation based on principle of maximum entropy is presented in this paper.The method uses threshold segmentation，boundary tracking and mathematical morphology in a comprehensive way，and it improves the speed and accuracy of segmentation.Analysis and experiment prove that this method can extract Regions Of Interest（ROI） in the liver tumor CT images.

摘要： 传统的分割方法难以实现医学图像准确地分割，提出了基于最大信息熵原理的医学图像分割方法。该方法集成了阈值分割、边界跟踪和数学形态学，提高了分割的精度和速度。分析和实验结果表明，采用该方法对肝肿瘤CT图像进行分割时，能自动准确地提取出医生感兴趣的区域。