中文摘要：

根据医学图像领域的相关知识,提出了关键像素区域ROI的概念及其特征属性针对ROI聚类后的图像数据进行挖掘,提出了一种基于对FP-GROWTH算法改进的Diff-FP-GROWTH算法。最后,给出了该算法的实例分析和该算法的实际研究意义。

英文摘要：

The term of ROI and its characteristic attributes by the kowledge of medical domain was presented. In addition, in order to extract the knowledge from the medical images which are clustered by the characteristic attributes of ROI, an improved association rule mining algorithm Diff_ FP-GROWTH is presented based on FP- GROWTH. Finally, the algorithm with an example and concluded the significance of it were analyzed.