数据挖掘在医学图像分类中的应用

Data Mining and Its Application in Medical Image Classification

医学图像的自动分类对于医院的信息自动化有重要的意义。本文通过提取医学影像特征的方法,利用数据挖掘中的决策树、神经网络,对图像特征进行分析,找到能够对图像进行分类的图像特征临界值。本文将数据挖掘的方法应用于正侧位胸片中,给出如何自动区分正侧位胸片的方法,并结合本院的PACS系统应用,证明了这种方法的有效性。

Medical image classification plays a central role in automation of information system of hospitals. This paper introduces the methods of extracting image features ,and two main data mining techniques : decision tree and neural networks. We investigate the use of these techniques for classification of medical images. The results show that the two approaches performed well in our project.