中文摘要：

随着计算机医学成像技术的发展,医学影像越来越多样化,医学影像的应用也越来越广泛,与此同时大量的医学设备的应用使得医学影像的数量也越来越多,大量的影像资料使医院迷失在信息的海洋.如何有效的对医学影像进行组织、管理,合理有效的对其进行分类,从而使其更好的辅助日常的医学诊断和医学研究.为了解决这一问题,本文在单特征研究的基础上提出一种综合了多特征融合和数据挖掘的医学图像分类的方法,该方法引入数据挖掘中集成学习的概念,利用改进的Adaboost M1算法针对单特征分类的弱分类器进行迭代训练,并采用SVM分类器开发设计了一个基于改进Adaboost M1算法的医学图像分类系统,以提高医学图像分类效率.

英文摘要：

Along with computer medicine image formation technology development, medical images are more and more diversified , the application of the medical image is more and more extensive, mean- while the application of a large amount of medical device makes the quantity of the medical image more and more , A large number of image data so that the hospital lost in the ocean of information. How ef- fective conduct of medical imaging, management, reasonable and effective for its classification, thus cause it better to support its daily medical diagnosis and medical research. In order to solve this prob- lem, this text proposed one kind of medicine image classification new method that synthesized the multi- characteristic merge and the data mining technology on the basis of single characteristic research. This method is through the introduction of data mining in the concept of Ensemble Learning,utilizingthe im- proved algorithm of Adaboost M1 to classify to the single characteristic the weak sorter to carry on the iterative training. A medical image classification system based on the improved algorithm of Adaboost M1 is designed by using SVM classifier to improve the efficiency of medical image classification.