Research On Medical Image Intelligent Classification Based On Association Rules

The image-oriented medical research as a medical diagnostic technology and the interdisciplinary computer science, become an important medical research direction at home and abroad. Using data mining and computer technology, excavating rich feature information and rules from medical images, assisting a medical image clinical diagnosis, the direction has high academic value and broad application prospects. Currently, medical image-oriented data mining study has just started, and the existing direct application of data mining methods in medical image still exist many problems. Research and explore suitable for medical image data mining method and algorithm is important and realistic significance.This study systematically reviewed the medical image Mining Association Rules purpose, significance and study conditions at home and abroad, for images of their own characteristics and medical image data mining application requirements, a detailed analysis of medical images Mining Association Rules. Association for the existing classification algorithm existing problems, based on association rules to the medical image classification algorithm, and the initial establishment of the characteristics of medical image-based Intelligent Image classifier The research work and the results list as follows: 1,According to the characteristics of medical images, respectively, based on the color histogram feature based on the co-occurrence matrix of gray texture characteristics, Gabor wavelet-based texture characteristics of the feature extraction and analysis methods to achieve.2,To extract the characteristics of medical images for use of rough set attribute reduction, and use of knowledge-based rules weighted method to determine the characteristics of the attributes weights.3,Against CMAR algorithm (Accurate and efficient classification based on multiple class-association rules, CMAR) when used for medical image Mining inefficient shortcomings, based on double support to the mining algorithm. This algorithm can reduce memory overhead, improve the accuracy of Mining, suitable for image database, and other mass Mining Association Rules.4,To establish and achieve intelligent medical image classifier model algorithm and its application CMAR improved by the Mining Association Rules on Medical Image algorithm to classify. Key words: Medical images, feature extraction, feature-weighted, association rules, intelligent classification...