Feature Extraction Based Classification of Magnetic Resonance Images Using Machine Learning
2019 6th International Conference on Signal Processing and Integrated Networks (SPIN)(2019)
摘要
Brain abnormalities are the most challenging fatality in the current scenario of health care society. Hence, accurate detection of any abnormality in the brain is highly essential for treatment planning. The clinical diagnosis of the any brain abnormality results in excess medical cost, errors and does not provide correct detection. This work analyzes the texture of images taken of the brain to find the values of various features of the images. Statistical features have been calculated of two types, first order statistics and second order statistics. The standard database has grouped the images into normal and abnormal scans and calculated features for these scans, it was determined which features are useful for categorization. The technique involves texture analysis of medical scans. GLCM and other first order statistical features are used to form a vector space of 675 features. This feature vector is subjected to Principal component analysis to find the optimal number of principal components to train SVM and KNN classifiers. Classification with success rate of 95.45% for SVM and 77.27% for KNN is obtained for the proposed method.
更多查看译文
关键词
Brain scans,GLCM,First order statistics,Second order statistics,KNN,SVM
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络