CONTENT-BASED IMAGE RETRIEVAL FOR MEDICAL IMAGING USING IRMA DATASET

Abstract

Content-Based Image Retrieval (CBIR) for medical imaging helps in efficient diagnosis and treatment planning can be supported by developing retrieval systems to provide high-quality healthcare. In this thesis, different approaches are proposed using IRMA dataset in each block of the CBIR system, And this includes pre-processing, image enhancement as Contrast Limited Adaptive Histogram Equalization (CLAHE ), median filter and gamma correction, feature extraction method as Local Binary Pattern (LBP) and Mesh Local Binary Pattern (MLBP) with different configurations and for the retrieval different methods as Support Vector Machine (SVM), Locality Sensitive Hashing (LSH),Fisher Discriminant Analysis (FDA), Linear Fisher Discriminant Analysis (LFDA), using fusion of them and by using conventional similarity based learning as Euclidean, Mahalanobis, Spearman, Correlation and Hamming Distance. And the evaluation is done using the specific evaluation of IRMA dataset. This thesis gives a detailed implementation for each mentioned step.