IMPROVEMENT IN MEDICAL IMAGE RETRIEVAL THROUGH COLOR HISTOGRAM AND MACHINE LEARNING

Abstract

A Content Based Image Retrieval (CBIR) framework plays an undeniable critical job in the field of medical diagnoses of images produced from various medical modalities. The present thesis experiments with a new image recovery system with the point of enhancing the outputs of color histograms. Additionally, we aimed to look into how to quantify the feasibility of such strategies employed. Therefore, in research we suggested an strategy for retrieval image dependent on various image characteristics such as image HSV-color-histogram examination, extracting color values from the selected image. The proposed technique produced good test results. In spite of the fact, whichever strategy one selects for the execution of CBIR techniques and mark the performances, it is to be noted of this is where it is hard to achieve absolute results and truths. A number of methods might be surely adept at recovering particular set of images, yet they may perform inadequately on others.