Optimization of CBMIR system performance using fuzzy clustering with random forest classifier

Abstract

During the past decades, the usage of image data has been permanently increasing, which leads to huge repositories. Content Based Image Retrieval (CBIR) methods have tried to alleviate the access to image data. CBIR also known as query by image content is to search for images with similar content in a large collection of images with computer vision technologies. In medical field, digital images are produced in ever increasing quantities and used for diagnostics and therapy. With medical imaging techniques such as X-Ray, computed tomography, magnetic resonance imaging, mammographic imaging and ultrasound, the amount of digital images that are produced in hospitals is increasing incredibly fast. Thus, the need for systems that can provide efficient retrieval of images of particular interest is becoming very high. The swift expansion of digital medical images has enforced the requirement of efficient CBIR for retrieving medical images that are visually similar to query image. Such systems provide great assistance to doctors in clinical care and research. CBIR systems for medical images retrieving based on similarity of their visual contents and help the doctors to diagnose the disease efficiently. Many research works were developed in content based medical image retrieval, but the techniques have the drawback of low efficiency and high computation cost. It aims at reducing the need for textual description and to provide the most appropriate images automatically and computationally faster.. To avoid such negative aspects, a new enhanced Content Based Medical Image Retrieval (CBMIR) is proposed in this research work. The proposed CBMIR technique consists of four stages which include (i) Feature Extraction (ii) Clustering Process, (iii) Image retrieval and (iv) Relevance Matching (RM). newline newline newline