Content-Based Medical Visual Information Retrieval

Abstract

Medical imaging gives essential visual data of interior body portions for the purpose of clinical investigation and restorative intercession which empowers doctors for analyzing and treating an assortment of maladies. Content-based medical visual information retrieval (CBMVIR) or content-based medical image retrieval (CBMIR) is an active computer vision and medical imaging research area. This is due to the steady increase in visual data production and the expanding assortment of restorative imaging information. CBMIR helps medical specialists in recovering related medical pictures plus case chronicles to understand the particular patient’s infection or damage status and furthermore clarifies about the therapeutic result to the patient. It likewise assists the medical specialist for setting up the report of specific determination all the more precisely. Other than the diagnostics, CBMIR likewise helps with instructing and research. This chapter briefly surveys the various techniques for retrieval of medical images and different methodologies adopted in various retrieval techniques are focused. For retrieval of medical images, a new feature descriptor based on Krawtchouk moment and histogram of oriented gradient (HOG) is also proposed in this chapter. Krawtchouk polynomial-based moments set defines the Krawtchouk moment, and it provides a fine description of the image shape. HOG features compute the gradient orientation happenings in the local region of an image. It mainly extracts local shape information from an image based on the intensity gradient distribution or edge direction. All the experimentations are conducted on high-resolution computed tomography (HRCT) lungs pictures and MRI images of different part of the body. Some other techniques are considered for the comparison with proposed method.