"Computer-aided Diagnosis in Ophthalmology, Dermatology, and Radiology"

点击翻译按钮获取中文摘要

Computer-aided diagnosis (CAD) of several types of diseases is possible by the application of specialized techniques of digital image processing and pattern recognition; this is the theme of my research program. I propose to develop CAD techniques for several applications in ophthalmology, dermatology, and radiology. Several conditions such as diabetes, hypertension, and retinopathy of prematurity (ROP) affect the blood vessels in the retina; specifically, the shape and width of the vessels are modified by pathological processes. Quantitative analysis of the architecture of the retinal vasculature could assist in monitoring the stage of a disease, evaluating the effects on the visual system, and assessing the response to treatment. I propose to develop image processing techniques to detect and analyze blood vessels in fundus images of the retina, to detect and model the major temporal arcade, and to develop measures to facilitate objective grading and followup of ROP. I propose to develop algorithms to analyze the shape, width, and tortuosity of blood vessels using shape factors, coherence, angular deviation, and fractal analysis. I will develop parametric methods to facilitate longitudinal quantitative analysis to aid followup of patients with ROP and assessment of the effect of treatment. The methods will be tested with images of patients at the Alberta Children's Hospital. I propose to continue the productive collaboration I have established with researchers in Brazil, UK, Spain, India, China, and Italy. The projects include CAD of musculoskeletal deformities (fracture of the vertebrae and distortion of the spine), dermatological ulcers (segmentation and analysis of tissue composition), and breast cancer (analysis of tumors, bilateral asymmetry, and architectural distortion). The proposed techniques of image processing and pattern recognition will facilitate objective and quantitative analysis of medical images. My research program will lead to efficient and accurate CAD and improved prognosis for patients.