The Key Technologies And Implementation Of Computer Aided Diagnosis System Of Fundus Images

The changes of people’s fundus distribution structure and morphological characteristics couldreflect the severity of diseases to a certain extent. Digital fundus camera is an effective tool that canprovide a clear picture of the eye, and can help doctors to make accurate, objective andcomprehensive analysis. Computer-Aided Diagnosis techniques have been widely used in the field ofmedical images recently. Due to the advanced computer technologies, Image processing and analysisof medical images significantly improved the clinical diagnosis and treatment.In this paper, the key image processing and analysis technologies of Computer-Aided Diagnosishave been studied, such as fundus image enhancement, image mosaicking, the segmentation andmeasurement of optic disc and cup of fundus image and so on. Through the practical application weselect basic Image processing method which suits fundus image best. In the actual clinical diagnosis,interactive mosaic sometimes is more practical, we carry out the interactive splicing of fundus imagesbased on template matching method. Basic image segmentation methods have been studied and opticdisk of fundus image has been extracted by some typical segmentation methods. The segmentation foroptic cup-disk of fundus image is a difficult point. Based on the research of C-V mode1, combinedwith interactive method, users set an initial contour and some other parameters so that getting theoptic disk and cup of the fundus image. Then we can get the correlative measurement parameterswhich have important accessory diagnostic value for some fundus disease. At last, taking C++Builder2010as a platform, a basic Computer-Aided Diagnosis system using fundus images has beenestablished.