Abstract:
Image processing has gained lot of importance over the last few years. Its applications
are used in almost all the fields especially medical field. The development of automated
techniques for detection of diabetic eye diseases has become a reality with the exponential
evolution of information processing system and the emergence of economical ophthalmic
imaging devices. Several automated techniques are being designed and used for practical
applications all over the world.
Digital retinal imaging uses high-resolution imaging system to take pictures of the inside of
eye. This helps the doctors assess and manage the health of retina. Retinal imaging providesopportunity for the diagnosis of several medical pathologies. A particular type of pathologywhich occurs due to diabetic retinopathy is bright lesions. The manual analysis of retinal
images is time-consuming and expensive. The automation of certain processing steps is thusimportant and facilitates the subsequent decisions by specialists to provide a basis for earlydiagnosis steps of specific diseases. The automatic segmentation of the vessel tree is animportant processing step which facilitates subsequent automatic processes that contributeto such diagnosis. This can help to point out the area of disease. This work propose anautomatic/semi-automatic technique to distinguish non-disease images and disease imagescontaining bright lesions.
The proposed work will contribute to the development of a system which can be used in
hospitals to pre diagnose the cases and refer the relevant cases to the ophthalmologist for
attention. The research can pre-diagnose the bright lesion cases in patients and automate thedisease detection.
