Diabetic Retinopathy (DR) is a prominent cause of blindness and visual problem that affects the eyes of humans who are affected by diabetics. Most of the time it does not show symptoms at an early stage and it is hard for the patient to identify the symptoms until a visual ability degrade and the treatment becomes is less effective. It becomes tough for medical experts (ophthalmologists) to identify DR at an early stage manually by observing the retinal image taken by a fundus camera. Thus, computer-aided image processing of retinal images taken by fundus camera has tremendous advantages to detect retinal lesions associated with Diabetic Retinopathy at an early stage. With less time and effort, the computer aid image processing examines a large number of images more accurately than the manual observer-driven techniques. It becomes important diagnostic aid to reduce the workload of ophthalmologists. However, the presence of various artifacts like the similarity of anatomical structures, movement of the patient eye during image capturing, device noise, and illumination makes the segmentation and processing of images of major pathological structures a difficult task. In this study, we have developed a retinal image segmentation algorithm and user-friendly software that can ease the task of the medical experts by automatically identifying Hard Exudates (HEs), which are the most prevalent characteristic features of Diabetic Retinopathy in its earliest stage. The algorithm first is written and tested using MATLAB then user-friendly software is developed using C# programming language in the Microsoft .Net framework. To classify and segment the retinal image taken by the fundus camera a general representation of images color in the three spaces (trinion) has been used and to extract image features a trinion based Fourier Transforms has also been applied. Neural Network (NN) based segmentation of Hard Exudates are included in the method for color space transformation and to extract features. The efficiency of the developed image processing has been tested in classifying and identifying hard exudated and it shows better results.

Keywords : Retinal Imaging, Image Processing, Image Segmentation, Neural Network, Diabetic Retinopathy.