Reliable detection of retinal lesions in fundus images is required for the creation of an automatic telemedicine system for computer-aided screening and grading of diabetic retinopathy. Propose an automatic approach for detecting microaneurysms and haemorrhages in retina pictures in this project. The most common and often the first lesions to occur as a result of diabetic retinopathy are microaneurysms and haemorrhages. As a result, their detection is required for both pathology screening and follow-up (progression measurement). This task, which is now done by hand, may be automated to improve objectivity and reproducibility. Previous approaches for detecting red lesions in retinal pictures failed to perform consistently across all regions. As a result, reliable detection of lesions in all parts of pictures remains a challenge. Consider the region in your suggested work. Consider region-based segmentation with previous DSF analysis in the suggested study. Red lesions are recognised in this way so that they may be graded simply. It is accomplished by splitting the retinal picture into separate zones in accordance with international standards. This approach adapts the search region for red lesion detection to the image size. Erosion and dilation are used to remove OD and vessels from the resulting image region by region. Finally, a hybrid kernel SVM classifier is used to distinguish between lesion and non-lesion cases. As a result, the sensitivity, specificity, and accuracy of the system can be increased.

fundus, haemorrhage,retinopathy