Microaneurysms segmentation in retinal images for early detection of diabetic retinopathy

Microaneurysms (MAs) are the tiny aneurysms which show the earliest sign of diabetic retinopathy (DR). MAs might progress and harm human eyes if not treated. This paper presents an automatic method for segmentation of MAs in order to control the progression of DR. MESSIDOR database of 40 random images were utilised for further processing. The proposed approach covered pre-processing steps, contrast enhancement, filtration and segmentation by h-maxima transform and multilevel thresholding. Some post-processing techniques also involved in this approach using morphological operation. The detected MAs determined the grade of disease severity. The result showed that the percentage of severity disease detected was 60%.