Enhancement of Retinal Images for Microaneurysms Detection in Diabetic Retinopathy

Abstract

Microaneurysms (MAs) are the earliest clinical sign of Diabetic retinopathy (DR). DR is known as microvascular complication of diabetes due to the damage of blood vessel (RBVs) in human retina. In extreme cases, DR may cause visual impairment such as temporary or permanent vision loss if untreated at the early stage. The detection of MAs can be done by processing the retinal images of DR patients captured by digital fundus photography. However, some of the retinal images have low image quality due to some imperfections such as low contrast, illumination and noises issues. Thus, some of the formation of MAs in retinal images is in visibly poor contrast regions, affecting the visual appearance of MAs. This study aims to enhance the quality of colour fundus retinal images for better detection of MAs. The Morphological Enhancement (ME) by combination of top and bottom hat was proposed in this study. The proposed method was compared with other existing method such as Contrast Limited Adaptive Histogram Equalization (CLAHE), adaptive Histogram Equalization (AHE) and Contrast Adjustment Method (CA). The performance was evaluated using Image Quality Assessment (IQA) e.g. Mean Square Error (MSE), Peak Signal to Noise Ratio (PSNR) and Structural Similarity Index Metric (SSIM). The ME method achieved 36.67, 17.56 and 0.93 for the MSE, PSNR and SSIM respectively.