Methods to enhance digital fundus image for diabetic retinopathy detection

Digital fundus images (DFIs) are crucial in detecting pathological phenomenon that would lead to various diseases. However, DFI has multiple contrast and illumination problems which makes enhancement a necessity. Consequently, DFI must be enhanced to allow for better visualization in order to facilitate ophthalmologists to carry out their diagnosis. In this work, an investigation of three enhancement methods namely, Histogram Equalization(HE), Contrast Limited Adaptive Histogram Equalization(CLAHE) and Mahalanobis Distance (MD), were conducted on the digital fundus images and the results are qualitatively presented using histogram representation and image product quality. The result shows that the MD is the best algorithm for the application of blood vessels image enhancement.