Enhancement of Colour Fundus Images by using Single and Multi-Scale Retinex

Lighting and image acquisition is main source that affect the image quality. Details of images are lost due improper acquisition of image modalities. There has been a renaissance of interest in the Retinex computation in the past decades, especially in its use for image enhancement. In this paper, we processed retinal colour images which suffered from varying-low contrast and shadows; images were processed through the single scale and multi-scales Retinex with gamma correction or a contrast adjustment factor to enhance, remove the shadows. The newly developed image enhancement technique gave better results with removal of the shadows and provided good contrast ratio with good improved noise level also. In this research work, 16 images were analysed with the proposed image enhancement technique. It gave better enhanced images and with more information as compared to their original images with improved 3db noise level. This result in improvement in the contrast leads to an analysis more features in the retinal colour images and especially for the analysis eye related disease and reduce invasive method such as Fundus Fluorescein angiogram (FFA) which takes injecting contrast agent.