Particle Swarm Optimization Based Retinal Image Enhancement

Rear portion of the eye can be pictured by the retinal fundus imaging technique. The retinal fundus images provide clear description about the retinal blood vessels and optic disc regions. Most of the retinal diseases have perceptible symptoms on these regions and hence this can facilitate the optometrist for early diagnosis and treatment of retinal diseases. Blur, uneven illumination and poor contrast in the retinal fundus images are the commonly seen challenges for both manual and automated image analyzing systems. In this paper, a new method is proposed for the enhancement of uneven illuminance and contrast fundus images by adjusting their luminance and contrast based on Particle swarm optimization (PSO). The proposed method incorporating gamma correction on the HSV color space and contrast adjustment on the LAB color space. The contrast adjustment is optimized by the PSO algorithm to improve the enhancement process. The proposed method is evaluated on the publicly available databases DIARET DB0 and DIARET DB1. The Structural Similarity Index and Peak Signal to Noise Ratio are the performance measures used to evaluate the enhancement method. From the results obtained it is inferred that the fundus image with uneven illumination and contrast is enhanced in a better way by the proposed method.