Swarm intelligence based adaptive gamma corrected (SIAGC) retinal image enhancement technique for early detection of diabetic retinopathy

Retinal images are low contrast, complex, suffer the problems of blurring and uneven illumination. So, it is very difficult to identify the vascular abnormalities. Some of the existing fundus sensors result noisy information, which makes scientific evaluation and diagnosis more complicated. So, for detecting the vascular abnormalities and for evaluating the early stages of Diabetic Retinopathy (DR), a robust retinal image enhancement technique (SIAGC) is proposed in this paper. To enhance the contrast of retinal images effectively and to make the detection process easier, the object of the retinal image is 1st extracted from the background. To avoid the over enhancement, plateau limit is applied separately to both regions. Then mapping function and adaptive gamma parameter have been evaluated using modified weighted probability density function (PDF) for enhancing the image quality. The plateau thresholds and exponentiation parameters used for gamma correction are automatically selected using swarm intelligence in order to maximize the proposed fitness function. It improves the adaptive-ness of the proposed technique. A multi-objective fitness function is proposed in this paper, which includes entropy, edge contents, AMBE and PSNR. Experimental results indicate the robustness of the proposed SIAGC based enhancement technique over other state of the art techniques. © 2021 Elsevier GmbH