Devignetting Fundus Images Via Bayesian Estimation of Illumination Component and Gamma Correction

Fundus photography is a type of ophthalmic imaging that is used to see structures such as macula, retina, and optic disc. The fundus camera has only one source of light, which is located in the middle. As a result, structures further from the center will look darker than they are in reality. 

Vignetting is the term for the negative impact generated by uneven light. This work proposes the Gamma Correction of Illumination Component technique, which is inspired from retinex theory for devignetting fundus images. 

The Gamma correction is applied to the estimated illumination component after normalisation to reduce its unevenness. In terms of performance measures, the suggested technique performed much better, especially peak-signal-to-noise-ratio (PSNR), structure similarity index (SSIM).

Keywords: Fundus imaging, Gamma correction, Vignetting