Retrospective illumination correction of retinal fundus images from gradient distribution sparsity

We present a novel technique for retrospective illumination correction of retinal fundus images from the sparsity property of image gradient distribution. It can automatically estimate the illumination inhomogeneity given an arbitrary retinal fundus image. Experimental results on 665 high resolution fundus images show both the efficiency of our algorithm on illumination correction and its high value on improving the accuracy of blood-vessel segmentation.