De-noising and contrast enhancement of fundus image through integration of filtering techniques with CLAHE

Abstract

Now-a-daysnmedical fundus imagesnare generally used innclinical diagnosis for the recognitionnof retinal disorders. Fundusnimages are generallyndegraded by noise and alsonsuffers from lowncontrast issue. These problems makenit difficultnfor ophthalmologist tondetect and interpretndiseases in fundus images. This thesisnwork presents a noisenremoval and contrast enhancementnalgorithm for fundus image enhancementnby combining filtersnwith Contrast LimitednAdaptive Histogram Equalizationn (CLAHE) technique. nPrimarily input fundus imagenis divided in tonred, greennand blue components. Furthernfiltering process is appliednon each componentnrespectively to remove noisy pixelsnand finally CLAHEntechnique is applied to increasencontrast of the de-noisednfundus image. The efficacynof the proposednmethod is claculated throughndifferent performancenparameters like Peak Signal tonNoise Ration (PSNR), StructuralnSimilarity Index (SSIM) andnCorrelation coefficientn (CoC) and the results indicatenthat the proposednalgorithm is better than simplenCLAHE technique. In this thesis work, Chapters are prearranged as follows.