Image Enhancement in Fundus Images using Modified Frangi Vesselness Filter in comparison with Hessian and Gaussian Filter

The main aim of this work to develop image enhancement algorithm using modified vesselness filter for fundus image. Image enhancement using modified vesselness filter has been designed and compared with hessian and Gaussian filters. Performance of proposed algorithm is measured using high values of Peak Signal to Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM), Normalized Correlation (NC), and Naturalness Image Quality Evaluator (NIQE). Modified frangi vesselness filter provides PSNR = 41.31db, SSIM = 92.9%, NC = 23.25%, NIQE = 30.70%. Hessian filter provides PSNR = 41.37db, SSIM = 10.85%, NC = 18.65%, NIQE = 23.64%. Gaussian filter provides PSNR = 11.33db, SSIM = 95.23%, NC = 99.24%, NIQE = 6.29%. The designed filters were tested against various noises like Gaussian, salt and pepper, speckle noise. All three filters performed good against Gaussian noise over other two types of noises. Based on the simulation results modified vesselness filter performed well when compare other two filters.