Modified Fractional Edge Detection Masks Based on Compass Gradient for Retinopathy Screening

Edge detection is pivotal in uncovering object boundaries and identifying discontinuities in digital images. This research paper proposes four novel edge detection techniques based on ABC-fractional integral using compass gradient masks, focusing on enhancing accuracy while maintaining reasonable computational effort. We thoroughly assess the results through visual evaluation by conducting extensive experiments on sample images from datasets such as BIPED, DRIVE, and HRF. Our findings demonstrate the superiority of the proposed methods over nine traditional and several recently introduced edge detection techniques. Moreover, we evaluate the performance of these algorithms using MSE, PSNR, SSIM, and EBIQA, demonstrating their promising potential for precise edge detection in Retinopathy images. Furthermore, concerning computational expenses, the fractional compass masks proposed in this study require nearly identical computational effort as the existing conventional masks.