PERBANDINGAN ALGORITMA SOBEL, PREWITT, ROBERT, CANNY, KIRSCH, DAN LAPLACIAN OF GAUSSIAN DALAM DETEKSI TEPI PADA CITRA RONTGEN BATU SALURAN KEMIH

Abstract

Urinary stones are a medical condition that occurs when the bladder cannot expel all the urine it contains, causing minerals in the urine to precipitate and harden. Several methods can be used to diagnose urinary stones, one of which in X-ray imaging. The results of X-ray images are usually blurry and have low contrast, making it difficult to read the edges of objects int the images. Therefore, edge detection in used, and in this study, six edge detection algorithms are employed to sharpen the edges of objects in an image, resulting in images with clear and sharp object edges. This study compares six edge detection algorithms: Sobel, Prewitt, Robert, Canny, Kirsch, and Laplacian of Gaussian on X-ray images of urinary stones. The research methodology involves collecting a dataset, preprocessing, converting images to grayscale, segmentation, and implementing the algorithms. The results of the edge detection algorithms are measured using MSE (Mean Squared Error) and PSNR (Peak Signal-to-Noise Ratio). The results show that the Canny algorithm has the highest average MSE of 3801.45 and the lowest average PSNR of 12.78, while the Sobel algorithm produces the worst results with the lowest average MSE of 3254.95 and the highest average PSNR of 13.72.
Key Words: urinary stones, edge detection, grayscale, segmentation, MSE, PSNR