Performance Analysis of Some Edge Detection Methods on Magnetic Resonance Images

Magnetic Resonance Imaging is used in many clinical and scientific applications due to its advantages such as the ability to obtain three-dimensional images, high resolution, and precise information. Image processing methods can be used to obtain numerical data from brain MR images and automatically obtain important information in computer environment. At the beginning of these methods, there are methods that reveal the edge information in which a large part of the information in the image is located. There are many standard methods used for edge detection in the literature. In this study, Sobel, Prewitt, Roberts, Gauss's Laplace Canny and Fuzzy Logic edge detection methods were applied on the brain MR images and the results were compared using different threshold values. In addition, in Fuzzy Logic method, the performance comparison of these functions was made by using different membership functions for input and output. When the results were examined, it was observed that the Robert, Prewitt and Sobel methods with a threshold value of 0.03 gave better results than other methods for brain MR images