Title: Fuzzy logic image processing

Authors: Adnan Shaout; Daniel Murray; Abdelwahed Motwakel

Addresses: The Department of Electrical and Computer Engineering, The University of Michigan – Dearborn, Dearborn, Michigan, USA ' The Department of Electrical and Computer Engineering, The University of Michigan – Dearborn, Dearborn, Michigan, USA ' Department of Computer Science, Omdurman Islamic University, Sudan; National University – Sudan, Khartoum, Sudan

Abstract: With facial recognition software becoming more widely used, especially in mobile apps such as Snapchat, boundary detection will continue to be one of the primary areas of interest in enhancing software performance. Edge detection is paramount in discriminating objects so they can be used and processed. This paper will present a fuzzy system for boundary detection. The proposed system will then compare it to traditional methods of edge detection using MATLAB's image processing toolbox. The proposed fuzzy system allows for more effective tool since the membership functions can be more defined and robust to accommodate different images, as well as tailored to result in a more effective processed image due to ambiguity. The results of the simulation of the fuzzy system shows that it was more capable of dealing with ambiguities in the input images, so in pictures where the edges were not so clear or required more detail, the fuzzy system was more capable since it can be better tailored for use.

Keywords: Roberts detection; Prewitt detection; Sobel detection; Canny detection; fuzzy-based detection.

DOI: 10.1504/IJKEDM.2019.102489

International Journal of Knowledge Engineering and Data Mining, 2019 Vol.6 No.3, pp.207 - 233

Received: 12 Sep 2018
Accepted: 20 Mar 2019
Published online: 27 Sep 2019 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article