An Expert System Based on Type-1 Fuzzy Logic and Digital Image Processing for Knowledge Based Edge and Contour Detection

Abstract

In computer vision, contour/edge detection is a crucial phenomenon. Edge detection is an important step in contour detection, which is helpful in the identification of important data. The accuracy of the edge detection process is heavily dependent on edge localization and orientation. In recent years, due to their versatility, soft computing approaches have been considered effective edge detection strategies. Broadly, edge detection accuracy is deeply affected by weak and dull edges. In recent works, edge detection based on fuzzy logic (FL) was proposed, and image edges were improved using guided filtering. However, guided image filtering (GIF) does not take into account the local features of an image. To include local features of an image for edge detection, an improved version, i.e., an offset enable sharpening-guided filter is used in this paper, and FL is used for edge detection. The figure of merit (FoM) and F-score are used to evaluate the method's accuracy. Using visual representations and performance metrics, the results are compared with those from cutting-edge techniques.