Sliding window based adaptative fuzzy measure for edge detection

Cedric Marco‐Detchart,Giancarlo Lucca,Miquéias Amorim Santos Silva,Jaime A. Rincon,Vicente Julian,Graçaliz Dimuro
DOI: https://doi.org/10.1111/exsy.13730
IF: 3.3
2024-10-06
Expert Systems
Abstract:In this work, we explore the impact of adaptive fuzzy measures on edge detection, aiming to enhance how computers interpret images by identifying edges more accurately. Traditional methods rely on analysing changes in image brightness to locate edges, but they often use fixed rules that do not account for the unique characteristics of each image. Our approach differs by adjusting fuzzy measures based on the information within specific areas of an image under a sliding window approach, utilizing a variety of fusion functions and generalizations of the Choquet integral to analyse and combine pixel data. The proposed method is flexible, allowing for the adaptation of measures in response to the image's local features. We put our method to the test against the well‐established Canny edge detector to evaluate its effectiveness. Our experimental results suggest that by adapting fuzzy measures for each image section, we can improve edge detection results.
computer science, artificial intelligence, theory & methods
What problem does this paper attempt to address?