Hierarchical Fuzzy Inference System for Edge Detection Using Adaptive Multi-Scale Gradient Analysis and Evolutionary Threshold Selection

Resumen

Edge detection remains a fundamental challenge in computer vision and image processing, requiring robust methods that can handle varying illumination conditions, noise levels, and image complexities. This paper presents a novel hierarchical fuzzy inference system (HFIS) that integrates adaptive multi-scale gradient analysis with evolutionary threshold selection for enhanced edge detection performance. The proposed system employs a three-tier hierarchical architecture where each level processes gradient information at different scales, utilizing fuzzy logic to handle uncertainty and imprecision inherent in edge detection tasks. An evolutionary algorithm optimizes threshold parameters across multiple scales, ensuring adaptive performance across diverse image types. Experimental validation on standard benchmark datasets demonstrates superior performance compared to conventional edge detection methods, achieving an average F-measure of 0.847 and significantly reduced false positive rates. The system exhibits robust performance across various noise conditions and image complexities, making it suitable for real-time applications in autonomous systems and medical imaging.