CBM3D: A New Approach to Edge Detection
Zhuoyue Wang, Yiyi Tao, Danqing Ma, Jiajing Chen · 2025
Context-based Multi-scale 3D is a novel edge detection method that enhances accuracy and robustness by integrating multi-scale and contextual information. Unlike traditional methods that rely on local features and are vulnerable to noise and lighting variations, CBM3D leverages multi-scale 3D context analysis to better capture image details and structure. By extracting texture information across multiple scales and incorporating it into the detection process, CBM3D improves edge coherence and clarity. Additionally, adaptive thresholding and noise suppression allow it to maintain high accuracy even with lowquality images. Experimental results demonstrate that CBM3D outperforms conventional methods, especially in complex backgrounds and low-contrast scenarios.