August 4, 2025

CMH‑Net:a structured and optimized network for real-time steel surface defect detection

View Full Paper

Ask AI

Bookmark

Authors

Discussion

Innovative CMH-Net improves detection accuracy in steel surface defects, suggesting enhanced real-time performance in industry applications.

CMH-Net significantly enhances steel surface defect detection accuracy while maintaining real-time performance.
The model incorporates a unique cascading downsampling module for improved feature extraction and representation.
Experimental results show CMH-Net outperforms previous methods, achieving a mAP50 of 0.814 on NEU-DET dataset.
With only 3.6M parameters and 6.4ms inference time, CMH-Net optimally balances detection and efficiency.

Li et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark

Innovative CMH-Net improves detection accuracy in steel surface defects, suggesting enhanced real-time performance in industry applications.