August 15, 2025Applied SciencesOpen Access

Steel Surface Defect Detection Algorithm Based on Improved YOLOv8 Modeling

View Full Paper

Ask AI

Bookmark

Authors

Discussion

The enhanced YOLOv8 model boosts detection accuracy to 80.2% on the NEU-DET dataset, highlighting improved steel defect identification.

Detection accuracy reached 80.2% on the NEU-DET dataset, surpassing previous methods.
The average precision on the Severstal dataset increased from 65.4% to 66.1%, indicating improved performance.
The model integrates an improved multi-scale attention mechanism to enhance feature extraction.
Advanced feature fusion techniques were employed to achieve better detection precision.

Peng et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark