Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
July 16, 2026Eng—Advances in EngineeringOpen Access

GEA-YOLO: Real-Time Steel Surface Defect Detection via Deformable Gated Attention and Enhanced Multi-Scale Feature Fusion

View Full Paper
Ask AI
Bookmark
Share

Authors

TWTianfei WangKZKun Zou

Discussion

Loading...

Member takes

Overview

Randomized trial evaluates GEA-YOLO's effectiveness in detecting steel defects, indicating strong potential for real-time inspection.

Key Points

  • The aim is to enhance real-time detection of steel surface defects using advanced deep learning techniques.
  • GEA-YOLO integrates deformable gated attention and enhanced multi-scale feature fusion.
  • Utilizes dynamic gating and auxiliary supervision during training to improve detection accuracy.
  • Evaluated on the NEU-DET and GC10-DET datasets for performance assessment.
  • Achieved 80.3% mAP@0.5 on NEU-DET, outperforming YOLOv11s by 2.6 percentage points.
  • Maintained 169.5 FPS inference speed while ensuring accuracy across datasets.
  • Confirmed robustness through cross-dataset validation, achieving 74.9% mAP@0.5 on GC10-DET.

Cite This Study

Wang et al. (2026) studied this question.

synapsesocial.com/papers/6a58752d2b46c88ba9ad101ehttps://doi.org/10.3390/eng7070344
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Edge-Enhance YOLO for Steel Surface Defect Detection2026 · 1 citations
  2. 2CRGF-YOLO: An Optimized Multi-Scale Feature Fusion Model Based on YOLOv5 for Detection of Steel Surface Defects2024 · 21 citations
  3. 3Squeeze-and-Excitation Networks2019 · 12,749 citations
  4. 4EfficientDet: Scalable and Efficient Object Detection2020 · 9,055 citations
  5. 5Auxiliary Detection Head for One-Stage Object Detection2020 · 22 citations