A Deformable Attention-Based Detection Transformer with Cross-Scale Feature Fusion for Industrial Coil Spring Inspection
Matteo Rossi, Pony Matt · 2026
Read A Deformable Attention-Based Detection Transformer with Cross-Scale Feature Fusion for Industrial Coil Spring Inspection by Matteo Rossi, Pony Matt (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-13429.txt.
Description
Matteo Rossi and Pony Matt propose MSD-DETR for visual inspection of locomotive coil springs. Defects vary in shape and scale, and industrial backgrounds are complex, so checks are hard. Their detection transformer uses structural re-parameterization to split training-time branches from fast inference. A deformable attention mechanism samples space in a content-adaptive way, focusing on defect zones. Cross-scale feature fusion with GSConv and VoVGSCSP blocks merges multi-resolution data. On a real dataset it reached 92.4 percent mAP at 98 FPS, beating YOLOv8 and RT-DETR.
Catalog record
| CREATOR | Matteo Rossi, Pony Matt |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 61772 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.13429