High-performance RT-DETR model for industrial defect detection with semantic guidance and hierarchical attention
Jin Huang, Bin Gu, Yunlong Tang, Bingyue Xu, Puxuan Li, Jiayu Ye, Jingwen Xu, Taihua Zhang · 2026
Read High-performance RT-DETR model for industrial defect detection with semantic guidance and hierarchical attention by Jin Huang, Bin Gu, Yunlong Tang, Bingyue Xu, Puxuan Li, Jiayu Ye, Jingwen Xu, Taihua Zhang (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/plos-pone-0348807.txt.
Description
Jin Huang and co-workers build a computer vision system that spots flaws on factory products. Finding small surface defects matters for quality and safety, but flaws vary in size and hide in busy textures. Their 2026 paper improves a real-time detection model by adding guidance from word-like meaning cues and a way to blend detail across picture scales.
Tested on three public sets of defect photos, covering steel, patterned cloth, and circuit boards, the new model beat strong rival systems. It found more flaws and drew tighter boxes around them, even in tangled textures. Visual maps showed the model looking at the right spots on each item. The work points toward faster, steadier checks on busy production lines.
Catalog record
| CREATOR | Jin Huang, Bin Gu, Yunlong Tang, Bingyue Xu, Puxuan Li, Jiayu Ye, Jingwen Xu, Taihua Zhang |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 59865 |
| LICENSE | CC_BY |
| RIGHTS | CC BY |
| SOURCE | PLOS |