Confidence-feedback-weighted graph matching network: online-offline laser-induced damage site matching under complex interference
Yueyue Han, Guanhua Chen, Hangcheng Dong, Kang Zhang, Fengdong Chen, Zhitao Peng, Fa Zeng, Qihua Zhu, Guodong Liu \corresp F. Chen and G. Liu, No. 92 Xidazhi Street, Harbin 150001, China · 2026
Read Confidence-feedback-weighted graph matching network: online-offline laser-induced damage site matching under complex interference by Yueyue Han, Guanhua Chen, Hangcheng Dong, Kang Zhang, Fengdong Chen, Zhitao Peng, Fa Zeng, Qihua Zhu, Guodong Liu \corresp F. Chen and G. Liu, No. 92 Xidazhi Street, Harbin 150001, China (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-2606-29255.txt.
Description
Han and eight colleagues build a network that matches laser damage sites between two kinds of inspection images. In high-power laser facilities, online images of final optics contain fake damage sites that look much like real ones. Telling them apart needs accurate matching to offline ground-truth sites, but features are weak, shapes distort locally, and many distractor sites confuse the match.
Their model needs only the centre coordinates of each site. After each matching round it estimates how matchable each site looks, and feeds that confidence back as a weight guiding the next round's feature mixing. That feedback suppresses distractors instead of letting them spread. A geometric consistency check reins in false high confidence, and a hard-example loss sharpens the split between look-alike sites. On their Complex-Scene dataset the method reached a matching F1 score of about ninety-six percent.
Catalog record
| CREATOR | Yueyue Han, Guanhua Chen, Hangcheng Dong, Kang Zhang, Fengdong Chen, Zhitao Peng, Fa Zeng, Qihua Zhu, Guodong Liu \corresp F. Chen and G. Liu, No. 92 Xidazhi Street, Harbin 150001, China |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 45 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2606.29255