An Efficient and Adaptive Watermark Detection System with Tile-based Error Correction
Xinrui Zhong, Xinze Feng, Jingwei Zuo, Fanjiang Ye, Yi Mu ⋄ \diamond , Junfeng Guo † \dagger , Heng Huang † \dagger , Myungjin Lee ‡ \ddagger , Yuke Wang Rice University ⋄ \diamond University of Illinois Urbana–Champaign † \dagger University of Maryland ‡ \ddagger Cisco Research · 2025
Read An Efficient and Adaptive Watermark Detection System with Tile-based Error Correction by Xinrui Zhong, Xinze Feng, Jingwei Zuo, Fanjiang Ye, Yi Mu ⋄ \diamond , Junfeng Guo † \dagger , Heng Huang † \dagger , Myungjin Lee ‡ \ddagger , Yuke Wang Rice University ⋄ \diamond University of Illinois Urbana–Champaign † \dagger University of Maryland ‡ \ddagger Cisco Research (2025) 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-2509-02447.txt.
Description
Xinrui Zhong and eight co-authors present a 2025 systems paper on finding watermarks hidden in generated images across huge collections. Accuracy under edits and attacks gets much attention, but raw detection speed at scale is often ignored.
Their QRMark system borrows error correction ideas from QR codes and splits images into tiles. A Reed-Solomon code repairs accuracy lost through tiling. On the system side, a resource-aware fusion policy gives more processing streams to heavy stages. Tile workloads are interleaved so data loading overlaps with computation, and kernels are scheduled across stages. End-to-end tests show detection running more than twice as fast as a sequential baseline, while keeping accuracy and robustness.
Catalog record
| CREATOR | Xinrui Zhong, Xinze Feng, Jingwei Zuo, Fanjiang Ye, Yi Mu ⋄ \diamond , Junfeng Guo † \dagger , Heng Huang † \dagger , Myungjin Lee ‡ \ddagger , Yuke Wang Rice University ⋄ \diamond University of Illinois Urbana–Champaign † \dagger University of Maryland ‡ \ddagger Cisco Research |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 60 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-02447