Research map / Label-efficient and robust learning / Recognition and long-tail
Revisiting Coding-Based Approaches to Overcome the Curse of Dimensionality in Learning-Based Watermarking
Yupeng Qiu, Han Fang, Ee-Chien Chang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| replace correlated projection with orthogonal projection (MSE loss) (ablation) | OrthoMark SED · 64-bit, 15 distortions | PSNR (dB) / average accuracy: PSNR 32.41 -> 39.20; AVG 93.82 -> 99.45 | Improved |
| replace MSE loss with QIM-style loss under orthogonal projection (ablation) | OrthoMark SED · 64-bit, 15 distortions | PSNR (dB): 39.20 -> 39.26; AVG 99.45 -> 99.59 | Improved |
| replace MSE loss with QIM-style loss under correlated projection (ablation) | OrthoMark SED · 64-bit, 15 distortions | average accuracy: 93.82 -> 92.02 | Hurt |
| use QIM-style loss instead of MSE for large payloads (ablation) | OrthoMark SED · bit accuracy vs payload (clean) | bit decoding accuracy | Improved |
| use orthogonal instead of correlated projection for capacity scaling (ablation) | OrthoMark SED · bit accuracy vs payload (clean) | bit decoding accuracy | Improved |
| increase inference optimization iterations from 500 to 4500 (ablation) | OrthoMark · 64-bit, 15 distortions | PSNR (dB): 37.40 -> 39.71 | Improved |
| increase inference optimization iterations beyond 1000 (ablation) | OrthoMark · 64-bit, 15 distortions | average accuracy: 99.45 -> 99.47 | No clear effect |
| use OrthoMark instead of deep watermarking baselines at 256 bits | OrthoMark · 256-bit, 15 distortions | average accuracy (%): 84.62 (CIN) -> 90.60 | Improved |
Closest papers by meaning
- END^2: Robust Dual-Decoder Watermarking Framework Against Non-Differentiable Distortions (AAAI 2025 · similarity 0.87)
- Lightweight-Mark: Rethinking Deep Learning-Based Watermarking (ICML 2025 · similarity 0.86)
- Decoder Gradient Shield: Provable and High-Fidelity Prevention of Gradient-Based Box-Free Watermark Removal (CVPR 2025 · similarity 0.85)
- RAW: A Robust and Agile Plug-and-Play Watermark Framework for AI-Generated Images with Provable Guarantees (NeurIPS 2024 · similarity 0.84)
- WaterMax: breaking the LLM watermark detectability-robustness-quality trade-off (NeurIPS 2024 · similarity 0.84)
- Mask Image Watermarking (NeurIPS 2025 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
