Research map / Video / Video generation
LoT-Pass: Long-term-robust Image Watermarking for Image to Video Generation
Guanjie Wang, Zehua Ma, Han Fang, Weiming Zhang
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 |
|---|---|---|---|
| add temporal evolution simulation (VAE reconstruction and random warping) to the training noise layer (ablation) | LoT-Pass · I2V FACC across models | frame accuracy | Improved |
| add temporal inversion (optical-flow realignment) at extraction (ablation) | LoT-Pass · I2V FACC across models | frame accuracy | Improved |
| drop increasing share of early video frames (0% to 20%) | LoT-Pass · CogVideoX | VACC / RDD: 0.966/49 -> 0.910/13 | Hurt |
| drop 20% of early frames on SVD (no prompt conditioning) | LoT-Pass · SVD | VACC / RDD: 0.823/1 -> 0.582/0 | Hurt |
| high-motion instead of low-motion videos | LoT-Pass · CogVideoX | VACC: 0.971 -> 0.955 | Hurt |
| LoT-Pass vs strongest classic baseline on random bending | LoT-Pass · Random bending | bit accuracy: 74.78 -> 97.53 | Improved |
| LoT-Pass vs WAM under real crop | LoT-Pass · RealCrop 40-80% | bit accuracy: 99.09 -> 98.09 | Hurt |
| commercial models (Kling, Hailuo) instead of open-source | LoT-Pass · RDD | RDD: Kling 120, Hailuo 47 vs CogVideoX 49 | Improved |
Closest papers by meaning
- Robust Watermarking Using Generative Priors Against Image Editing: From Benchmarking to Advances (ICLR 2025 · similarity 0.89)
- T2SMark: Balancing Robustness and Diversity in Noise-as-Watermark for Diffusion Models (NeurIPS 2025 · similarity 0.88)
- SleeperMark: Towards Robust Watermark against Fine-Tuning Text-to-image Diffusion Models (CVPR 2025 · similarity 0.87)
- BitMark: Watermarking Bitwise Autoregressive Image Generative Models (NeurIPS 2025 · similarity 0.87)
- Safe-Sora: Safe Text-to-Video Generation via Graphical Watermarking (NeurIPS 2025 · similarity 0.87)
- VideoShield: Regulating Diffusion-based Video Generation Models via Watermarking (ICLR 2025 · similarity 0.86)
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.
