Research map / Video / Video generation
Hiding Video in Audio via Reversible Generative Models
Hyukryul Yang, Hao Ouyang, Vladlen Koltun, Qifeng Chen
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 flip-rate and bit-importance optimized embedding instead of naive encoding | WaveGlow steganography · DAVIS 128x128 | PSNR: 11.5683 -> 25.4318 | Improved |
| add optimized embedding instead of naive encoding | WaveGlow steganography · DAVIS 128x128 | MS-SSIM: 0.4690 -> 0.9648 | Improved |
| add optimized embedding instead of naive encoding | WaveGlow steganography · DAVIS 128x128 | audio samples per one-second video: 13680 -> 13440 | Improved |
| use binary codes in flow-based latent instead of continuous bottleneck | video hiding in audio · DAVIS 128x128 | PSNR: 23.2836 -> 25.4318 | Improved |
| use binary codes in flow-based latent instead of continuous bottleneck | video hiding in audio · DAVIS 128x128 | audio samples per one-second video: 30720 -> 13440 | Improved |
| use flow-based hiding instead of ConvNet spectrogram hiding | video hiding in audio · DAVIS 128x128 | MS-SSIM: 0.9049 -> 0.9648 | Improved |
| use flow-based hiding instead of ConvNet spectrogram hiding | video hiding in audio · DAVIS 128x128 | audio samples per one-second video: 330750 -> 13440 | Improved |
| increase ICN iterations from K=4 to K=8 (ablation) | high-resolution video hiding · 480x848 video | MS-SSIM: 0.9631 -> 0.9838 | Mixed |
| encode cover audio into stereo with encoded audio and listen | encoded audio · User study | share cannot tell difference: 95.0% | Improved |
Closest papers by meaning
- Seeing and Hearing: Open-domain Visual-Audio Generation with Diffusion Latent Aligners (CVPR 2024 · similarity 0.85)
- XAttnMark: Learning Robust Audio Watermarking with Cross-Attention (ICML 2025 · similarity 0.85)
- MMDisCo: Multi-Modal Discriminator-Guided Cooperative Diffusion for Joint Audio and Video Generation (ICLR 2025 · similarity 0.84)
- MDSGen: Fast and Efficient Masked Diffusion Temporal-Aware Transformers for Open-Domain Sound Generation (ICLR 2025 · similarity 0.83)
- Silence is Golden: Leveraging Adversarial Examples to Nullify Audio Control in LDM-based Talking-Head Generation (CVPR 2025 · similarity 0.83)
- Tri-Ergon: Fine-Grained Video-to-Audio Generation with Multi-Modal Conditions and LUFS Control (AAAI 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.
