Research map / Image generation / Image restoration
UDH: Universal Deep Hiding for Steganography, Watermarking, and Light Field Messaging
Chaoning Zhang, Philipp Benz, Adil Karjauv, Geng Sun, In So Kweon
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 |
|---|---|---|---|
| cover-independent hiding (UDH) vs cover-dependent DDH | deep steganography · ImageNet | cover PSNR: 35.87 -> 39.13 | Improved |
| UDH vs DDH under uniform noise on cover | deep steganography · noise 50/255 | secret APD: 5.9 -> 52.3 | Hurt |
| UDH vs DDH under constant shift on container | deep steganography · shift 50/255 | secret APD: 32.4 -> 3.5 | Improved |
| train with combined distortions vs no distortions (ablation) | UDH watermarking · JPEG | secret APD: 57.0 -> 23.6 | Improved |
| train with combined distortions vs no distortions (ablation) | UDH watermarking · no distortion | secret APD: 3.5 -> 9.6 | Hurt |
| decrease patch size (more hidden bits) (ablation) | UDH barcode hiding · JPEG | bit accuracy: 91.5% -> 60.2% | Hurt |
| UDH vs HiDDeN | watermarking · JPEG | bit accuracy: 63.0% -> 91.5% | Improved |
| UDH vs LFM | photographic steganography · unseen camera-display pairs | BER: 13.62% -> 4.41% | Improved |
| split mini-batch into groups with different distortions (ablation) | UDH training · combined distortions | convergence/performance | Improved |
Closest papers by meaning
- SSHR: More Secure Generative Steganography with High-Quality Revealed Secret Images (ICML 2025 · similarity 0.86)
- StegFormer: Rebuilding the Glory of Autoencoder-Based Steganography (AAAI 2024 · similarity 0.85)
- Purified and Unified Steganographic Network (CVPR 2024 · similarity 0.85)
- Neural Cover Selection for Image Steganography (NeurIPS 2024 · similarity 0.85)
- Robust Message Embedding via Attention Flow-Based Steganography (CVPR 2025 · similarity 0.84)
- Efficient and Separate Authentication Image Steganography Network (ICML 2025 · similarity 0.84)
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.
