Research map / Learning theory and algorithms / Prediction and decision losses
StegaStamp: Invisible Hyperlinks in Physical Photographs
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Combine pixelwise and spatial training perturbations (ablation) | StegaStamp · Cellphone camera + consumer printer | mean bit accuracy: 0.89 -> 0.99 | Helped |
| Add spatial perturbations during training (ablation) | StegaStamp · Cellphone camera + consumer printer | mean bit accuracy: 0.51 -> 0.89 | Helped |
| Add pixelwise perturbations to spatial perturbations during training (ablation) | StegaStamp · Cellphone camera + consumer printer | mean bit accuracy: 0.89 -> 0.99 | Helped |
| Train with no image perturbations (ablation) | StegaStamp · Cellphone camera + consumer printer | mean bit accuracy: 0.49 | Hurt |
| Train with pixelwise perturbations only (ablation) | StegaStamp · Cellphone camera + consumer printer | mean bit accuracy: 0.51 | Hurt |
| Train with spatial perturbations only (ablation) | StegaStamp · Cellphone camera + consumer printer | mean bit accuracy: 0.89 | Hurt |
| Increase message length from 50 bits to 100 bits (ablation) | StegaStamp · ImageNet images | PSNR: 29.88 -> 28.50 | Hurt |
| Increase message length from 100 bits to 150 bits (ablation) | StegaStamp · ImageNet images | PSNR: 28.50 -> 26.47 | Hurt |
| Increase message length from 100 bits to 200 bits (ablation) | StegaStamp · ImageNet images | PSNR: 28.50 -> 21.79 | Hurt |
| Use StegaStamp rather than LFM for printed-image transmission | StegaStamp · Cellphone camera + consumer printer | mean bit accuracy: 0.61 -> 0.99 | Helped |
Closest papers by meaning
- Robust Message Embedding via Attention Flow-Based Steganography (CVPR 2025 · similarity 0.85)
- Physical Marker: Revealing Invisible Hyperlinks Hidden in Printed Trademarks (AAAI 2025 · similarity 0.84)
- StegoGAN: Leveraging Steganography for Non-Bijective Image-to-Image Translation (CVPR 2024 · similarity 0.83)
- Hiding Images in Diffusion Models by Editing Learned Score Functions (CVPR 2025 · similarity 0.82)
- StegFormer: Rebuilding the Glory of Autoencoder-Based Steganography (AAAI 2024 · similarity 0.82)
- SSHR: More Secure Generative Steganography with High-Quality Revealed Secret Images (ICML 2025 · similarity 0.82)
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.