Research map / Image generation / Autoregressive and latent generation
Training-Free Coverless Multi-Image Steganography with Access Control
Minyeol Bae, Si-Hyeon Lee
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 |
|---|---|---|---|
| use Random Basis plus Latent Vector Fusion instead of concatenation with Noise Flip (MIDAS vs DiffStega*) | training-free coverless steganography · Stego260 (N=2) stego diversity | PSNR between stego and secret (lower is more diverse): 17.065 -> 9.885 | Improved |
| use MIDAS instead of DiffStega* for access control | multi-image hiding · Stego260 (N=2) wrong-key reconstruction | PSNR (lower is better): 18.137 -> 9.964 | Improved |
| increase number of secret images N from 1 to 4 for DiffStega* | DiffStega* · Stego260 | MANIQA: 0.450 -> 0.364 | Hurt |
| increase number of secret images N from 1 to 4 for MIDAS | MIDAS · Stego260 | MANIQA: 0.429 -> 0.479 | Improved |
| MIDAS vs modification-based IIS under Gaussian noise | access-controlled steganography · Stego260 (N=2) robustness | PSNR (dB): 12.439 -> 20.046 | Improved |
| MIDAS vs DiffStega* under Gaussian noise | robustness · Stego260 (N=2) | PSNR (dB): 20.076 -> 20.046 | No clear effect |
| MIDAS vs modification-based IIS/AIS for reconstruction on clean images | reconstruction quality · Stego260 (N=2) | PSNR (correct key): 41.360 (IIS) -> 23.903 | Hurt |
| MIDAS vs CRoSS*/DiffStega* in steganalysis resistance | stego images · XuNet and SiaStegNet accuracy | steganalysis accuracy: about 20% lower than baselines | Improved |
| use Random Basis strength gamma 0.4 in security analysis (ablation) | Random Basis · Stego260 | information leakage | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- CRoSS* — CRoSS: Diffusion Model Makes Controllable, Robust and Secure Image Steganography (NeurIPS 2023) · better on 5 of 5 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- SSHR: More Secure Generative Steganography with High-Quality Revealed Secret Images (ICML 2025 · similarity 0.85)
- Hiding Images in Diffusion Models by Editing Learned Score Functions (CVPR 2025 · similarity 0.85)
- Efficient and Separate Authentication Image Steganography Network (ICML 2025 · similarity 0.84)
- Neural Cover Selection for Image Steganography (NeurIPS 2024 · similarity 0.84)
- StegFormer: Rebuilding the Glory of Autoencoder-Based Steganography (AAAI 2024 · similarity 0.83)
- PID: Prompt-Independent Data Protection Against Latent Diffusion Models (ICML 2024 · 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.
