Research map / Architectures and efficiency / RNNs and normalization
Steganography of Steganographic Networks
Read the paperCheck 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 |
|---|---|---|---|
| Disguise a secret DNN as a functional model for an ordinary task, then recover it using side information and a key | HiDDeN secret decoder · HiDDeN secret decoder recovery | BER (×10^-5): 6.68 -> 6.68 | No clear effect |
| Disguise the secret model while preserving its selected filters | ResNet18 · Fashion-MNIST classification | ACC (%): 93.99 -> 93.80 | Hurt |
| Disguise the secret model while preserving its selected filters | U-Net · ImageNet denoising | PSNR (dB): 28.98 -> 28.92 | Hurt |
| Disguise the secret decoder as a GTSRB classifier | HiDDeN secret decoder · GTSRB classification | ACC (%): 99.97 -> 99.81 | Hurt |
| Disguise ResNet18 as a CIFAR10 classifier | ResNet18 · CIFAR10 classification | ACC (%): 93.46 -> 93.02 | Hurt |
| Disguise U-Net as an Oxford-Pet segmenter | U-Net · Oxford-Pet segmentation | mIOU (%): 87.68 -> 87.05 | Hurt |
| Use the proposed model-disguising scheme instead of Jing et al.'s image-based embedding | HiDDeN secret decoder · HiDDeN secret decoder recovery | BER (%): 50.13 -> 6.68 $\times 10^{-3}$ | Helped |
| Use the proposed model-disguising scheme instead of Jing et al.'s image-based embedding | ResNet18 · ResNet18 recovery on Fashion-MNIST | ACC (%): 10.00 -> 93.80 | Helped |
| Hide the model in a functional stego DNN | Proposed method · Stego DNN detection | Detection accuracy (%): 47.50 (SVM); 50.00 (MLP) | No clear effect |
Closest papers by meaning
- Purified and Unified Steganographic Network (CVPR 2024 · similarity 0.86)
- Hiding Images in Diffusion Models by Editing Learned Score Functions (CVPR 2025 · similarity 0.85)
- StegoGAN: Leveraging Steganography for Non-Bijective Image-to-Image Translation (CVPR 2024 · similarity 0.85)
- StegaStyleGAN: Towards Generic and Practical Generative Image Steganography (AAAI 2024 · similarity 0.84)
- Neural Cover Selection for Image Steganography (NeurIPS 2024 · similarity 0.83)
- SSHR: More Secure Generative Steganography with High-Quality Revealed Secret Images (ICML 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.