Hiding Images in Deep Probabilistic Models
Authors: Haoyu Chen, Linqi Song, Zhenxing Qian, Xinpeng Zhang, Kede Ma
NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In this section, we perform a series of experiments to verify the promise of our Sin GAN approach. First, we evaluate secret image extraction accuracy both quantitatively and qualitatively in comparison to image-in-image hiding methods based on autoencoders. Second, we probe the security of the stego Sin GAN by comparing it to the original one in terms of 1) quality and diversity of generated cover images, 2) marginal distribution similarity of model parameters [56], and 3) possibility of secret image leakage. |
| Researcher Affiliation | Academia | Haoyu Chen Department of Computer Science City University of Hong Kong haoychen3-c@my.cityu.edu.hk; Linqi Song Department of Computer Science City University of Hong Kong linqi.song@cityu.edu.hk; Zhenxing Qian School of Computer Science Fudan University zxqian@fudan.edu.cn; Xinpeng Zhang School of Computer Science Fudan University zhangxinpeng@fudan.edu.cn; Kede Ma Department of Computer Science City University of Hong Kong kede.ma@cityu.edu.hk |
| Pseudocode | No | The paper describes the methodology using text and mathematical equations but does not include structured pseudocode or algorithm blocks. |
| Open Source Code | Yes | 3. If you ran experiments... (a) Did you include the code, data, and instructions needed to reproduce the main experimental results (either in the supplemental material or as a URL)? [Yes] |
| Open Datasets | No | The paper mentions "a cover image dataset D" and "training cover images" but does not provide specific access information (link, DOI, repository, formal citation with authors/year) for a publicly available or open dataset used for training in the main text. |
| Dataset Splits | No | The paper does not provide specific dataset split information (exact percentages, sample counts, citations to predefined splits, or detailed splitting methodology) for validation sets in the main text. |
| Hardware Specification | Yes | 3. If you ran experiments... (d) Did you include the total amount of compute and the type of resources used (e.g., type of GPUs, internal cluster, or cloud provider)? [Yes] In the appendix. |
| Software Dependencies | No | The paper mentions using PyTorch in a referenced implementation and specific loss functions, but it does not provide specific version numbers for key software components or libraries in the main text. |
| Experiment Setup | Yes | 3. If you ran experiments... (b) Did you specify all the training details (e.g., data splits, hyperparameters, how they were chosen)? [Yes] In supplemental material. |