GS-Hider: Hiding Messages into 3D Gaussian Splatting

Authors: Xuanyu Zhang, Jiarui Meng, Runyi Li, Zhipei Xu, yongbing zhang, Jian Zhang

NeurIPS 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments demonstrated that the proposed GS-Hider can effectively conceal multimodal messages without compromising rendering quality and possesses exceptional security, robustness, capacity, and flexibility. Our project is available at: https://xuanyuzhang21. github.io/project/gshider/.
Researcher Affiliation Academia School of Electronic and Computer Engineering, Peking University 2 Guangdong Provincial Key Laboratory of Ultra High Definition Immersive Media Technology, Peking University Shenzhen Graduate School 3 School of Computer Science and Technology, Harbin Institute of Technology (Shenzhen)
Pseudocode No The paper describes its methods in text and diagrams but does not include structured pseudocode or algorithm blocks.
Open Source Code No Our project is available at: https://xuanyuzhang21. github.io/project/gshider/. This link leads to a project page, not directly to a source-code repository as strictly required.
Open Datasets Yes We conduct experiments on 9 original scenes taken from the public Mip-Ne RF360 dataset [2].
Dataset Splits No The paper mentions 'training views' but does not explicitly provide percentages, sample counts, or citations to predefined validation splits for the dataset used in its experiments.
Hardware Specification Yes We conduct all our experiments on a NVIDIA RTX 4090 Server.
Software Dependencies No The paper mentions 'CUDA rasterizer' but does not provide specific version numbers for it or any other software dependencies crucial for replication.
Experiment Setup Yes λ is set to 0.5 when hiding 3D scenes and set to 0.1 when hiding a single image. β and γ in Eq. 6 and Eq. 7 are respectively set to 0.2. The feature dimension M is set to 16.