Research map / 3D vision / Gaussian splatting
All That Glitters Is Not Gold: Key-Secured 3D Secrets within 3D Gaussian Splatting
Yan Ren, Shilin Lu, Adams Wai-Kin Kong
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 |
|---|---|---|---|
| update only opacity vs also rotation, scale and position (ablation) | KeySS · Mip-NeRF360 hiding | composite score: 42.717 -> 44.389 | Improved |
| add SH update to the four-feature combination (ablation) | KeySS · Mip-NeRF360 hiding | secret PSNR / score: 26.427 -> 20.961 (score 44.389 -> 34.810) | Hurt |
| update only SH vs only opacity (ablation) | KeySS · Mip-NeRF360 hiding | secret PSNR: 26.138 -> 12.530 | Hurt |
| add rotation to opacity update (ablation) | KeySS · Mip-NeRF360 hiding | score: 42.717 -> 43.008 | Improved |
| add position to opacity update (ablation) | KeySS · Mip-NeRF360 hiding | score: 42.717 -> 31.795 | Hurt |
| combine cover and secret SfM points for initialization (ablation) | KeySS · Mip-NeRF360 hiding | avg PSNR: cover-only 26.275 vs combined 26.384 | Mixed |
| use combined densification instead of cover-only (ablation) | KeySS · Mip-NeRF360 hiding | avg PSNR: 26.249 -> 26.384 | Improved |
| add wrong-key loss for incorrect keys (ablation) | KeySS · wrong key attack | PSNR vs secret: 10.341 vs secret (25.916 vs cover) | Improved |
| replace GS-Hider decoder pipeline with KeySS | GS-Hider · Mip-NeRF360 hiding | cover PSNR: 25.817 -> 25.980 | Improved |
| replace GS-Hider with KeySS | GS-Hider · Mip-NeRF360 hiding | SSIM and LPIPS (cover): 0.782 -> 0.726; 0.246 -> 0.330 | Hurt |
| use standard 3DGS format instead of GS-Hider decoder rendering | KeySS · rendering speed | FPS: 45 -> 130 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- GS-Hider — GS-Hider: Hiding Messages into 3D Gaussian Splatting (NeurIPS 2024) · better on 3 of 6 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
- SecureGS: Boosting the Security and Fidelity of 3D Gaussian Splatting Steganography (ICLR 2025 · similarity 0.94)
- GS-Hider: Hiding Messages into 3D Gaussian Splatting (NeurIPS 2024 · similarity 0.92)
- Geometry Cloak: Preventing TGS-based 3D Reconstruction from Copyrighted Images (NeurIPS 2024 · similarity 0.89)
- GaussianMarker: Uncertainty-Aware Copyright Protection of 3D Gaussian Splatting (NeurIPS 2024 · similarity 0.89)
- GuardSplat: Efficient and Robust Watermarking for 3D Gaussian Splatting (CVPR 2025 · similarity 0.89)
- 3D-GSW: 3D Gaussian Splatting for Robust Watermarking (CVPR 2025 · similarity 0.88)
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.
