AI Paper Atlas

Research map / 3D vision / Gaussian splatting

CompMarkGS: Robust Watermarking for Compressed 3D Gaussian Splatting

Sumin In, Youngdong Jang, Utae Jeong, MinHyuk Jang, Hyeongcheol Park, Eunbyung Park, Sangpil Kim

ICLR 2026 · 3D vision › Gaussian splatting · approach: Regularization and primitive design for 3DGS

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 changedWhereMeasuredEffect
Add quantization distortion layer (QDL) (ablation)CompMarkGS · HAC after compressionbit accuracy (%): 90.75 -> 95.92Improved
Add quantization distortion layer (QDL) (ablation)CompMarkGS · ContextGS after compressionbit accuracy (%): 88.09 -> 94.03Improved
Add frequency-aware anchor growing (FAG) (ablation)CompMarkGS · HAC after compressionbit accuracy (%): 93.95 -> 95.92Improved
Add frequency-aware anchor growing (FAG) (ablation)CompMarkGS · ContextGS after compressionPSNR: 27.15 -> 27.55Improved
Add HSV loss (ablation)CompMarkGS · HAC after compressionbit accuracy (%): 92.57 -> 95.92Improved
Add HSV loss (ablation)CompMarkGS · HAC after compressionLPIPS: 0.179 -> 0.177Improved
Embed watermark in anchor feature instead of position (ablation)CompMarkGS · HAC after compressionPSNR: 12.05 -> 27.65Improved
Embed watermark in anchor feature instead of offsets (ablation)CompMarkGS · HAC after compressionPSNR: 24.61 -> 27.65Improved
Add dedicated watermark embedding feature f' (ablation)CompMarkGS · ContextGS after compressionPSNR: 26.79 -> 27.55Improved
Increase message length from 32 to 64 bitsCompMarkGS · before compressionbit accuracy (%): 96.52 -> 91.29Hurt
Use CompMarkGS instead of 3D-GSWanchor-based 3DGS watermark · ContextGSPSNR: 19.83 -> 27.55Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 4 on most 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

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy