CAdam: Context-Adaptive Moment Estimation for 3D Gaussian Densification in Generative Distillation Conference: Special Interest Group on Computer Graphics and Interactive Techniques Conference Conference Papers; July 19–23, 2026; Los Angeles, CA, USA Special Interest Group on Computer Graphics and Interactive Techniques Conference Conference Papers (SIGGRAPH Conference Papers ’26), July 19–23, 2026, Los Angeles, CA, USA DOI: 10.1145/3799902.3811215 ISBN: 979-8-4007-2554-8/2026/07 CCS: Computing methodologies Rendering
SeungJeh Chung, Geonho Park, Misong Kim, HyeongYeop Kang · 2026
Read CAdam: Context-Adaptive Moment Estimation for 3D Gaussian Densification in Generative Distillation Conference: Special Interest Group on Computer Graphics and Interactive Techniques Conference Conference Papers; July 19–23, 2026; Los Angeles, CA, USA Special Interest Group on Computer Graphics and Interactive Techniques Conference Conference Papers (SIGGRAPH Conference Papers ’26), July 19–23, 2026, Los Angeles, CA, USA DOI: 10.1145/3799902.3811215 ISBN: 979-8-4007-2554-8/2026/07 CCS: Computing methodologies Rendering by SeungJeh Chung, Geonho Park, Misong Kim, HyeongYeop Kang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-20872.txt.
Description
SeungJeh Chung and three colleagues propose CAdam, a context-adaptive moment estimation method for 3D Gaussian densification. Adaptive thickening drives 3D Gaussian Splatting, but in generative distillation it yields cluttered, extra parts. The authors diagnose a dilemma: magnitude-based summing gathers transient noise with geometric signals. That blurs the line between over-densification and under-fitting. CAdam reframes this process as signal checking, using the first moment of gradients to exploit interference cancellation.
Catalog record
| CREATOR | SeungJeh Chung, Geonho Park, Misong Kim, HyeongYeop Kang |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 39819 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.20872