Home | Publications | LWL+26

SuperGSeg: Open-Vocabulary 3D Segmentation With Structured Super-Gaussians

MCML Authors

Abstract

3D Gaussian Splatting has recently gained traction for its efficient training and real-time rendering. While its vanilla representation is mainly designed for view synthesis, recent works extended it to scene understanding with language features. However, storing additional high-dimensional features per Gaussian for semantic information is memory-intensive, which limits their ability to segment and interpret challenging scenes. To this end, we introduce SuperGSeg, a novel approach that fosters cohesive, context-aware hierarchical scene representation by disentangling segmentation and language field distillation. SuperGSeg first employs neural 3D Gaussians to learn geometry, instance and hierarchical segmentation features from multi-view images with the aid of off-the-shelf 2D masks. These features are then leveraged to create a sparse set of Super-Gaussians. Super-Gaussians facilitate the lifting and distillation of 2D language features into 3D space. They enable hierarchical scene understanding with high-dimensional language feature rendering at moderate GPU memory costs. Extensive experiments demonstrate that SuperGSeg achieves remarkable performance on both open-vocabulary object selection and semantic segmentation tasks. More results at supergseg.github.io.

inproceedings LWL+26

3DV 2026

13th International Conference on 3D Vision. Vancouver, Canada, Mar 20-23, 2026. To be published. Preprint available.

Authors

S. Liang • S. Wang • K. Li • M. Niemeyer • S. Gasperini • H. P. A. Lensch • N. Navab • F. Tombari

Links

URL GitHub

In Collaboration

 Google
 VisualAIs Labs

Research Area

 C1 | Medicine

BibTeXKey: LWL+26

#p-navab #p-tombari
Back to Top