Saved in:
Bibliographic Details
Main Authors: Zhao, Lingjun, Luo, Yandong, Hays, James, Gan, Lu
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.03370
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914462370365440
author Zhao, Lingjun
Luo, Yandong
Hays, James
Gan, Lu
author_facet Zhao, Lingjun
Luo, Yandong
Hays, James
Gan, Lu
contents We introduce ShelfGaussian, an open-vocabulary multi-modal Gaussian-based 3D scene understanding framework supervised by off-the-shelf vision foundation models (VFMs). Gaussian-based methods have demonstrated superior performance and computational efficiency across a wide range of scene understanding tasks. However, existing methods either model objects as closed-set semantic Gaussians supervised by annotated 3D labels, neglecting their rendering ability, or learn open-set Gaussian representations via purely 2D self-supervision, leading to degraded geometry and limited to camera-only settings. To fully exploit the potential of Gaussians, we propose a Multi-Modal Gaussian Transformer that enables Gaussians to query features from diverse sensor modalities, and a Shelf-Supervised Learning Paradigm that efficiently optimizes Gaussians with VFM features jointly at 2D image and 3D scene levels. We evaluate ShelfGaussian on various perception and planning tasks. Experiments on Occ3D-nuScenes demonstrate its state-of-the-art zero-shot semantic occupancy prediction performance. ShelfGaussian is further evaluated on an unmanned ground vehicle (UGV) to assess its in the-wild performance across diverse urban scenarios. Project website: https://lunarlab-gatech.github.io/ShelfGaussian/.
format Preprint
id arxiv_https___arxiv_org_abs_2512_03370
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle ShelfGaussian: Shelf-Supervised Open-Vocabulary Gaussian-based 3D Scene Understanding
Zhao, Lingjun
Luo, Yandong
Hays, James
Gan, Lu
Computer Vision and Pattern Recognition
We introduce ShelfGaussian, an open-vocabulary multi-modal Gaussian-based 3D scene understanding framework supervised by off-the-shelf vision foundation models (VFMs). Gaussian-based methods have demonstrated superior performance and computational efficiency across a wide range of scene understanding tasks. However, existing methods either model objects as closed-set semantic Gaussians supervised by annotated 3D labels, neglecting their rendering ability, or learn open-set Gaussian representations via purely 2D self-supervision, leading to degraded geometry and limited to camera-only settings. To fully exploit the potential of Gaussians, we propose a Multi-Modal Gaussian Transformer that enables Gaussians to query features from diverse sensor modalities, and a Shelf-Supervised Learning Paradigm that efficiently optimizes Gaussians with VFM features jointly at 2D image and 3D scene levels. We evaluate ShelfGaussian on various perception and planning tasks. Experiments on Occ3D-nuScenes demonstrate its state-of-the-art zero-shot semantic occupancy prediction performance. ShelfGaussian is further evaluated on an unmanned ground vehicle (UGV) to assess its in the-wild performance across diverse urban scenarios. Project website: https://lunarlab-gatech.github.io/ShelfGaussian/.
title ShelfGaussian: Shelf-Supervised Open-Vocabulary Gaussian-based 3D Scene Understanding
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.03370