Saved in:
Bibliographic Details
Main Authors: Jiang, Haoyi, Liu, Liu, Cheng, Tianheng, Wang, Xinjie, Lin, Tianwei, Su, Zhizhong, Liu, Wenyu, Wang, Xinggang
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2412.13193
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915210836574208
author Jiang, Haoyi
Liu, Liu
Cheng, Tianheng
Wang, Xinjie
Lin, Tianwei
Su, Zhizhong
Liu, Wenyu
Wang, Xinggang
author_facet Jiang, Haoyi
Liu, Liu
Cheng, Tianheng
Wang, Xinjie
Lin, Tianwei
Su, Zhizhong
Liu, Wenyu
Wang, Xinggang
contents 3D Semantic Occupancy Prediction is fundamental for spatial understanding, yet existing approaches face challenges in scalability and generalization due to their reliance on extensive labeled data and computationally intensive voxel-wise representations. In this paper, we introduce GaussTR, a novel Gaussian-based Transformer framework that unifies sparse 3D modeling with foundation model alignment through Gaussian representations to advance 3D spatial understanding. GaussTR predicts sparse sets of Gaussians in a feed-forward manner to represent 3D scenes. By splatting the Gaussians into 2D views and aligning the rendered features with foundation models, GaussTR facilitates self-supervised 3D representation learning and enables open-vocabulary semantic occupancy prediction without requiring explicit annotations. Empirical experiments on the Occ3D-nuScenes dataset demonstrate GaussTR's state-of-the-art zero-shot performance of 12.27 mIoU, along with a 40% reduction in training time. These results highlight the efficacy of GaussTR for scalable and holistic 3D spatial understanding, with promising implications in autonomous driving and embodied agents. The code is available at https://github.com/hustvl/GaussTR.
format Preprint
id arxiv_https___arxiv_org_abs_2412_13193
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle GaussTR: Foundation Model-Aligned Gaussian Transformer for Self-Supervised 3D Spatial Understanding
Jiang, Haoyi
Liu, Liu
Cheng, Tianheng
Wang, Xinjie
Lin, Tianwei
Su, Zhizhong
Liu, Wenyu
Wang, Xinggang
Computer Vision and Pattern Recognition
3D Semantic Occupancy Prediction is fundamental for spatial understanding, yet existing approaches face challenges in scalability and generalization due to their reliance on extensive labeled data and computationally intensive voxel-wise representations. In this paper, we introduce GaussTR, a novel Gaussian-based Transformer framework that unifies sparse 3D modeling with foundation model alignment through Gaussian representations to advance 3D spatial understanding. GaussTR predicts sparse sets of Gaussians in a feed-forward manner to represent 3D scenes. By splatting the Gaussians into 2D views and aligning the rendered features with foundation models, GaussTR facilitates self-supervised 3D representation learning and enables open-vocabulary semantic occupancy prediction without requiring explicit annotations. Empirical experiments on the Occ3D-nuScenes dataset demonstrate GaussTR's state-of-the-art zero-shot performance of 12.27 mIoU, along with a 40% reduction in training time. These results highlight the efficacy of GaussTR for scalable and holistic 3D spatial understanding, with promising implications in autonomous driving and embodied agents. The code is available at https://github.com/hustvl/GaussTR.
title GaussTR: Foundation Model-Aligned Gaussian Transformer for Self-Supervised 3D Spatial Understanding
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2412.13193