Saved in:
Bibliographic Details
Main Authors: Song, Qi, Li, Chenghong, Lin, Haotong, Peng, Sida, Huang, Rui
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2504.00437
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912895381536768
author Song, Qi
Li, Chenghong
Lin, Haotong
Peng, Sida
Huang, Rui
author_facet Song, Qi
Li, Chenghong
Lin, Haotong
Peng, Sida
Huang, Rui
contents We present a novel approach, termed ADGaussian, for generalizable street scene reconstruction. The proposed method enables high-quality rendering from merely single-view input. Unlike prior Gaussian Splatting methods that primarily focus on geometry refinement, we emphasize the importance of joint optimization of image and depth features for accurate Gaussian prediction. To this end, we first incorporate sparse LiDAR depth as an additional input modality, formulating the Gaussian prediction process as a joint learning framework of visual information and geometric clue. Furthermore, we propose a Multi-modal Feature Matching strategy coupled with a Multi-scale Gaussian Decoding model to enhance the joint refinement of multi-modal features, thereby enabling efficient multi-modal Gaussian learning. Extensive experiments on Waymo and KITTI demonstrate that our ADGaussian achieves state-of-the-art performance and exhibits superior zero-shot generalization capabilities in novel-view shifting.
format Preprint
id arxiv_https___arxiv_org_abs_2504_00437
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle ADGaussian: Generalizable Gaussian Splatting for Autonomous Driving via Multi-modal Joint Learning
Song, Qi
Li, Chenghong
Lin, Haotong
Peng, Sida
Huang, Rui
Computer Vision and Pattern Recognition
We present a novel approach, termed ADGaussian, for generalizable street scene reconstruction. The proposed method enables high-quality rendering from merely single-view input. Unlike prior Gaussian Splatting methods that primarily focus on geometry refinement, we emphasize the importance of joint optimization of image and depth features for accurate Gaussian prediction. To this end, we first incorporate sparse LiDAR depth as an additional input modality, formulating the Gaussian prediction process as a joint learning framework of visual information and geometric clue. Furthermore, we propose a Multi-modal Feature Matching strategy coupled with a Multi-scale Gaussian Decoding model to enhance the joint refinement of multi-modal features, thereby enabling efficient multi-modal Gaussian learning. Extensive experiments on Waymo and KITTI demonstrate that our ADGaussian achieves state-of-the-art performance and exhibits superior zero-shot generalization capabilities in novel-view shifting.
title ADGaussian: Generalizable Gaussian Splatting for Autonomous Driving via Multi-modal Joint Learning
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2504.00437