VR-Drive: Viewpoint-Robust End-to-End Driving with Feed-Forward 3D Gaussian Splatting

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Cho, Hoonhee, Kang, Jae-Young, Lee, Giwon, Yang, Hyemin, Park, Heejun, Jung, Seokwoo, Yoon, Kuk-Jin
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866914116321411072
author Cho, Hoonhee
Kang, Jae-Young
Lee, Giwon
Yang, Hyemin
Park, Heejun
Jung, Seokwoo
Yoon, Kuk-Jin
author_facet Cho, Hoonhee
Kang, Jae-Young
Lee, Giwon
Yang, Hyemin
Park, Heejun
Jung, Seokwoo
Yoon, Kuk-Jin
contents End-to-end autonomous driving (E2E-AD) has emerged as a promising paradigm that unifies perception, prediction, and planning into a holistic, data-driven framework. However, achieving robustness to varying camera viewpoints, a common real-world challenge due to diverse vehicle configurations, remains an open problem. In this work, we propose VR-Drive, a novel E2E-AD framework that addresses viewpoint generalization by jointly learning 3D scene reconstruction as an auxiliary task to enable planning-aware view synthesis. Unlike prior scene-specific synthesis approaches, VR-Drive adopts a feed-forward inference strategy that supports online training-time augmentation from sparse views without additional annotations. To further improve viewpoint consistency, we introduce a viewpoint-mixed memory bank that facilitates temporal interaction across multiple viewpoints and a viewpoint-consistent distillation strategy that transfers knowledge from original to synthesized views. Trained in a fully end-to-end manner, VR-Drive effectively mitigates synthesis-induced noise and improves planning under viewpoint shifts. In addition, we release a new benchmark dataset to evaluate E2E-AD performance under novel camera viewpoints, enabling comprehensive analysis. Our results demonstrate that VR-Drive is a scalable and robust solution for the real-world deployment of end-to-end autonomous driving systems.
format Preprint
id arxiv_https___arxiv_org_abs_2510_23205
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle VR-Drive: Viewpoint-Robust End-to-End Driving with Feed-Forward 3D Gaussian Splatting
Cho, Hoonhee
Kang, Jae-Young
Lee, Giwon
Yang, Hyemin
Park, Heejun
Jung, Seokwoo
Yoon, Kuk-Jin
Computer Vision and Pattern Recognition
End-to-end autonomous driving (E2E-AD) has emerged as a promising paradigm that unifies perception, prediction, and planning into a holistic, data-driven framework. However, achieving robustness to varying camera viewpoints, a common real-world challenge due to diverse vehicle configurations, remains an open problem. In this work, we propose VR-Drive, a novel E2E-AD framework that addresses viewpoint generalization by jointly learning 3D scene reconstruction as an auxiliary task to enable planning-aware view synthesis. Unlike prior scene-specific synthesis approaches, VR-Drive adopts a feed-forward inference strategy that supports online training-time augmentation from sparse views without additional annotations. To further improve viewpoint consistency, we introduce a viewpoint-mixed memory bank that facilitates temporal interaction across multiple viewpoints and a viewpoint-consistent distillation strategy that transfers knowledge from original to synthesized views. Trained in a fully end-to-end manner, VR-Drive effectively mitigates synthesis-induced noise and improves planning under viewpoint shifts. In addition, we release a new benchmark dataset to evaluate E2E-AD performance under novel camera viewpoints, enabling comprehensive analysis. Our results demonstrate that VR-Drive is a scalable and robust solution for the real-world deployment of end-to-end autonomous driving systems.
title VR-Drive: Viewpoint-Robust End-to-End Driving with Feed-Forward 3D Gaussian Splatting
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2510.23205