ReCamDriving: LiDAR-Free Camera-Controlled Novel Trajectory Video Generation

TL;DR

ReCamDriving is a novel vision-based framework that generates controllable, high-quality videos of driving trajectories without LiDAR, using dense 3D scene guidance and a two-stage training process, supported by a large new dataset.

Contribution

It introduces a two-stage training paradigm and a 3DGS-based data curation strategy, enabling scalable multi-trajectory video generation from monocular videos.

Findings

Achieves state-of-the-art camera controllability.

Maintains high structural consistency in generated videos.

Constructed the ParaDrive dataset with over 110K trajectory pairs.

Abstract

We propose ReCamDriving, a purely vision-based, camera-controlled novel-trajectory video generation framework. While repair-based methods fail to restore complex artifacts and LiDAR-based approaches rely on sparse and incomplete cues, ReCamDriving leverages dense and scene-complete 3DGS renderings for explicit geometric guidance, achieving precise camera-controllable generation. To mitigate overfitting to restoration behaviors when conditioned on 3DGS renderings, ReCamDriving adopts a two-stage training paradigm: the first stage uses camera poses for coarse control, while the second stage incorporates 3DGS renderings for fine-grained viewpoint and geometric guidance. Furthermore, we present a 3DGS-based cross-trajectory data curation strategy to eliminate the train-test gap in camera transformation patterns, enabling scalable multi-trajectory supervision from monocular videos. Based on this strategy, we construct the ParaDrive dataset, containing over 110K parallel-trajectory video pairs. Extensive experiments demonstrate that ReCamDriving achieves state-of-the-art camera controllability and structural consistency.