GeoV2V: Geometry-Grounded Video Diffusion Model for Driving Scene Generation

Key points

• Novel view synthesis under trajectory changes is essential for autonomous driving simulation, yet existing methods struggle to generate consistent videos when extrapolating to unseen viewpoints such as lane shifts.1
• Reconstruction-based approaches maintain geometric consistency but degrade in visual quality under large viewpoint changes, while reconstruct-then-restore methods often produce visually plausible frames yet suffer from geometric and temporal inconsistencies.1
• A key challenge in learning cross-lane transformations is the lack of synchronized multi-trajectory supervision in real-world datasets.1
• To address this, we also introduce Para4D, a synthetic dataset capturing synchronized multi-lane driving videos within dynamic scenes, which provides explicit supervision for lane-shift generation and serves as a strong training prior for geometry-consistent view synthesis.1

Sources (1)

• [1]GeoV2V: Geometry-Grounded Video Diffusion Model for Driving Scene Generation
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 18, 01:05 PM

  We present GeoV2V, a geometry-grounded video-to-video diffusion framework that synthesizes temporally coherent driving videos along novel trajectories by integrating LiDAR and depth-derived geometric priors with full-video conditioning in the Wan 2.1 backbone.

  Novel view synthesis under trajectory changes is essential for autonomous driving simulation, yet existing methods struggle to generate consistent videos when extrapolating to unseen viewpoints such as lane shifts.

Extractive summary: sentences quoted from the sources.