StreetCrafter: Street View Synthesis with Controllable Video Diffusion Models

TL;DR

StreetCrafter introduces a controllable video diffusion model leveraging LiDAR data for photorealistic, flexible, and editable view synthesis in autonomous driving scenes, outperforming existing methods.

Contribution

The paper presents StreetCrafter, a novel diffusion-based approach that uses pixel-level LiDAR conditions for improved view synthesis and scene editing in autonomous driving environments.

Findings

Enables flexible viewpoint control and scene editing.

Outperforms existing view synthesis methods.

Supports real-time rendering with dynamic scenes.

Abstract

This paper aims to tackle the problem of photorealistic view synthesis from vehicle sensor data. Recent advancements in neural scene representation have achieved notable success in rendering high-quality autonomous driving scenes, but the performance significantly degrades as the viewpoint deviates from the training trajectory. To mitigate this problem, we introduce StreetCrafter, a novel controllable video diffusion model that utilizes LiDAR point cloud renderings as pixel-level conditions, which fully exploits the generative prior for novel view synthesis, while preserving precise camera control. Moreover, the utilization of pixel-level LiDAR conditions allows us to make accurate pixel-level edits to target scenes. In addition, the generative prior of StreetCrafter can be effectively incorporated into dynamic scene representations to achieve real-time rendering. Experiments on Waymo Open Dataset and PandaSet demonstrate that our model enables flexible control over viewpoint changes, enlarging the view synthesis regions for satisfying rendering, which outperforms existing methods.