AnyScene: Towards Highly Controllable Driving Scene Generation at Anywhere and Beyond
Researchers have introduced several new frameworks for generating realistic and controllable driving scenes, crucial for training autonomous vehicles. DriveWAM adapts video diffusion transformers to create autoregressive action policies, incorporating scene understanding and memory for long-horizon planning. AnyScene offers a unified occupancy-centric approach, enabling precise control from arbitrary BEV layouts and generating temporally consistent multi-view videos. DriveGen3D combines efficient video diffusion with 3D scene reconstruction for high-quality, controllable dynamic scenes, supporting long driving videos and 3D representations. Additionally, a new dataset, Nuplan-Occ, has been curated to facilitate large-scale generative modeling and downstream applications in autonomous driving. AI
IMPACT These advancements in synthetic data generation could accelerate the development and testing of autonomous driving systems by providing more realistic and controllable training environments.