AutoScape: Geometry-Consistent Long-Horizon Scene Generation
Publication Date: 10/19/2025
Event: ICCV 2025
Reference: pp. 25700-25711, 2025
Authors: Jiacheng Chen, NEC Laboratories America, Inc., Simon Fraser University; Ziyu Jiang, NEC Laboratories America, Inc.; Mingfu Liang, NEC Laboratories America, Inc., Northwestern University; Bingbing Zhuang, NEC Laboratories America, Inc.; Jong-Chyi Su, NEC Laboratories America, Inc.; Sparsh Garg, NEC Laboratories America, Inc.; Ying Wu, NEC Laboratories America, Inc., Northwestern University; Manmohan Chandraker, NEC Laboratories America, Inc., UC San Diego
Abstract: This paper proposes AutoScape, a long-horizon driving scene generation framework. At its core is a novel RGB-D diffusion model that iteratively generates sparse, geometrically consistent keyframes, serving as reliable anchors for the scenes appearance and geometry. To maintain long-range geometric consistency, the model 1) jointly handles image and depth in a shared latent space, 2) explicitly conditions on the existing scene geometry (i.e., rendered point clouds) from previously generated keyframes, and 3) steers the sampling process with a warp-consistent guidance. Given high-quality RGB-D keyframes, a video diffusion model then interpolates between them to produce dense nd coherent video frames. AutoScape generates realistic and geometrically consistent driving videos of over 20 seconds, improving the long-horizon FID and FVD scores over the prior state-of-the-art by 48.6% and 43.0%, respectively.
Project page: https://auto-scape.github.io.
Publication Link: https://openaccess.thecvf.com/content/ICCV2025/html/Chen_AutoScape_Geometry-Consistent_Long-Horizon_Scene_Generation_ICCV_2025_paper.html
Additional Publication Link: https://arxiv.org/pdf/2510.20726
