AutoScape: Geometry-Consistent Long-Horizon Scene Generation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chen, Jiacheng, Jiang, Ziyu, Liang, Mingfu, Zhuang, Bingbing, Su, Jong-Chyi, Garg, Sparsh, Wu, Ying, Chandraker, Manmohan
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914109521395712
author Chen, Jiacheng
Jiang, Ziyu
Liang, Mingfu
Zhuang, Bingbing
Su, Jong-Chyi
Garg, Sparsh
Wu, Ying
Chandraker, Manmohan
author_facet Chen, Jiacheng
Jiang, Ziyu
Liang, Mingfu
Zhuang, Bingbing
Su, Jong-Chyi
Garg, Sparsh
Wu, Ying
Chandraker, Manmohan
contents 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 scene's 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 and 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.
format Preprint
id arxiv_https___arxiv_org_abs_2510_20726
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle AutoScape: Geometry-Consistent Long-Horizon Scene Generation
Chen, Jiacheng
Jiang, Ziyu
Liang, Mingfu
Zhuang, Bingbing
Su, Jong-Chyi
Garg, Sparsh
Wu, Ying
Chandraker, Manmohan
Computer Vision and Pattern Recognition
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 scene's 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 and 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.
title AutoScape: Geometry-Consistent Long-Horizon Scene Generation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2510.20726