Enregistré dans:
Détails bibliographiques
Auteurs principaux: Wang, Jiamin, Yao, Yichen, Feng, Xiang, Wu, Hang, Wang, Yaming, Huang, Qingqiu, Ma, Yuexin, Zhu, Xinge
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2506.13138
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866914014898946048
author Wang, Jiamin
Yao, Yichen
Feng, Xiang
Wu, Hang
Wang, Yaming
Huang, Qingqiu
Ma, Yuexin
Zhu, Xinge
author_facet Wang, Jiamin
Yao, Yichen
Feng, Xiang
Wu, Hang
Wang, Yaming
Huang, Qingqiu
Ma, Yuexin
Zhu, Xinge
contents The generation of temporally consistent, high-fidelity driving videos over extended horizons presents a fundamental challenge in autonomous driving world modeling. Existing approaches often suffer from error accumulation and feature misalignment due to inadequate decoupling of spatio-temporal dynamics and limited cross-frame feature propagation mechanisms. To address these limitations, we present STAGE (Streaming Temporal Attention Generative Engine), a novel auto-regressive framework that pioneers hierarchical feature coordination and multi-phase optimization for sustainable video synthesis. To achieve high-quality long-horizon driving video generation, we introduce Hierarchical Temporal Feature Transfer (HTFT) and a novel multi-stage training strategy. HTFT enhances temporal consistency between video frames throughout the video generation process by modeling the temporal and denoising process separately and transferring denoising features between frames. The multi-stage training strategy is to divide the training into three stages, through model decoupling and auto-regressive inference process simulation, thereby accelerating model convergence and reducing error accumulation. Experiments on the Nuscenes dataset show that STAGE has significantly surpassed existing methods in the long-horizon driving video generation task. In addition, we also explored STAGE's ability to generate unlimited-length driving videos. We generated 600 frames of high-quality driving videos on the Nuscenes dataset, which far exceeds the maximum length achievable by existing methods.
format Preprint
id arxiv_https___arxiv_org_abs_2506_13138
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle STAGE: A Stream-Centric Generative World Model for Long-Horizon Driving-Scene Simulation
Wang, Jiamin
Yao, Yichen
Feng, Xiang
Wu, Hang
Wang, Yaming
Huang, Qingqiu
Ma, Yuexin
Zhu, Xinge
Computer Vision and Pattern Recognition
The generation of temporally consistent, high-fidelity driving videos over extended horizons presents a fundamental challenge in autonomous driving world modeling. Existing approaches often suffer from error accumulation and feature misalignment due to inadequate decoupling of spatio-temporal dynamics and limited cross-frame feature propagation mechanisms. To address these limitations, we present STAGE (Streaming Temporal Attention Generative Engine), a novel auto-regressive framework that pioneers hierarchical feature coordination and multi-phase optimization for sustainable video synthesis. To achieve high-quality long-horizon driving video generation, we introduce Hierarchical Temporal Feature Transfer (HTFT) and a novel multi-stage training strategy. HTFT enhances temporal consistency between video frames throughout the video generation process by modeling the temporal and denoising process separately and transferring denoising features between frames. The multi-stage training strategy is to divide the training into three stages, through model decoupling and auto-regressive inference process simulation, thereby accelerating model convergence and reducing error accumulation. Experiments on the Nuscenes dataset show that STAGE has significantly surpassed existing methods in the long-horizon driving video generation task. In addition, we also explored STAGE's ability to generate unlimited-length driving videos. We generated 600 frames of high-quality driving videos on the Nuscenes dataset, which far exceeds the maximum length achievable by existing methods.
title STAGE: A Stream-Centric Generative World Model for Long-Horizon Driving-Scene Simulation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2506.13138