Guardado en:
Detalles Bibliográficos
Autores principales: Zhu, Yiyao, Xue, Ying, Zhang, Haiming, Jiang, Guangfeng, Zhou, Wending, Yan, Xu, Gao, Jiantao, Cai, Yingjie, Liu, Bingbing, Li, Zhen, Shen, Shaojie
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:https://arxiv.org/abs/2604.00969
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866917377747189760
author Zhu, Yiyao
Xue, Ying
Zhang, Haiming
Jiang, Guangfeng
Zhou, Wending
Yan, Xu
Gao, Jiantao
Cai, Yingjie
Liu, Bingbing
Li, Zhen
Shen, Shaojie
author_facet Zhu, Yiyao
Xue, Ying
Zhang, Haiming
Jiang, Guangfeng
Zhou, Wending
Yan, Xu
Gao, Jiantao
Cai, Yingjie
Liu, Bingbing
Li, Zhen
Shen, Shaojie
contents Vision-based autonomous driving has gained much attention due to its low costs and excellent performance. Compared with dense BEV (Bird's Eye View) or sparse query models, Gaussian-centric method is a comprehensive yet sparse representation by describing scene with 3D semantic Gaussians. In this paper, we introduce DLWM, a novel paradigm with Dual Latent World Models specifically designed to enable holistic gaussian-centric pre-training in autonomous driving using two stages. In the first stage, DLWM predicts 3D Gaussians from queries by self-supervised reconstructing multi-view semantic and depth images. Equipped with fine-grained contextual features, in the second stage, two latent world models are trained separately for temporal feature learning, including Gaussian-flow-guided latent prediction for downstream occupancy perception and forecasting tasks, and ego-planning-guided latent prediction for motion planning. Extensive experiments in SurroundOcc and nuScenes benchmarks demonstrate that DLWM shows significant performance gains across Gaussian-centric 3D occupancy perception, 4D occupancy forecasting and motion planning tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2604_00969
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle DLWM: Dual Latent World Models enable Holistic Gaussian-centric Pre-training in Autonomous Driving
Zhu, Yiyao
Xue, Ying
Zhang, Haiming
Jiang, Guangfeng
Zhou, Wending
Yan, Xu
Gao, Jiantao
Cai, Yingjie
Liu, Bingbing
Li, Zhen
Shen, Shaojie
Computer Vision and Pattern Recognition
Vision-based autonomous driving has gained much attention due to its low costs and excellent performance. Compared with dense BEV (Bird's Eye View) or sparse query models, Gaussian-centric method is a comprehensive yet sparse representation by describing scene with 3D semantic Gaussians. In this paper, we introduce DLWM, a novel paradigm with Dual Latent World Models specifically designed to enable holistic gaussian-centric pre-training in autonomous driving using two stages. In the first stage, DLWM predicts 3D Gaussians from queries by self-supervised reconstructing multi-view semantic and depth images. Equipped with fine-grained contextual features, in the second stage, two latent world models are trained separately for temporal feature learning, including Gaussian-flow-guided latent prediction for downstream occupancy perception and forecasting tasks, and ego-planning-guided latent prediction for motion planning. Extensive experiments in SurroundOcc and nuScenes benchmarks demonstrate that DLWM shows significant performance gains across Gaussian-centric 3D occupancy perception, 4D occupancy forecasting and motion planning tasks.
title DLWM: Dual Latent World Models enable Holistic Gaussian-centric Pre-training in Autonomous Driving
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2604.00969