GeoState: Learning View-Decoupled Geometry States for 3D World Modeling

Hongchang Liu; Jian Liu; Yujuan Ding; Liang Peng; Xun Jiang; Fei Ma; Xing Xu; Yi Bin

cs.CVOct 15, 2026

GeoState: Learning View-Decoupled Geometry States for 3D World Modeling

Authors: Hongchang Liu, Jian Liu, Yujuan Ding, Liang Peng, Xun Jiang, Fei Ma, Xing Xu, Yi Bin

Organizations: Tongji University · Hong Kong Polytechnic University · Tencent · Guangdong Laboratory of Artificial Intelligence and Digital Economy (SZ)

Abstract

Recent 3D world models increasingly use features from geometry foundation models (GFMs) as predictive world states, benefiting from their rich geometric priors. However, the view dependence of these features entangles scene dynamics with camera-induced variation, making the resulting state evolution harder to predict. We propose GeoState, a view-decoupled geometry representation for predictive 3D world modeling that separates scene-state evolution from camera-dependent observation. Specifically, a scene tokenizer aggregates multi-view GFM features into a view-decoupled scene state, which a geometry decoder queries using camera rays. An autoregressive world model then predicts future scene states, while camera motion is modeled separately. The predicted states can therefore be decoded under either predicted or externally specified cameras, while retaining a single context-derived scale throughout the rollout without per-frame scale alignment. Experiments show that GeoState outperforms prior 3D world models on Waymo, supports geometrically consistent cross-view querying, and generalizes zero-shot to KITTI and nuScenes. Project page: https://hongchang-liu.github.io/GeoState-project-page/

Figures & tables

Figure 1: Overview of GeoState. GeoState Learning constructs view-decoupled scene states with a scene tokenizer and query-conditioned geometry decoder. GeoState World Modeling predicts future states directly in this representation space.

Figure 2: Qualitative multi-view depth forecasting. Red boxes highlight prediction differences.

Figure 3: Cross-query geometric consistency on Waymo. We evaluate symmetric reprojection AbsRel (

\downarrow

) across perturbation levels.

Method	$t{+}0.2$ s		$t{+}0.6$ s		$t{+}1.0$ s		$t{+}2.0$ s
	AbsRel $\downarrow$	$\delta_{1}\uparrow$	AbsRel $\downarrow$	$\delta_{1}\uparrow$	AbsRel $\downarrow$	$\delta_{1}\uparrow$	AbsRel $\downarrow$	$\delta_{1}\uparrow$
VGGT- $\Omega$	0.074	0.934	0.079	0.932	0.076	0.932	0.078	0.930
GeoState Reconstruction	0.068	0.934	0.073	0.932	0.072	0.932	0.072	0.929

Table 3: Geometric reconstruction on Waymo.

Table 5

Appendix figures & tables3 assets

Supplementary material from the paper’s appendix.

Appendix

Component	Configuration
Scene Tokenizer	512 queries of width 512; eight Perceiver layers; eight attention heads; MLP expansion ratio 4; 70.46M parameters
Geometry decoder	Width 512; two scene-conditioning blocks; 16 joint scene–ray blocks; four output branches with two Perceiver layers each; 135.55M parameters
World Model	Four-state context; 12 factorized blocks; width 1,152; eight heads; MLP expansion ratio 4; dropout 0.1; 256.55M parameters

Appendix

Table 6: Detailed architecture of the model.

Horizon (s)	ATE (m) $\downarrow$	RTE (m) $\downarrow$	RRE ( ∘ ) $\downarrow$
0.2	0.0125	0.0125	0.1894
0.6	0.0379	0.0246	0.2462
1.0	0.0840	0.0361	0.2783
2.0	0.3117	0.0785	0.3755

Appendix

Table 7: Camera prediction results on Waymo.

Quantity	Measure	15 vs. 33 images
GFM patch features
Layer-4 patch features	Relative L2	$1.38\pm 0.59\%$
Layer-11 patch features	Relative L2	$11.21\pm 1.96\%$
Layer-17 patch features	Relative L2	$33.55\pm 8.71\%$
Layer-23 patch features	Relative L2	$31.15\pm 5.97\%$
Four-layer mean	Relative L2	$\mathbf{19.32\pm 4.16\%}$

Appendix

Table 8: Context dependence of frozen VGGT-

\Omega

. All metrics come from the same image while varying only its jointly processed context. Values are sample mean

\pm

standard deviation.

World Model Learning

World model learning builds predictive models of how environments change under actions to support planning, simulation, and decision-making.

See all 399 papers

GeoWM: Efficient Direct World Modeling in Explicit Geometry
Oct 5, 2026Mehrdad Noori, Guile Wu, Sam Hosseini +1World Model Learning World Models for Robotics
Huawei Noah’s Ark Lab, Canada
VGGT-World: Transforming VGGT into an Autoregressive Geometry World Model
Mar 13, 2026Xiangyu Sun, Shijie Wang, Fengyi Zhang +53D Scene Prediction World Model Learning
UQMM Lab, The University of Queensland · Beijing Jiaotong University
GeoWorld: Providing Full-frame Geometry Features to Facilitate 3D Scene Generation
Nov 28, 2025Yuhao Wan, Lijuan Liu, Jingzhi Zhou +6Video Diffusion Models 3D Scene Generation
1VCIP & AAIS, Nankai University · 2ByteDance Inc. · 3Renmin University of China +2

Oct 5, 2026J/K move · Enter open · S save

GeoWM: Efficient Direct World Modeling in Explicit Geometry

Mehrdad Noori, Guile Wu, Sam Hosseini, Dongfeng Bai

Huawei Noah’s Ark Lab, Canada

World Model Learning World Models for Robotics World Models for Autonomous Driving

Open paper AskPDF arXiv

Modeling 3D scene geometry and its evolution over time is essential for autonomous driving and robotics. A common paradigm is to use world models to predict future images or latent representations of the environment and subsequently recover geometry from these predictions. However, this paradigm does not explicitly model geometric structure and typically relies on recursive rollouts to reach longer prediction horizons, leading to error accumulation and increasing computational cost. To address these limitations, we present GeoWM, a geometry world model that directly forecasts future scene geometry at specified future horizons without recursive rollout. The key idea is to leverage a geometry foundation model to transform observed RGB frames into a geometric history, which conditions a flow-matching transformer to predict the scene geometry at a specified future horizon. We further show that a lightweight camera-motion predictor can accurately estimate the future viewpoint, and that projecting the observed geometry into the predicted viewpoint provides an effective geometric prior for future geometry forecasting. Extensive experiments on four datasets spanning urban driving, aerial flight, and dynamic manipulation demonstrate that GeoWM outperforms the evaluated world models in forecasting depth, camera pose, and 3D scene geometry, while substantially reducing inference time at longer horizons.

Abstract

World Model Learning

Explore similar work

GeoWM: Efficient Direct World Modeling in Explicit Geometry

VGGT-World: Transforming VGGT into an Autoregressive Geometry World Model

GeoWorld: Providing Full-frame Geometry Features to Facilitate 3D Scene Generation