[Paper] Generation Models Know Space: Unleashing Implicit 3D Priors for Scene Understanding
Source: arXiv - 2603.19235v1
Overview
While Multimodal Large Language Models demonstrate impressive semantic capabilities, they often suffer from spatial blindness, struggling with fine-grained geometric reasoning and physical dynamics. Existing solutions typically rely on explicit 3D modalities or complex geometric scaffolding, which are limited by data scarcity and generalization challenges. In this work, we propose a paradigm shift by leveraging the implicit spatial prior within large-scale video generation models. We posit that to synthesize temporally coherent videos, these models inherently learn robust 3D structural priors and physical laws. We introduce VEGA-3D (Video Extracted Generative Awareness), a plug-and-play framework that repurposes a pre-trained video diffusion model as a Latent World Simulator. By extracting spatiotemporal features from intermediate noise levels and integrating them with semantic representations via a token-level adaptive gated fusion mechanism, we enrich MLLMs with dense geometric cues without explicit 3D supervision. Extensive experiments across 3D scene understanding, spatial reasoning, and embodied manipulation benchmarks demonstrate that our method outperforms state-of-the-art baselines, validating that generative priors provide a scalable foundation for physical-world understanding. Code is publicly available at https://github.com/H-EmbodVis/VEGA-3D.
Key Contributions
This paper presents research in the following areas:
- cs.CV
- cs.RO
Methodology
Please refer to the full paper for detailed methodology.
Practical Implications
This research contributes to the advancement of cs.CV.
Authors
- Xianjin Wu
- Dingkang Liang
- Tianrui Feng
- Kui Xia
- Yumeng Zhang
- Xiaofan Li
- Xiao Tan
- Xiang Bai
Paper Information
- arXiv ID: 2603.19235v1
- Categories: cs.CV, cs.RO
- Published: March 19, 2026
- PDF: Download PDF