VLM-3R：基於指令對齊三維重建增強的視覺-語言模型

VLM-3R: Vision-Language Models Augmented with Instruction-Aligned 3D Reconstruction

May 26, 2025

作者: Zhiwen Fan, Jian Zhang, Renjie Li, Junge Zhang, Runjin Chen, Hezhen Hu, Kevin Wang, Huaizhi Qu, Dilin Wang, Zhicheng Yan, Hongyu Xu, Justin Theiss, Tianlong Chen, Jiachen Li, Zhengzhong Tu, Zhangyang Wang, Rakesh Ranjan

cs.AI

摘要

大型多模态模型（LMMs）在二维图像和视频领域的快速发展，推动了这些模型向理解三维场景的延伸，旨在实现类人的视觉空间智能。然而，达到与人类能力相媲美的深度空间理解，在模型编码和数据获取方面仍面临重大挑战。现有方法常依赖外部深度传感器进行几何捕捉，或利用现成算法预先构建三维地图，这限制了其可扩展性，特别是在普遍的单目视频输入及对时间敏感的应用场景中。本研究提出了VLM-3R，一个融合三维重建指令调优的视觉语言模型（VLMs）统一框架。VLM-3R通过几何编码器处理单目视频帧，生成代表空间理解的隐式三维标记。借助我们的空间-视觉-视图融合技术及超过20万条精心策划的三维重建指令调优问答对，VLM-3R有效对齐了现实世界的空间语境与语言指令，实现了单目三维空间辅助与具身推理。为促进时间推理的评估，我们引入了视觉-空间-时间智能基准，包含超过13.86万条问答对，覆盖五个专注于空间关系演变的独特任务。大量实验表明，我们的模型VLM-3R不仅促进了稳健的视觉空间推理，还能理解三维语境的时间变化，在准确性和可扩展性上均表现出色。

English

The rapid advancement of Large Multimodal Models (LMMs) for 2D images and videos has motivated extending these models to understand 3D scenes, aiming for human-like visual-spatial intelligence. Nevertheless, achieving deep spatial understanding comparable to human capabilities poses significant challenges in model encoding and data acquisition. Existing methods frequently depend on external depth sensors for geometry capture or utilize off-the-shelf algorithms for pre-constructing 3D maps, thereby limiting their scalability, especially with prevalent monocular video inputs and for time-sensitive applications. In this work, we introduce VLM-3R, a unified framework for Vision-Language Models (VLMs) that incorporates 3D Reconstructive instruction tuning. VLM-3R processes monocular video frames by employing a geometry encoder to derive implicit 3D tokens that represent spatial understanding. Leveraging our Spatial-Visual-View Fusion and over 200K curated 3D reconstructive instruction tuning question-answer (QA) pairs, VLM-3R effectively aligns real-world spatial context with language instructions. This enables monocular 3D spatial assistance and embodied reasoning. To facilitate the evaluation of temporal reasoning, we introduce the Vision-Spatial-Temporal Intelligence benchmark, featuring over 138.6K QA pairs across five distinct tasks focused on evolving spatial relationships. Extensive experiments demonstrate that our model, VLM-3R, not only facilitates robust visual-spatial reasoning but also enables the understanding of temporal 3D context changes, excelling in both accuracy and scalability.