Spatial 3D-LLM: Progressive Spatial Awareness for Advanced 3D Vision-Language Understanding

Xiaoyan Wang, Zeju Li, Yifan Xu, Jiaxing Qi, Zhifei Yang, Ruifei Ma, Xiangde Liu, Chao Zhang

Keywords: 3D-MLLM, 3D scene understanding

Abstract:

New era has unlocked exciting possibilities for extending Large Language Models (LLMs) to tackle 3D vision-language tasks. However, most existing 3D Multimodal LLMs (MLLMs) rely on holistic 3D scene information or specifically designated regions for 3D vision-language tasks, failing to capture multi-level location-based information.
Addressing these concerns, we present Spatial 3D-LLM, a 3D MLLM specifically designed to enhance spatial perception and reasoning for 3D vision-language tasks by enriching the spatial embeddings of 3D scenes.
Spatial 3D-LLM incorporates an LLM backbone and a meticulously designed progressive spatial awareness scheme that captures spatial information as the perception field expands, generating location-enriched 3D scene embeddings that serve as visual prompt.
Additionally, we introduce two novel tasks, namely 3D object distance measurement and 3D layout editing, and construct a 3D instruction dataset MODEL, to inspire more profound 3D spatial perception capabilities.
Experimental results demonstrate that Spatial 3D-LLM achieves state-of-the-art performance across a wide range of 3D vision-language tasks, revealing the improvements stemmed from our progressive spatial awareness scheme of mining more profound spatial information and the proposed dataset.

Primary Area: foundation or frontier models, including LLMs

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 8974