Struct2D: A Perception-Guided Framework for Spatial Reasoning in Large Multimodal Models

• URL: http://arxiv.org/abs/2506.04220v1
• Date: Wed, 04 Jun 2025 17:58:04 GMT
• Title: Struct2D: A Perception-Guided Framework for Spatial Reasoning in Large Multimodal Models
• Authors: Fangrui Zhu, Hanhui Wang, Yiming Xie, Jing Gu, Tianye Ding, Jianwei Yang, Huaizu Jiang,
• Abstract summary: Struct2D is a perception-guided prompting framework that combines bird's-eye-view (BEV) images with object marks and object-centric metadata.<n>We find that LMMs exhibit surprisingly strong spatial reasoning abilities when provided with structured 2D inputs.<n>We fine-tune an open-source LMM (Qwen2.5VL) on Struct2D-Set, achieving competitive performance on multiple benchmarks.
• Score: 24.897120944471347
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Unlocking spatial reasoning in Large Multimodal Models (LMMs) is crucial for enabling intelligent interaction with 3D environments. While prior efforts often rely on explicit 3D inputs or specialized model architectures, we ask: can LMMs reason about 3D space using only structured 2D representations derived from perception? We introduce Struct2D, a perception-guided prompting framework that combines bird's-eye-view (BEV) images with object marks and object-centric metadata, optionally incorporating egocentric keyframes when needed. Using Struct2D, we conduct an in-depth zero-shot analysis of closed-source LMMs (e.g., GPT-o3) and find that they exhibit surprisingly strong spatial reasoning abilities when provided with structured 2D inputs, effectively handling tasks such as relative direction estimation and route planning. Building on these insights, we construct Struct2D-Set, a large-scale instruction tuning dataset with 200K fine-grained QA pairs across eight spatial reasoning categories, generated automatically from 3D indoor scenes. We fine-tune an open-source LMM (Qwen2.5VL) on Struct2D-Set, achieving competitive performance on multiple benchmarks, including 3D question answering, dense captioning, and object grounding. Our approach demonstrates that structured 2D inputs can effectively bridge perception and language reasoning in LMMs-without requiring explicit 3D representations as input. We will release both our code and dataset to support future research.

Related papers

• 3D-MOOD: Lifting 2D to 3D for Monocular Open-Set Object Detection [58.78881632019072]
  We introduce the first end-to-end 3D Monocular Open-set Object Detector (3D-MOOD)<n>We lift the open-set 2D detection into 3D space through our designed 3D bounding box head.<n>We condition the object queries with geometry prior and overcome the generalization for 3D estimation across diverse scenes.
  arXiv  Detail & Related papers  (2025-07-31T13:56:41Z)
• E3D-Bench: A Benchmark for End-to-End 3D Geometric Foundation Models [78.1674905950243]
  We present the first comprehensive benchmark for 3D geometric foundation models (GFMs)<n>GFMs directly predict dense 3D representations in a single feed-forward pass, eliminating the need for slow or unavailable precomputed camera parameters.<n>We evaluate 16 state-of-the-art GFMs, revealing their strengths and limitations across tasks and domains.<n>All code, evaluation scripts, and processed data will be publicly released to accelerate research in 3D spatial intelligence.
  arXiv  Detail & Related papers  (2025-06-02T17:53:09Z)
• seg_3D_by_PC2D: Multi-View Projection for Domain Generalization and Adaptation in 3D Semantic Segmentation [2.4549463031236396]
  3D semantic segmentation plays a pivotal role in autonomous driving and road infrastructure analysis.<n>We propose a novel multi-view projection framework that excels in both domain generalization (DG) and unsupervised domain adaptation (UDA)<n>We achieve state-of-the-art results in UDA and close to state-of-the-art in DG, with particularly large gains on large, static classes.
  arXiv  Detail & Related papers  (2025-05-21T14:08:42Z)
• xMOD: Cross-Modal Distillation for 2D/3D Multi-Object Discovery from 2D motion [4.878192303432336]
  DIOD-3D is the first baseline for multi-object discovery in 3D data using 2D motion.<n>xMOD is a cross-modal training framework that integrates 2D and 3D data while always using 2D motion cues.<n>Our approach yields a substantial performance improvement compared with the 2D object discovery state-of-the-art on all datasets.
  arXiv  Detail & Related papers  (2025-03-19T09:20:35Z)
• MM-Spatial: Exploring 3D Spatial Understanding in Multimodal LLMs [13.678235444299286]
  Multimodal large language models (MLLMs) excel at 2D visual understanding but remain limited in their ability to reason about 3D space.<n>In this work, we leverage large-scale high-quality 3D scene data with open-set annotations to introduce 1) a novel supervised fine-tuning dataset and 2) a new evaluation benchmark, focused on indoor scenes.
  arXiv  Detail & Related papers  (2025-03-17T12:34:22Z)
• GEAL: Generalizable 3D Affordance Learning with Cross-Modal Consistency [50.11520458252128]
  Existing 3D affordance learning methods struggle with generalization and robustness due to limited annotated data.<n>We propose GEAL, a novel framework designed to enhance the generalization and robustness of 3D affordance learning by leveraging large-scale pre-trained 2D models.<n>GEAL consistently outperforms existing methods across seen and novel object categories, as well as corrupted data.
  arXiv  Detail & Related papers  (2024-12-12T17:59:03Z)
• 3DSRBench: A Comprehensive 3D Spatial Reasoning Benchmark [15.33844114999498]
  3D spatial reasoning is the ability to analyze and interpret the positions, orientations, and spatial relationships of objects within the 3D space.<n>Large multi-modal models (LMMs) have achieved remarkable progress in a wide range of image and video understanding tasks.<n>We present the first comprehensive 3D spatial reasoning benchmark, 3DSRBench, with 2,772 manually annotated visual question-answer pairs.
  arXiv  Detail & Related papers  (2024-12-10T18:55:23Z)
• MMScan: A Multi-Modal 3D Scene Dataset with Hierarchical Grounded Language Annotations [55.022519020409405]
  This paper builds the first largest ever multi-modal 3D scene dataset and benchmark with hierarchical grounded language annotations, MMScan. The resulting multi-modal 3D dataset encompasses 1.4M meta-annotated captions on 109k objects and 7.7k regions as well as over 3.04M diverse samples for 3D visual grounding and question-answering benchmarks.
  arXiv  Detail & Related papers  (2024-06-13T17:59:30Z)
• Language-Image Models with 3D Understanding [59.499585515469974]
  We develop a large-scale pre-training dataset for 2D and 3D called LV3D. Next, we introduce a new MLLM named Cube-LLM and pre-train it on LV3D. We show that pure data scaling makes a strong 3D perception capability without 3D specific architectural design or training objective.
  arXiv  Detail & Related papers  (2024-05-06T17:57:27Z)
• Fully Sparse Fusion for 3D Object Detection [69.32694845027927]
  Currently prevalent multimodal 3D detection methods are built upon LiDAR-based detectors that usually use dense Bird's-Eye-View feature maps. Fully sparse architecture is gaining attention as they are highly efficient in long-range perception. In this paper, we study how to effectively leverage image modality in the emerging fully sparse architecture.
  arXiv  Detail & Related papers  (2023-04-24T17:57:43Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.