SSR: Pushing the Limit of Spatial Intelligence with Structured Scene Reasoning

TL;DR

SSR introduces a novel framework for structured scene reasoning that effectively integrates 2D and 3D representations, enabling advanced spatial reasoning and achieving state-of-the-art results on spatial intelligence benchmarks.

Contribution

The paper presents a lightweight alignment mechanism and a scene graph generation pipeline that enhance spatial reasoning without extensive pre-training or large-scale alignment.

Findings

Achieves 73.9 on VSI-Bench, outperforming larger models.

Effectively integrates 2D and 3D data for spatial reasoning.

Demonstrates state-of-the-art performance on multiple benchmarks.

Abstract

While Multimodal Large Language Models (MLLMs) excel in semantic tasks, they frequently lack the "spatial sense" essential for sophisticated geometric reasoning. Current models typically suffer from exorbitant modality-alignment costs and deficiency in fine-grained structural modeling precision.We introduce SSR, a framework designed for Structured Scene Reasoning that seamlessly integrates 2D and 3D representations via a lightweight alignment mechanism. To minimize training overhead, our framework anchors 3D geometric features to the large language model's pre-aligned 2D visual semantics through cross-modal addition and token interleaving, effectively obviating the necessity for large-scale alignment pre-training. To underpin complex spatial reasoning, we propose a novel scene graph generation pipeline that represents global layouts as a chain of independent local triplets defined by relative coordinates. This is complemented by an incremental generation algorithm, enabling the model to construct "language-model-friendly" structural scaffolds for complex environments. Furthermore, we extend these capabilities to global-scale 3D global grounding task, achieving absolute metric precision across heterogeneous data sources. At a 7B parameter scale, SSR achieves state-of-the-art performance on multiple spatial intelligence benchmarks, notably scoring 73.9 on VSI-Bench. Our approach significantly outperforms much larger models, demonstrating that efficient feature alignment and structured scene reasoning are the cornerstones of authentic spatial intelligence.