Search papers, labs, and topics across Lattice.
The paper introduces Think, Act, Build (TAB), a novel agentic framework that reformulates 3D visual grounding as a generative 2D-to-3D reconstruction problem using raw RGB-D streams and VLMs. TAB uses a VLM agent to dynamically track and reconstruct the target across 2D frames, guided by a 3D-VG skill and visual tools. To address multi-view coverage deficits, they propose Semantic-Anchored Geometric Expansion, which anchors the target in a reference clip and propagates its location using multi-view geometry.
Zero-shot 3D visual grounding can be achieved by decoupling spatial semantics (resolved by 2D VLMs) from 3D structure instantiation (handled by deterministic multi-view geometry), outperforming even fully supervised methods.
3D Visual Grounding (3D-VG) aims to localize objects in 3D scenes via natural language descriptions. While recent advancements leveraging Vision-Language Models (VLMs) have explored zero-shot possibilities, they typically suffer from a static workflow relying on preprocessed 3D point clouds, essentially degrading grounding into proposal matching. To bypass this reliance, our core motivation is to decouple the task: leveraging 2D VLMs to resolve complex spatial semantics, while relying on deterministic multi-view geometry to instantiate the 3D structure. Driven by this insight, we propose"Think, Act, Build (TAB)", a dynamic agentic framework that reformulates 3D-VG tasks as a generative 2D-to-3D reconstruction paradigm operating directly on raw RGB-D streams. Specifically, guided by a specialized 3D-VG skill, our VLM agent dynamically invokes visual tools to track and reconstruct the target across 2D frames. Crucially, to overcome the multi-view coverage deficit caused by strict VLM semantic tracking, we introduce the Semantic-Anchored Geometric Expansion, a mechanism that first anchors the target in a reference video clip and then leverages multi-view geometry to propagate its spatial location across unobserved frames. This enables the agent to"Build"the target's 3D representation by aggregating these multi-view features via camera parameters, directly mapping 2D visual cues to 3D coordinates. Furthermore, to ensure rigorous assessment, we identify flaws such as reference ambiguity and category errors in existing benchmarks and manually refine the incorrect queries. Extensive experiments on ScanRefer and Nr3D demonstrate that our framework, relying entirely on open-source models, significantly outperforms previous zero-shot methods and even surpasses fully supervised baselines.
