Z3D: Zero-Shot 3D Visual Grounding from Images
This addresses the problem of localizing objects in 3D scenes using natural language queries for applications in robotics and AR/VR, representing a novel method for a known bottleneck.
The paper tackles zero-shot 3D visual grounding from multi-view images without geometric supervision, achieving state-of-the-art performance on benchmarks like ScanRefer and Nr3D.
3D visual grounding (3DVG) aims to localize objects in a 3D scene based on natural language queries. In this work, we explore zero-shot 3DVG from multi-view images alone, without requiring any geometric supervision or object priors. We introduce Z3D, a universal grounding pipeline that flexibly operates on multi-view images while optionally incorporating camera poses and depth maps. We identify key bottlenecks in prior zero-shot methods causing significant performance degradation and address them with (i) a state-of-the-art zero-shot 3D instance segmentation method to generate high-quality 3D bounding box proposals and (ii) advanced reasoning via prompt-based segmentation, which utilizes full capabilities of modern VLMs. Extensive experiments on the ScanRefer and Nr3D benchmarks demonstrate that our approach achieves state-of-the-art performance among zero-shot methods. Code is available at https://github.com/col14m/z3d .