Search papers, labs, and topics across Lattice.
The paper introduces CoordRefer, a novel framework for 3D visual grounding that separates the processes of coordinate frame selection and box regression to mitigate coordinate-relative box ambiguity. By first establishing a reference frame and then conditioning the 3D box prediction on this frame, the method enhances grounding performance significantly. Experiments on the ScanRefer dataset demonstrate that CoordRefer outperforms existing coordinate-agnostic approaches, achieving notable accuracy improvements of 11% at Acc@0.25 and 7% at Acc@0.5.
Decoupling coordinate frame selection from box regression leads to a remarkable 11% accuracy boost in 3D visual grounding tasks.
Multiview image-based 3D visual grounding predicts a coordinate frame to define a coordinate system and then regresses a 3D bounding box for localization. However, existing methods jointly optimize coordinate frame selection and box regression, leading to coordinate-relative box ambiguity and degraded grounding performance. This ambiguity arises because the same box admits different numerical representations across coordinate frames, creating multiple optimization targets and yielding invalid compromise predictions. To tackle this challenge, we propose CoordRefer, a coordinate-aware framework that decouples coordinate frame selection from coordinate-conditioned grounding. CoordRefer first selects a reference frame to define the coordinate system and then conditions 3D box prediction on the coordinate system. We perform coordinate-aware supervised fine-tuning to establish coordinate frame selection and coordinate-conditioned box regression, followed by Group Relative Policy Optimization with 3D IoU-based rewards to align both stages with downstream grounding quality. On ScanRefer with Qwen3-VL-2B, CoordRefer achieves gains of 11% in Acc@0.25 and 7% in Acc@0.5 over the coordinate-agnostic baseline, while its geometrically refined variant surpasses methods using explicit 3D inputs.
