Voxel-informed Language Grounding
This addresses the challenge of grounding natural language in 3D scenes for tasks like object reference, though it is incremental as it builds on existing methods with specific geometric enhancements.
The paper tackles the problem of language grounding in 3D contexts by introducing VLG, a model that uses voxel maps from volumetric reconstruction, resulting in a 2.0% absolute improvement and SOTA performance on the SNARE benchmark.
Natural language applied to natural 2D images describes a fundamentally 3D world. We present the Voxel-informed Language Grounder (VLG), a language grounding model that leverages 3D geometric information in the form of voxel maps derived from the visual input using a volumetric reconstruction model. We show that VLG significantly improves grounding accuracy on SNARE, an object reference game task. At the time of writing, VLG holds the top place on the SNARE leaderboard, achieving SOTA results with a 2.0% absolute improvement.