ZERO-SHOT 3D VISUAL GROUNDING

Summary from the TLDR AI newsletter

VLM-Grounder is a new method for 3D visual grounding that overcomes the limitations of traditional approaches by using vision-language models (VLMs) and 2D images.