About this document
CVPR24 Laso by lyc071719 is a document available to read on EtoBox.
The document introduces Language-guided Affordance Segmentation on 3D Object (LASO), a novel task aimed at segmenting parts of 3D objects based on affordance-related questions, addressing limitations in existing models that primarily focus on visual aspects. It presents a dataset of 19,751 point-question pairs and proposes a baseline model, PointRefer, which utilizes language cues to enhance segmentation accuracy. The goal is to improve the integration of 3D affordance knowledge with large language models,
- Author
- lyc071719
- Language
- EN