Learning Visual Proxy for Compositional Zero-Shot Learning
Published in ArXiv.org • Jan 23, 2025
NobleIDNI3P77W35R26S53
Authors:,,
Shiyu Zhang
Cheng Yan
Yang Liu
Abstract
Compositional Zero-Shot Learning (CZSL) aims to recognize novel attribute-object compositions by leveraging knowledge from seen compositions. Current methods align textual prototypes with visual features via Vision-Language Models (VLMs), but suffer from two limitations: (1) modality gaps hinder the...
Finding related papers...
Discussions
(0)No comments yet
Be the first to share your thoughts!