Return
Progressive language-guided visual learning for multi-task visual grounding
DOI:10.1016/j.patcog.2026.115117.png)
Abstract
En 中文
• Progressive language guidance enriches visual features with low overhead.
• Shared representations and grouped feature routing support joint REC and RES.
• PLVL achieves competitive REC and RES performance across benchmark datasets.
Keywords:
Multi-modality
Multi-task visual grounding
Progressive learning
Journal
IF:
7.6
Papers:
1.4W
Citations:
6.1W
Organization
Cited Papers
No cited papers available

