OmniGF: A Dual-Branch Vision-Language Framework for Unified Gaze Following 文章

ArXiv CS.CV2026-05-27NEWSen作者: Qiaomu Miao, Haoyu Wu, Jingyi Xu, Minh Hoai, Dimitris Samaras
大语言模型
查看原文 →
关系图谱

OmniGF: A Dual-Branch Vision-Language Framework for Unified Gaze Following · 相关技术

相关技术

ODElanguage modelUnifiedTeraStanScalaSSESPASEMReferring expression comprehension (REC)Parts-of-Speech (POS) tagsMITVision-Language ModelsVLMICTHISGranular Alignment ParadigmForFOLEffort Metric AttentionDiTDAPTCISARGANN