AI RESEARCH
GroundVTS: Visual Token Sampling in Multimodal Large Language Models for Video Temporal Grounding
arXiv CS.CV
•
ArXi:2604.02093v1 Announce Type: new Video temporal grounding (VTG) is a critical task in video understanding and a key capability for extending video large language models (Vid-LLMs) to broader applications. However, existing Vid-LLMs rely on uniform frame sampling to extract video information, resulting in a sparse distribution of key frames and the loss of crucial temporal cues. To address this limitation, we propose Grounded Visual Token Sampling (GroundVTS), a Vid-LLM architecture that focuses on the most informative temporal segments.