A CLIP-scored, Gumbel-Max frame sampler with per-frame multi-resolution allocation improves long-video question answering in Video-LLMs under a fixed token budget.
Robust motion-guided frame sam- pler with interpretive evaluation for video action recognition
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Q-Frame: Query-aware Frame Selection and Multi-Resolution Adaptation for Video-LLMs
A CLIP-scored, Gumbel-Max frame sampler with per-frame multi-resolution allocation improves long-video question answering in Video-LLMs under a fixed token budget.