DynImg represents a video snippet as a keyframe plus four resized neighboring frames as temporal prompts, with a 4D rotary position embedding, and reports improved video QA accuracy.
Cores: Orchestrating the dance of reasoning and seg- mentation
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
DynImg: Key Frames with Visual Prompts are Good Representation for Multi-Modal Video Understanding
DynImg represents a video snippet as a keyframe plus four resized neighboring frames as temporal prompts, with a 4D rotary position embedding, and reports improved video QA accuracy.