CueKFS: Agentic Cue-Driven Keyframe Selection for Long Video Understanding
A training-free method that reformulates question--frame matching as comparing frames against a set of dynamically generated visual cues, showing that agentic cue refinement drives active re-exploration of the video, yielding relative similarity gains of up to 92% over the initial context.
W. Kang, Hanieh Deilamsalehy, Yumo Xu et al.
· 0 citations