ReFoCUS: Reinforcement-guided Frame Optimization for Contextual Understanding

Hosu Lee, Junho Kim, Hyunjun Kim, Yong Man Ro

Keywords: Video: Action and Event Understanding