Learning frame-level affinity with video-level labels for weakly supervised temporal action detection

Bairong Li, Yuesheng Zhu, Ruixin Liu, Zhenyu Weng