HACS: Human Action Clips and Segments Dataset for Recognition and Temporal Localization.

Hang Zhao, Antonio Torralba, Lorenzo Torresani, Zhicheng Yan

Browse the full ICCV paper archive.