Sketch, Ground, and Refine: Top-Down Dense Video Captioning

Publication
CVPR