Search papers, labs, and topics across Lattice.
Korea University
1
0
2
Aligning text with action-relevant video segments, TF-CADE significantly boosts performance in zero-shot action detection, outperforming existing methods.
