Search papers, labs, and topics across Lattice.
1
0
3
16
Forget training: SemanticMoments achieves state-of-the-art motion-based video retrieval by simply computing temporal statistics over features from pre-trained semantic models.
