Search papers, labs, and topics across Lattice.
3
0
4
0
Avatar V achieves unprecedented avatar video fidelity by directly conditioning on full video references, capturing dynamic behaviors that previous methods missed.
Ditch the static image: this method generates realistic talking avatars by learning from *videos* of the subject in completely different scenes.
Injecting optical flow into VLMs lets them spot subtle video transitions that other methods miss, opening the door to more robust video understanding.
