Search papers, labs, and topics across Lattice.
1
0
3
By using text as an anchor to filter noisy audio-visual signals, TAEMI achieves state-of-the-art emotional mimicry intensity estimation, demonstrating that a stable semantic prior can significantly improve multimodal fusion robustness.
