Audio-Visual Active Speaker Extraction for Sparsely Overlapped Multi-talker Speech

Publication
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), South Korea, 2024

Related