VCSE: Time-Domain Visual-Contextual Speaker Extraction Network.

Junjie Li, Meng Ge, Zexu Pan, Longbiao Wang, Jianwu Dang

Browse the full Interspeech paper archive.