arXiv - CSCL: "Lip2Vec: Efficient and Robust Visual Speech Recog…" - Qoto Mastodon

arXiv - CSCL @arxiv_cscl@qoto.org

Lip2Vec: Efficient and Robust Visual Speech Recognition via Latent-to-Latent Visual to Audio Representation Mapping. (arXiv:2308.06112v1 [cs.SD])

http://arxiv.org/abs/2308.06112 #arXiv #NLProc

Aug 14, 2023, 03:17 · · arxiv-cscl · · ·

Sign in to participate in the conversation