Lip-2Wav

The Lip2Wav dataset is a benchmark used to evaluate speaker-dependent Lip-to-Speech synthesis by providing silent videos of talking faces paired with corresponding audio waveforms.

Papers using Lip-2Wav (3)

Robustl2s: Speaker-specific Lip-to-speech Synthesis Exploiting Self-supervised Representations2023 · 3 citesOn the Audio-visual Synchronization for Lip-to-Speech Synthesis2023RobustL2S: Speaker-Specific Lip-to-Speech Synthesis exploiting Self-Supervised Representations2023