LRS2
PulseAugur coverage of LRS2 — every cluster mentioning LRS2 across labs, papers, and developer communities, ranked by signal.
1 day(s) with sentiment data
-
FLARE framework enhances lipreading with cross-modal distillation and joint face-lip modeling
Researchers have developed FLARE, a new framework designed to improve visual speech representation learning for lipreading. FLARE utilizes cross-modal distillation, leveraging audio representations from Whisper as a tea…
-
New WYS framework enhances video-to-speech synthesis with textual conditioning
Researchers have developed a new video-to-speech synthesis framework called Watch Your Speech (WYS) to address the challenge of visual ambiguity in generating speech from silent talking-face videos. WYS incorporates tex…
-
New VSR method uses head pose to improve accuracy
Researchers have developed a new framework called HP-VSR-ResFiLM to improve visual speech recognition (VSR) by explicitly incorporating head-pose information. This method uses a pose-conditioned residual Feature-wise Li…