News for hackers. Stuff that matters.

geevi

Re:Wav2Lip: Accurately Lip-Sync Videos to Any Speech on Sunday November 08, 2020 @11:22PM (#25031437)
Glad to see works from Indian AI researchers featured in Hackernews. There is an excellent AI community in India.

Unfortunately their works are overshadowed by some of the Ivy Leagues and Silicon Valley companies because they do not have that sort of publicity machine or brand name. Even within research comm unity their works dont get cited relative to the importance. If the same work comes from an Ivy League Fac or Company they get cited much more.

Re:Wav2Lip: Accurately Lip-Syncing Videos in the Wild on Sunday September 13, 2020 @11:15PM (#24466400)
Exciting work having application to Video Dubbing, creating social media memes!..

Some of the previous work by the same set of people: https://www.youtube.com/watch?v=HziA-jmlk_4

Actually an equally simple proof (if not simpler IMHO) was known many 1000s of years ago. See http://www.geom.uiuc.edu/~demo5337/Group3/Bhaskara.html
There is also an India Driving Dataset: http://idd.insaan.iiit.ac.in/