search

jixinya / EVP

Code for paper 'Audio-Driven Emotional Video Portraits'.
301 stars 49 forks source link

how did you get the data in train.zip？ #24

Open wudidecc opened 1 year ago

wudidecc commented 1 year ago

Thank you for your excellent work, but how did you get the data in train.zip? Looking forward to your reply

jixinya commented 1 year ago

Hi, we use a landmark detector to detect the 106 facial key points of each frame. However, we can not provide the detector here due to copyright reasons. You can replace it with other open-source detectors.