how did you get the data in train.zip？

[wudidecc]

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

[jixinya]

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.