XgTu / 2DASL

The code (pytorch for testing & matlab for 3D plot and evaluation) for our project: Joint 3D Face Reconstruction and Dense Face Alignment from A Single Image with 2D-Assisted Self-Supervised Learning (2DASL)
MIT License
467 stars 116 forks source link

How to get the forth channel of input data? #8

Closed xinghuokang closed 5 years ago

xinghuokang commented 5 years ago

Hi, I want to test my own image, but input of model is four channels, and how to get the forth channel of my own data? Thanks

XgTu commented 5 years ago

@xinghuokang The additional channel is a binary-value image, where the locations corresponding to facial landmarks take the value of 1 and others take the value of −1

xinghuokang commented 5 years ago

Thank you for your reply, how to get facial landmarks of my own image, by others libarary?

XgTu commented 5 years ago

@xinghuokang You can use mtcnn or 2D FAN or dlib to obtain the landmarks.

xinghuokang commented 5 years ago

Thanks