search

filby89 / spectre

Official Pytorch Implementation of SPECTRE: Visual Speech-Aware Perceptual 3D Facial Expression Reconstruction from Videos
Other
236 stars 21 forks source link

Type of image input #21

Closed EvelynQIN closed 4 months ago

EvelynQIN commented 5 months ago

Hello, from your repo, the image inputs are within the value of [0, 255], but in DECA, they divide 255 in dataset preprocessing, right?

filby89 commented 4 months ago

Hey, we use the same range as DECA [0,1). This division takes place in the warp function in the dataloader (https://scikit-image.org/docs/stable/api/skimage.transform.html#skimage.transform.warp).