bbaaii/hfa-gp

Official implementation of High-fidelity Facial Avatar Reconstruction from Monocular Video with Generative Priors

hfa-gp's Issues

There is no training json files.

Hi!

Thank you so much for the brilliant work and for sharing the code.

I want to run your model. I tried audio-driven method.
But, in training, there are no cropping_params.json and transforms_train.json.
How can I make these files according to my custom data? or Can you share these files?

Thank you for your consideration.

3DMM Features

Hi!

Thank you so much for the brilliant work and for sharing the code.

May I ask if it's possible to also share the preprocessing code for extracting 3DMM features that are used to train the model on 3DMM and for 3DMM based reenactment?

Many thanks.

Cross-person renactment

Great work! How would you suggest doing cross-person renactment (i.e. face, expression and pose of person A into video of person B)

Dataset used for training and data prep

A bit confused on what dataset & methods did you use for training?

You mentioned in the "Preprocessing data" part that we should run "process_test_video.py" but then in the training you mention to just take nerfface dataset?

If we use nerfface dataset, do we need to run preprocessing?

About code release

Great work! I wonder when you will release the code, I think it can inspire more future works!