TachibanaYoshino/AnimeGANv3

Dataset Release

Atharva-Phatak opened this issue · 5 comments

Atharva-Phatak commented

Thanks for the work. Will you releasing the dataset used to train the models ?

TachibanaYoshino commented

I think I'll release the dataset, but this will be decided at least after the paper is accepted by the conference.
The paper describes the concrete implementation and experimental data of my algorithm. It's really a good article.

Atharva-Phatak commented

also can you tell how did you generate the crops from movies ?? is there a script to generate hug res crops ?

TachibanaYoshino commented

The pictures are manually selected by myself, some face data still need to be adjusted with Photoshop, and finally the denoising algorithm is used to improve the image quality.
Regarding the generated sample video, I used Premiere to do video clips and add background music.

keshav1990 commented

Can you please release code for Linux machine and for Disney code set too, please

TachibanaYoshino commented

For extended research on portrait-style models, data, and code, you can contact Asher Chan via email to inquire about commercial licenses.