Hangz-nju-cuhk / Talking-Face-Generation-DAVS

Code for Talking Face Generation by Adversarially Disentangled Audio-Visual Representation (AAAI 2019)
MIT License
819 stars 173 forks source link

train code #3

Open xiaoyun4 opened 6 years ago

xiaoyun4 commented 6 years ago

@Hangz-nju-cuhk Hi, This work is amazing and I run the code. But I have two questions:

  1. could you release the train code? and How can we prepare our train data? Could you describe these steps in details?
  2. After test, I find the video generated is jitters. But the demo you release is steady. Video Stabilization is important for this task. I want to know how you process this phenomenon?

Thanks a lot!

Hangz-nju-cuhk commented 6 years ago

Thanks for reaching out. I already add the answers to your questions to README. Our training code is preparing and will be released later. As for stabilization, we use the `Subspace Video Stabilization' (http://web.cecs.pdx.edu/~fliu/papers/tog2010.pdf) for postprocessing. Sorry that we forgot to mention this part before.

JianxiangWang commented 6 years ago

Hi,when will the training code be released? Have estimated time? Thx~

Hangz-nju-cuhk commented 6 years ago

@JianxiangWang I will be working on sorting out the train code and upgrading the project to Pytorch 0.4.1 within this month. Hopefully, they will be released by December ><.

zzzzhuque commented 6 years ago

Thanks for reaching out. I already add the answers to your questions to README. Our training code is preparing and will be released later. As for stabilization, we use the `Subspace Video Stabilization' (http://web.cecs.pdx.edu/~fliu/papers/tog2010.pdf) for postprocessing. Sorry that we forgot to mention this part before.

I'm sorry, but there is no code URL on the website. How can I get the code about 'Subspace Video Stabilization'?

Hangz-nju-cuhk commented 6 years ago

@ZHUTAO142857 We used a commercial non-sharable implementation, but we believe there are a great number of open-source projects on stabilization. This is a mature technique which has been adopted largely for example in youtube. We encourage you to find out yourself.

mathmax12 commented 4 years ago

@Hangz-nju-cuhk I wonder where can I get this "commercial non-sharable implementation"?

Hangz-nju-cuhk commented 4 years ago

@mathmax12 actually you can find it in Adobe Premiere X)