Generating image sequences

NVlabs / stylegan2-ada

StyleGAN2 with adaptive discriminator augmentation (ADA) - Official TensorFlow implementation

https://arxiv.org/abs/2006.06676

Other

1.8k stars 501 forks source link

Generating image sequences #49

Open ffabi opened 3 years ago

ffabi commented 3 years ago

Hi, thanks for the contribution!

Using stylegan2-ada, is it possible to generate consecutive frames of a video? (such as sequences in the KITTI/Cityscapes dataset) I am not talking about style transfer, I would like to generate completely new sequences. Or do you happen to know a better GAN-based solution for this problem?

Thanks, Fabian

dvschultz commented 3 years ago

you wouldn’t use StyleGAN for this. Check out Imaginaire—there are a few models there that will do next frame predicition.