samuel930930/Dual-Manifold-Adversarial-Training

Projection operation in Fig.1

Closed this issue · 1 comments

zheng547 commented

image

I could not find the position of relating projection operation codes in the project. Could you point it out?

Thanks.

samuel930930 commented

The projection described in Fig. 1 is GAN inversion, which is computationally costly. Therefore, we preprocess all the images and save all the corresponding latent vectors and construct the proposed OM-ImageNet dataset. Hence, there is no code about the projection operation.