Closed dajinstory closed 5 years ago
It is possible. The method works similarly to others: first you extract the pose in term of a keypoints Y. Then you combine this pose Y and appearance A using GAN. In paper images of the same person show for comparison with ground truth. You can youse demo.py
for this.
Thanks for quick answer! :) The same person in both input images is just for training.
I want to use pose-gan in this way.
But, i read the paper and it was quite different from what i thought...... In paper, two input images must have same person. just with different pose.
Do you think i can get the result what i want?
Sorry for asking too personal questions..