search

tengteng95 / Pose-Transfer

Code for the paper Progressive Pose Attention for Person Image Generation in CVPR19 (Oral).
MIT License
734 stars 160 forks source link

doubt about the equation(5) in original paper #45

Closed venuszhou closed 4 years ago

venuszhou commented 4 years ago

@tengteng95 hi, Could you please help me with the quation(5) in your original paper? I don't exactly know what the Pt is and how you can get it. Thanks.

jessemelpolio commented 4 years ago

Pt means the target image. You could simply take it as the ground truth image.

venuszhou commented 4 years ago

@jessemelpolio thanks, but how can you get the target (pose) image? What the author does is to try to utilize the model to generate the image with target pose, but how can you get the ground truth? Please point it out if I made any misunderstanding.

jessemelpolio commented 4 years ago

Eq. 5 is used in the training procedure. It doesn't necessarily mean to have this target image in the testing phase.