Open CrossLee1 opened 4 years ago
Thanks for your awesome work~
I conducted experiments with my own data with your provided code, and find that when human legs in driving video overlap, the generated ones cannot determine which one should be at front, leading to a confused result.
Do you have some advices on solving this? Thanks~
It is not possible to understand for the network if you are using 2d keypoints, you should consider using 3d keypoints.
Thanks for your awesome work~
I conducted experiments with my own data with your provided code, and find that when human legs in driving video overlap, the generated ones cannot determine which one should be at front, leading to a confused result.
Do you have some advices on solving this? Thanks~