SDF-StyleGAN icon indicating copy to clipboard operation
SDF-StyleGAN copied to clipboard

How to generate mesh from image

Open VladStojDev opened this issue 2 years ago • 2 comments

Hi,

Thank you for making this awesome code publicly available! I would just like to know how to perform the shape generation from single-view images, as illustrated in your paper?

Specifically what is the example code and with what parameters should the generator script be called when you want to provide an input image to generate e.g., a car model from?

VladStojDev avatar Jan 03 '23 14:01 VladStojDev

As described in the paper, we use a ResNet to map the input image into the hidden space of the generator on the airplane category. This part of the code is not publicly available for some reasons. But as I described, the core of this application is training an encoder, which is not the contribution of our work. If you are particularly interested, I believe you can complete this part of the code with a simple effort.

Zhengxinyang avatar Jan 03 '23 15:01 Zhengxinyang

Thank you for your diligent reply, I will look into this further. Best of luck with your future research :)

VladStojDev avatar Jan 03 '23 16:01 VladStojDev