search

youngLBW / HRN

[CVPR2023] A Hierarchical Representation Network for Accurate and Detailed Face Reconstruction from In-The-Wild Images.
https://younglbw.github.io/HRN-homepage/
Apache License 2.0
468 stars 42 forks source link

Multi view reconstruction #34

Open yumikim381 opened 1 year ago

yumikim381 commented 1 year ago

Dear author, When i run the demo for multiview reconstruction, i get results for each view - 5 different meshes. Is there a way that i can obtain the one final mesh for each subject?

yumikim381 commented 1 year ago

Or more specifically is the last view's reconstruction the final result?

Nibenitez commented 1 year ago

Dear author, When i run the demo for multiview reconstruction, i get results for each view - 5 different meshes. Is there a way that i can obtain the one final mesh for each subject?

I have the same question, thank you.

youngLBW commented 11 months ago

The model generates coeffs and displacement map for each view. You can average all the results to get the final displacement map and coeffs. Then by combining the final displacement map, coeffs and the shared deformation map, you can obtain a final mesh for each subject.

Nabuyay commented 4 months ago

The model generates coeffs and displacement map for each view. You can average all the results to get the final displacement map and coeffs. Then by combining the final displacement map, coeffs and the shared deformation map, you can obtain a final mesh for each subject.

Hello, I am quite new to doing coding and I am not quite sure on how to averaging the results and combining the displacement map. May I ask how to do so? Because now what I can see is that the results generated from single-view and multi-view is the same. I am using photos of a same person from three angles, one from the front and two other sides. And I still get three seperate mesh when I use single-view mode or multi-view mode. Thank you so much for the help.

nvargass commented 3 months ago

The model generates coeffs and displacement map for each view. You can average all the results to get the final displacement map and coeffs. Then by combining the final displacement map, coeffs and the shared deformation map, you can obtain a final mesh for each subject.

Hello, I am quite new to doing coding and I am not quite sure on how to averaging the results and combining the displacement map. May I ask how to do so? Because now what I can see is that the results generated from single-view and multi-view is the same. I am using photos of a same person from three angles, one from the front and two other sides. And I still get three seperate mesh when I use single-view mode or multi-view mode. Thank you so much for the help.

Did you could solve it ?

Nabuyay commented 3 months ago

The model generates coeffs and displacement map for each view. You can average all the results to get the final displacement map and coeffs. Then by combining the final displacement map, coeffs and the shared deformation map, you can obtain a final mesh for each subject.

Hello, I am quite new to doing coding and I am not quite sure on how to averaging the results and combining the displacement map. May I ask how to do so? Because now what I can see is that the results generated from single-view and multi-view is the same. I am using photos of a same person from three angles, one from the front and two other sides. And I still get three seperate mesh when I use single-view mode or multi-view mode. Thank you so much for the help.

Did you could solve it ?

Unfortunately not yet