Pix2mesh wrong number of gt points in loader?

autonomousvision / occupancy_networks

This repository contains the code for the paper "Occupancy Networks - Learning 3D Reconstruction in Function Space"

https://avg.is.tuebingen.mpg.de/publications/occupancy-networks

MIT License

1.49k stars 291 forks source link

Pix2mesh wrong number of gt points in loader? #49

Closed frankshen07 closed 4 years ago

frankshen07 commented 4 years ago

Thank you for the great work! There is one part that I'm not clear about:

https://github.com/autonomousvision/occupancy_networks/blob/39b772e265e8cbb5a2aadf90541eff80417a2d37/im2mesh/pix2mesh/config.py#L82-L83

Should here be pointcloud_target_n? This might explain why pix2mesh was reported worse than AtlasNet, which contradicts the results I saw in other paper.

m-niemeyer commented 4 years ago

Hi @frankshen07 ,

Thanks a lot for your great observation! pointcloud_n doesn't make sense here as we don't use point cloud input - here want to specify how many points are subsampled for calculating the Chamfer distance, hence pointcloud_target_n is correct!

We checked our internal repository and here we used the correct naming - hence we trained with the correct number of points. Still, thanks a lot for your observation!