NOTE: Please look at the existing list of Issues tagged with the label 'enhancement`. Only open a new issue if you do not see your feature request there.
Hi, I have a huge amount of mesh data (in .off format) and would like to use Pytorch3D to process it because it has most of the functions I want. Are you adding support for .off mesh file?
Do you think I can modify the functionalities of .ply format (in Pytorch3D) to work on .off format? Because .ply structure is closer to .off than .obj
Thank you.
NOTE: we only consider adding new features if they are useful for many users.
Hi @jibrilmuhammadadam
Supporting .off would be great. If you would like to submit a PR that supports load and save functionality for .off files we would be more than happy to review it.
I would like to work on adding .off io if @jibrilmuhammadadam or anyone else isn't already working on it.
I am not working at it @raincrash . You can work on it, it will be of great help. Thanks.
@raincrash that's exciting!!
When adding a new feature to PyTorch3D, please follow our guidelines which can be found here
https://github.com/facebookresearch/pytorch3d/blob/master/.github/CONTRIBUTING.md
Hi, @gkioxari @nikhilaravi, I have a quick question on the design of the io module, do you recommend preserving the same structure for the other files - ply_io.py and obj_io.py and adding a new off_io.py file, or refactor the module to support offs? I am asking because the off support can use some private functions of ply_io and will have some formatting and property (eg ASCII-only) differences.
@raincrash have you already started working on .off support? We are happy to work on this internally if not.
@nikhilaravi I have just started adding load_off calls and the raw version of it and realized it would make more sense to refactor ply_io to support more input/output types other than obj. But I wasn't sure if I should change the ply_io since it has dependencies across. I can try working on this if there is a clearer way on how to handle this. If there is another good-first-issue, I am happy to do that instead as well. :)
I am happy to make this change and I intend to do it soon. If you have done a lot on it and want to make a PR please do so by the end of September.
The ascii OFF format is clear to me, and I can see datasets which use it, but I cannot find examples of binary OFF files. I am wondering if anyone has examples of the binary version?
I think occupancy network used that .off file format. Hope it can help.
This has just landed in https://github.com/facebookresearch/pytorch3d/commit/0345f860d4d44e0a36b3a366644f6432458ae5cc . I am closing this issue.