Hi, I'd like to work on this issue. Do tell me if there's anything I need to know apart from what you have mentioned.

Awesome @mehulbhat! The above pointers should be enough, but feel free to ask any questions you have along the way :)

@osanseviero, For adding the ScanNet dataset, do I need to place the whole ScanNet Repository in the required format in the datasets folder?

@mehulbhat, the objective is to add utilities to process the data with TensorFlow datasets as already done for ModelNet. The ScanNet repository has many other tools (an app, server, etc) that we don't care about.

The linked repository has instruction on getting the data, which should be the first step. Once that's done, you can start seeing how to process it.