search

CCInc / 3d-ml

A versatile framework for 3D machine learning built on Pytorch Lightning and Hydra [looking for contributors!]
15 stars 3 forks source link

Add Paris-Lille-3D Dataset #28

Open CCInc opened 1 year ago

CCInc commented 1 year ago

The Paris-Lille-3D is a Dataset and a Benchmark on Point Cloud Classification. The data has been produced by a Mobile Laser System (MLS) in two different cities in France (Paris and Lille).

The Point Cloud has been labeled entirely by hand with 50 different classes to help the research community on automatic point cloud segmentation and classification algorithms.

https://npm3d.fr/paris-lille-3d

CCInc commented 1 year ago

See: https://github.com/isl-org/Open3D-ML/tree/master/scripts/download_datasets

aaronfderybel commented 1 year ago

Hi @CCInc , This seems like a rather straightforward dataset to implement. I have not seen examples of implementation using torch_geo or torch dataloader classes of this dataset. But I can get started on this. If you have any suggestions for implementing or direction feel free to comment.

CCInc commented 1 year ago

@aaronfderybel sounds great! I have not used this dataset before so I'm not sure how exactly it's stored but feel free to follow along with the Pytorch Geometric tutorial https://pytorch-geometric.readthedocs.io/en/latest/notes/create_dataset.html#creating-larger-datasets and the examples within this repo. You can also check out this if you haven't seen it yet.

Let me know if you have any questions!