Missing intensity from semantic-kitti dataset.

[john-philipp]

Checklist

• I have searched for similar issues.
• I have tested with the latest development wheel.
• I have checked the release documentation and the latest documentation (for main branch).

Describe the issue

I noticed that the semantic kitti dataset in ml3d/datasets/semantickitti.py drops any non coordinate features (which would include the intensity).

Which seems odd. I've included it in a PR. And tested it against some labels with associated intensities on a flat plan.

With feat: None. It's forced to hallucinate. With feat set to remaining point data. It accurately describes the areas after a few cycles.

Notably. The testing for this was done with three datasets for training that hold these labels in geometrically random locations.

Then tested against a forth - entirely unseen - dataset.

No intensity entering model:

Adding feature to remaining point coords:

Steps to reproduce the bug

Just train on an intensity based sample in which no correlation exists between label and location.

Something like a flat plan with areas of differing intensity.

Error message

No response

Expected behavior

Features should be allowed into the model.

Open3D, Python and System information

Based on this commit: `c4cbf85 2024-10-08 08:40 -0700 Sameer Sheorey            o Python 3.12 support (#657)`

Additional information

No response