---
language:
- en
license: bsd
size_categories:
- 1K<n<10K
pretty_name: arkit_labelmaker
viewer: false
tags:
- 3D semantic segmentation
- indoor 3D scene dataset
- pointcloud-segmentation
task_categories:
- image-segmentation
---

# ARKit Labelmaker: A New Scale for Indoor 3D Scene Understanding

[[arxiv]](https://arxiv.org/abs/2410.13924)  [[website]](https://labelmaker.org/) [[checkpoints]](https://huggingface.co/labelmaker/PTv3-ARKit-LabelMaker) [[code]](https://github.com/cvg/LabelMaker)

We complement ARKitScenes dataset with dense semantic annotations that are automatically generated at scale. This produces the first large-scale, real-world 3D dataset with dense semantic annotations.
Training on this auto-generated data, we push forward the state-of-the-art performance on ScanNet and ScanNet200 with prevalent 3D semantic segmentation models.