Awesome

Packaged version of Mask3D to be used in LabelMaker

Installation

# Some users experienced issues on Ubuntu with an AMD CPU
# Install libopenblas-dev (issue #115, thanks WindWing)
# sudo apt-get install libopenblas-dev

export TORCH_CUDA_ARCH_LIST="6.0 6.1 6.2 7.0 7.2 7.5 8.0 8.6"

conda env create -f environment.yml

conda activate mask3d_cuda113

pip3 install torch==1.12.1+cu113 torchvision==0.13.1+cu113 --extra-index-url https://download.pytorch.org/whl/cu113
pip3 install torch-scatter -f https://data.pyg.org/whl/torch-1.12.1+cu113.html
pip3 install 'git+https://github.com/facebookresearch/detectron2.git@710e7795d0eeadf9def0e7ef957eea13532e34cf' --no-deps

mkdir third_party
cd third_party

git clone --recursive "https://github.com/NVIDIA/MinkowskiEngine"
cd MinkowskiEngine
git checkout 02fc608bea4c0549b0a7b00ca1bf15dee4a0b228
python setup.py install --force_cuda --blas=openblas

cd ..
git clone https://github.com/ScanNet/ScanNet.git
cd ScanNet/Segmentator
git checkout 3e5726500896748521a6ceb81271b0f5b2c0e7d2
make

cd third_party/pointnet2
python setup.py install

cd ../../
pip3 install pytorch-lightning==1.7.2

pip install .

To use the model in your code you need to download a checkpoint from the list below. Afterwards, the basic model can be used like:

from mask3d import get_model

model = get_model(checkpoint_path='checkpoints/scannet200/scannet200_benchmark.ckpt')

Here is a minimal example assuming you have a pointcloud in the folder data.

from mask3d import get_model, load_mesh, prepare_data, map_output_to_pointcloud, save_colorized_mesh 

model = get_model('checkpoints/scannet200/scannet200_benchmark.ckpt')
model.eval()
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model.to(device)

# load input data
pointcloud_file = 'data/pcl.ply'
mesh = load_mesh(pointcloud_file)

# prepare data
data, points, colors, features, unique_map, inverse_map = prepare_data(mesh, device)

# run model
with torch.no_grad():
    outputs = model(data, raw_coordinates=features)
    
# map output to point cloud
labels = map_output_to_pointcloud(mesh, outputs, inverse_map)

# save colorized mesh
save_colorized_mesh(mesh, labels, 'data/pcl_labelled.ply', colormap='scannet200')

So far, only Scannet200 checkpoints are supported. We are working on the ScanNet checkpoints.

Original Information

Mask3D: Mask Transformer for 3D Instance Segmentation

¹RWTH Aachen University ²ETH AI Center ³ETH Zurich ⁴NVIDIA

Mask3D predicts accurate 3D semantic instances achieving state-of-the-art on ScanNet, ScanNet200, S3DIS and STPLS3D.

<a href="https://pytorch.org/get-started/locally/"><img alt="PyTorch" src="https://img.shields.io/badge/PyTorch-ee4c2c?logo=pytorch&logoColor=white"></a> <a href="https://pytorchlightning.ai/"><img alt="Lightning" src="https://img.shields.io/badge/-Lightning-792ee5?logo=pytorchlightning&logoColor=white"></a> <a href="https://hydra.cc/"><img alt="Config: Hydra" src="https://img.shields.io/badge/Config-Hydra-89b8cd"></a>

[Project Webpage] [Paper] [Demo]

News

• 17. January 2023: Mask3D is accepted at ICRA 2023. :fire:
• 14. October 2022: STPLS3D support added.
• 10. October 2022: Mask3D ranks 2nd on the STPLS3D Challenge hosted by the Urban3D Workshop at ECCV 2022.
• 6. October 2022: Mask3D preprint released on arXiv.
• 25. September 2022: Code released.

Code structure

We adapt the codebase of Mix3D which provides a highly modularized framework for 3D Semantic Segmentation based on the MinkowskiEngine.

├── mix3d
│   ├── main_instance_segmentation.py <- the main file
│   ├── conf                          <- hydra configuration files
│   ├── datasets
│   │   ├── preprocessing             <- folder with preprocessing scripts
│   │   ├── semseg.py                 <- indoor dataset
│   │   └── utils.py        
│   ├── models                        <- Mask3D modules
│   ├── trainer
│   │   ├── __init__.py
│   │   └── trainer.py                <- train loop
│   └── utils
├── data
│   ├── processed                     <- folder for preprocessed datasets
│   └── raw                           <- folder for raw datasets
├── scripts                           <- train scripts
├── docs
├── README.md
└── saved                             <- folder that stores models and logs

Dependencies :memo:

The main dependencies of the project are the following:

python: 3.10.9
cuda: 11.3

You can set up a conda environment as follows

# Some users experienced issues on Ubuntu with an AMD CPU
# Install libopenblas-dev (issue #115, thanks WindWing)
# sudo apt-get install libopenblas-dev

export TORCH_CUDA_ARCH_LIST="6.0 6.1 6.2 7.0 7.2 7.5 8.0 8.6"

conda env create -f environment.yml

conda activate mask3d_cuda113

pip3 install torch==1.12.1+cu113 torchvision==0.13.1+cu113 --extra-index-url https://download.pytorch.org/whl/cu113
pip3 install torch-scatter -f https://data.pyg.org/whl/torch-1.12.1+cu113.html
pip3 install 'git+https://github.com/facebookresearch/detectron2.git@710e7795d0eeadf9def0e7ef957eea13532e34cf' --no-deps

mkdir third_party
cd third_party

git clone --recursive "https://github.com/NVIDIA/MinkowskiEngine"
cd MinkowskiEngine
git checkout 02fc608bea4c0549b0a7b00ca1bf15dee4a0b228
python setup.py install --force_cuda --blas=openblas

cd ..
git clone https://github.com/ScanNet/ScanNet.git
cd ScanNet/Segmentator
git checkout 3e5726500896748521a6ceb81271b0f5b2c0e7d2
make

cd ../../pointnet2
python setup.py install

cd ../../
pip3 install pytorch-lightning==1.7.2

Data preprocessing :hammer:

After installing the dependencies, we preprocess the datasets.

ScanNet / ScanNet200

First, we apply Felzenswalb and Huttenlocher's Graph Based Image Segmentation algorithm to the test scenes using the default parameters. Please refer to the original repository for details. Put the resulting segmentations in ./data/raw/scannet_test_segments.

python -m datasets.preprocessing.scannet_preprocessing preprocess \
--data_dir="PATH_TO_RAW_SCANNET_DATASET" \
--save_dir="data/processed/scannet" \
--git_repo="PATH_TO_SCANNET_GIT_REPO" \
--scannet200=false/true

S3DIS

The S3DIS dataset contains some smalls bugs which we initially fixed manually. We will soon release a preprocessing script which directly preprocesses the original dataset. For the time being, please follow the instructions here to fix the dataset manually. Afterwards, call the preprocessing script as follows:

python -m datasets.preprocessing.s3dis_preprocessing preprocess \
--data_dir="PATH_TO_Stanford3dDataset_v1.2" \
--save_dir="data/processed/s3dis"

STPLS3D

python -m datasets.preprocessing.stpls3d_preprocessing preprocess \
--data_dir="PATH_TO_STPLS3D" \
--save_dir="data/processed/stpls3d"

Training and testing :train2:

Train Mask3D on the ScanNet dataset:

python main_instance_segmentation.py

Please refer to the config scripts (for example here) for detailed instructions how to reproduce our results. In the simplest case the inference command looks as follows:

python main_instance_segmentation.py \
general.checkpoint='PATH_TO_CHECKPOINT.ckpt' \
general.train_mode=false

Trained checkpoints :floppy_disk:

We provide detailed scores and network configurations with trained checkpoints.

S3DIS (pretrained on ScanNet train+val)

Following PointGroup, HAIS and SoftGroup, we finetune a model pretrained on ScanNet (config and checkpoint).

S3DIS (from scratch)

ScanNet v2

ScanNet 200

STPLS3D

BibTeX :pray:

@article{Schult23ICRA,
  title     = {{Mask3D: Mask Transformer for 3D Semantic Instance Segmentation}},
  author    = {Schult, Jonas and Engelmann, Francis and Hermans, Alexander and Litany, Or and Tang, Siyu and Leibe, Bastian},
  booktitle = {{International Conference on Robotics and Automation (ICRA)}},
  year      = {2023}
}