MedShapeNetCore

MedShapeNetCore is a subset of MedShapeNet, containing more lightweight 3D anatomical shapes in the format of mask, point cloud and mesh. The shape data are stored as numpy arrays in nested dictonaries in npz format (Zenodo). This API provides means to downloading, accessing and processing the shape data via Python, which integrates MedShapeNetCore seamless into Python-based machine learning workflows. For details, visit the Github Repository.  If you use the dataset in your research, please also cite the original datasets (which can be found using command <!python -m MedShapeNetCore info>), besides MedShapeNet.

Make sure to check out the MedShapeNetCore Show Case for the latest update.

@article{li2023medshapenet,
  title={MedShapeNet--A Large-Scale Dataset of 3D Medical Shapes for Computer Vision},
  author={Li, Jianning and Zhou, Zongwei and Yang, Jiancheng and others},
  journal={arXiv preprint arXiv:2308.16139},
  year={2023}
}

To install the MedShapeNetCore python package:

 pip install MedShapeNetCore

You can search the database using anatomy nomenclature, such as liver, aorta, skull, instrument etc.  Fore more commands, visit the Github Repository.

python -m MedShapeNetCore search_by_organ ORGAN

You can also directly download .stl files using the following command (replace ORGAN with the organ you want to search, e.g., liver, skull):

python -m MedShapeNetCore search_and_download ORGAN

Check the available .npz files:

python -m MedShapeNetCore info

Download .npz files to be used in python:

python -m MedShapeNetCore download FaceVR

Import the python packages to load, visualize and process the .npz files:

from MedShapeNetCore.MedShapeNetCore import MyDict,MSNLoader,MSNVisualizer,MSNSaver,MSNTransformer

Import shape dataloader:

from MedShapeNetCore.MedShapeNetCore import BatchLoader
facial_point=BatchLoader('facialVR','point',batch_size=2,shuffle=True)
for batch in facial_point:
    print(batch.shape)

Use the fast load function to load a specified number of  samples  (2) of a given format ('point') from a dataset ('facialVR'):

msn_loader=MSNLoader()
data=msn_loader.fast_load('facialVR','point',2,shuffle=True)
print(data.shape)