This is the official code repository for our paper, "Skin3D: Detection and Longitudinal Tracking of Pigmented Skin Lesions in 3D Total-Body Textured Meshes", published in Medical Imaging Analysis, 2021.
In this work, we propose an automated approach to detect and longitudinally track skin lesions on the skin surface of 3D total-body scans.
As there currently is no other large-scale publicly available dataset of 3D total-body skin lesions, we publicly release over 25,000
3DBodyTex manual annotations, which we hope will further encourage research on total-body skin lesion analysis.
This repo provides the data for 3DBodyTex bounding boxes annotations, and shows how to visualize and use the code in your project.
If you use the data or any part of the code, please consider citing our work, and the original 3DBodyTex dataset.
@article{zhao2022skin3d,
title={Skin3D: Detection and longitudinal tracking of pigmented skin lesions in 3D total-body textured meshes},
author={Zhao, Mengliu and Kawahara, Jeremy and Abhishek, Kumar and Shamanian, Sajjad and Hamarneh, Ghassan},
journal={Medical Image Analysis},
volume={77},
pages={102329},
year={2022},
publisher={Elsevier}
}
The 3D total-body scans can be downloaded here: 3DBodyTex.v1.
You need to sign an agreement in order to access the data.
Once you get the data, you can start using skin3d
.
NOTE: The bounding boxes provided by skin3d
correspond to the high resolution meshes.
To use skin3d
:
- Fill out the form to request and download 3DBodyTex.v1
- Clone this repository:
git clone https://github.com/jeremykawahara/skin3d.git
- Navigate to the repository:
cd skin3d
- Create a new environment using conda:
conda create -n skin3d python=3.7
- Activate the environment:
conda activate skin3d
- Install the requirements:
pip install -r requirements.txt
We provide example notebooks to help explain the format of the data.
- bodytex_annotations_data shows the format of the train, valid, and multiple annotator test data
- bodytex_longitudinal_data shows the format of the IDs used to track the same lesion across scans
- bodytex_annotations_visual illustrates how to visualize multiple annotators on a texture image
This repository is structured as follows:
- data/ contains the 25,000+ manual lesion annotations for 3DBodyTex
- notebooks/ contains example notebooks that load and process the annotations
- skin3d/ contains the Python package to load and process the annotations
If you use this data or code, please cite the following works:
Description of the annotated bounding boxes [pdf] [doi]:
@article{zhao2022skin3d,
title={Skin3D: Detection and longitudinal tracking of pigmented skin lesions in 3D total-body textured meshes},
author={Zhao, Mengliu and Kawahara, Jeremy and Abhishek, Kumar and Shamanian, Sajjad and Hamarneh, Ghassan},
journal={Medical Image Analysis},
volume={77},
pages={102329},
year={2022},
publisher={Elsevier}
}
Description of the meshes [pdf] [doi]:
@inproceedings{saint20183dbodytex,
title={3dbodytex: Textured 3d body dataset},
author={Saint, Alexandre and Ahmed, Eman and Cherenkova, Kseniya and Gusev, Gleb and Aouada, Djamila and Ottersten, Bjorn and others},
booktitle={2018 International Conference on 3D Vision (3DV)},
pages={495--504},
year={2018},
organization={IEEE}
}