A novel framework to solve data deficiency problems caused by privacy issues and time-consuming and costly medical data annotation processes.
Main objectives of Deepsynthbody are:
- Overcome the privacy-related limitations of medical data by producing open-access deep synthetic data.
- Reduce the time-consuming and resource-consuming process of medical data labeling and annotation.
- Find inter-correlations of human body organs (how one organ affects other organs) and functions and reproduce them to produce a new model for the human body.
Additionaly, Deepsynthbody works as:
- A repository for deep generative models used in medicine.
- A data compression mechanism to keep big medical data sets in small storage without any privacy concerns and space to save large amounts of the data
Generative models are organized under 11 categories which may contain subcategories and sub-sub categories.
- Cardiovascular
- Digestive
- Endocrine
- Integumentary
- Lymphatic
- Muscular
- Nervous
- Urinary
- Reproductive
- Respiratory
- Skeletal
How to contribute to Deepsynthbody as a researcher?
You can research and implement DeepGenerative models for medical data and publish models, datasets, and applications here. Remember to include references to your original study and papers in the readme files. Make your research more visible to a broader audience.
Citation:
@inproceedings{deepsynthbody,
title={DeepSynthBody: the beginning of the end for data deficiency in medicine},
author={Thambawita, Vajira and Hicks, Steven A. and Isaksen, Jonas, Stensen, Mette Haug and Haugen, Trine B. and Kanters, Jørgen and Parasa, Sravanthi and Lange, Thomas de and Johansen, Håvard D. and Johanse, Dag and Hammer, Hugo L. and Halvorsen, P{\aa}l and Riegler, Michael A.},
booktitle={In Proceedings of the International Conference on Applied Artificial Intelligence (ICAPAI 2021)},
year={2021}
}
Contact us
deepsynthbody@gmail.com
vajira@simula.no
michael@simula.no