Overcoming Data Scarcity in Biomedical Imaging with a Foundational Multi-Task Model - Model Archive
Year: 2024
Type: dataset
Abstract: This archive contains additional material to reproduce the study Overcoming Data Scarcity in Biomedical Imaging with a Foundational Multi-Task Model. It contains archived versions of the code available at https://github.com/FraunhoferMEVIS/MedicalMultitaskModeling. Here, the respective archived duplicates are named "mtl-torch". For practical use, we recommend using the latest version of the training strategy. The software is provided on "AS IS" basis, i.e. it comes without any warranty, express or implied including (without limitations) any warranty of merchantability and warranty of fitness for a particular purpose. Please note that this software is licensed under the LICENSE FOR SCIENTIFIC NON-COMMERCIAL RESEARCH PURPOSES, see `code/original_version/mtl-torch/license.md` and `code/medmnist_version/mtl-torch/license.md`. Folders and files code/original_version -> The version of the code used to run the experiments with UMedPT and its benchmark. code/medmnist_version -> The version of the code used to run the MedMNIST experiments separately from UMedPT. umedpt_checkpoint.zip -> The set of weights required to reproduce the study results with UMedPT. umedpt-large.pt -> The native PyTorch export of the weights required to reproduce the supplementary results with UMedPT-large. See the nested readme files for more information.
Authors: Schäfer, Jan Raphael, Nicke, Till, Höfener, Henning, Lange, Annkristin, Merhof, Dorit, Feuerhake, Friedrich, Schulz, Volkmar, Lotz, Johannes, Kiessling, Fabian
Institution: Fraunhofer Institute for Digital Medicine
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W6911240678