fastMONAI: A low-code deep learning library for medical image analysis
Peer reviewed, Journal article
Published version
Permanent lenke
https://hdl.handle.net/11250/3099143Utgivelsesdato
2023Metadata
Vis full innførselSamlinger
Sammendrag
We introduce fastMONAI, an open-source Python-based deep learning library for 3D medical imaging. Drawing upon the strengths of fastai, MONAI, and TorchIO, fastMONAI simplifies the use of advanced techniques for tasks like classification, regression, and segmentation. The library’s design addresses domain-specific demands while promoting best practices, facilitating efficient model development. It offers newcomers an easier entry into the field while keeping the option to make advanced, lower-level customizations if needed. This paper describes the library’s design, impact, limitations, and plans for future work.