The paper titled “OpenMEDLab: An Open-source Platform for Multi-modality Foundation Models in Medicine” presents a set of open source AI models and tools for the medical sector and its related fields.
Some of these interesting projects include:
- PULSE: A medical large language model.
- MIS-FM: 3D medical image segmentation model.
- SAM-Med2D and Sam-Med3D: Additional models for medical image segmentation. They work on 2D and 3D medical images.
- RTEFound: Retinal image recognition model for detecting diseases in human eyes.
They also provide medical datasets and model evaluation tools for training AI models on solving medical-related issues.
Most of these projects are licensed under Apache2.0, AGPL or the CC-BY-NC license. You should double-check the license of each project separately before you decide to build upon it.
Their collective GitHub profile can be found at the following URL: https://github.com/openmedlab
Apparently, the researchers are from China, but every project is well-documented and explained in English on its GitHub page.
These models, tools and datasets are superhelpful for anyone wishing to build an open-source medical application or develop an existing solution further. I highly recommend reading the introductory paper of the project that I mentioned at the beginning of this post to get a general insight.