NVIDIA recently released Clara Train 4.0, an application framework for medical imaging that includes pre-trained models, AI-Assisted Annotation, AutoML, and Federated Learning. In this 4.0 release, there are three new features to help get you started training quicker.
NVIDIA recently released Clara Train 4.0, an application framework for medical imaging that includes pre-trained models, AI-Assisted Annotation, AutoML, and Federated Learning. In this 4.0 release, there are three new features to help get you started training quicker.
Clara Train has upgraded its underlying infrastructure from TensorFlow to MONAI. MONAI is an open-source, PyTorch-based framework that provides domain-optimized foundational capabilities for healthcare. By leveraging MONAI, users now have access to a comprehensive list of medical image–specific transformations and reference networks. Clara Train has also updated its DeepGrow model to work on 3D CT images. This updated model gives you the ability to segment an organ in 3D with only a few clicks across the organ.
Expanding into Digital Pathology, Clara Train helps users navigate these new workloads by providing a Digital Pathology pipeline that includes data loading and training optimizations. These data loading optimizations involve using the new cuCIM library included in RAPIDS.
Clara Train 4.0 continues to improve on its Federated Learning framework by adding homomorphic encryption tools. Homomorphic encryption allows you to compute data while the data is still encrypted. It can play an important role in healthcare in ensuring that patient data stays secure at each hospital while still benefiting from using federated learning with other institutions.
To learn more about these new features, check out our Clara Train 4.0 features highlight video below or the latest blog posts which includes a walkthrough of how you can Bring-Your-Own-Components to Clara Train.
Download Clara Train 4.0 from NGC, and try out the newly updated Jupyter notebooks on GitHub.