The most widely-used open-source informatics platform for imaging research, announced the beta release of XNAT Machine Learning (XNAT ML). Building on the initial announcement at RSNA 2019, XNAT ML will accelerate the creation of AI models by providing an end-to-end development platform enabling faster collaborations between data science and clinical teams.
Last week XNAT, in collaboration with NVIDIA, Radiologics, and the ICR Imaging Informatics group, announced the beta availability of XNAT Machine Learning, adding support for Machine Learning and model training workflows. This integration was first announced at the RSNA 2019 conference, where a proof of concept was demonstrated using models and APIs from the NVIDIA Clara Imaging framework with accelerated GPU computing.
The XNAT ML beta release introduces new capabilities with the XNAT imaging research platform, accelerated by GPUs:
- Assemble training-specific collections of imaging data files into specific training projects to build balanced data cohorts
- Draw new segmentations and annotations on that data, using NVIDIA Clara Train’s AI-assisted annotation
- Install and configure pre-trained models from Clara Train, available through NVIDIA NGC into the XNAT ML environment
- Train models with annotated datasets using training framework provided by NVIDIA Clara Train
Future releases in the XNAT ML line will enable model sharing and inference. This beta version includes documented AI workflows with known limitations.
NVIDIA Clara Imaging is an application framework that accelerates the development and deployment of AI in medical imaging. Built for data scientists and researchers, Clara Imaging offers easy-to-use, domain-optimized tools to create high-quality, labeled datasets, collaborative techniques to train robust AI models, and end-to-end software for scalable and modular AI deployments.