AI is ready to unlock massive potential in hospital systems, especially in one of the areas where deep learning holds the most promise: medical imaging.
That’s why NVIDIA introduced Clara AI, a toolkit that includes 13 state-of-the-art classification and segmentation AIs, and software tools built for radiologists.
Leading medical institutions around the world are already using the Clara platform to put the power of AI into the hands of radiologists and take advantage of the growing ecosystem of researchers and startups.
How NVIDIA Clara Lowers Barriers of AI
Labeled data is critical to build safe and robust AI, but radiologists’ time is too precious to spend hours labeling datasets. The Clara AI assisted annotation capability speeds up the creation of structured datasets, enabling annotations in minutes instead of hours.
In fact, MITK (Medical Imaging Interaction Toolkit) developers from the German Cancer Research Center (DKFZ) already integrated Clara AI into its open source viewer used by thousands worldwide.
Transfer learning, another capability in the Clara AI toolkit, adapts existing models to fit local variables. It customizes deep learning algorithms to data that includes local demographics and imaging devices, without having to move or share patient data. As a result, doctors can create models for their own patients with 10x less data than starting from scratch.
It takes a significant amount of technical expertise to integrate AI models and applications into hospital IT systems. The toolkit facilitates the integration of AI models into existing radiology workflows using industry standards, like DICOM.
Major Medical Institutions Use NVIDIA Clara AI
Ohio State University
Using Clara AI, The Ohio State University radiologists, quickly incorporated a model developed at another institution, validated it, and annotated a local dataset to adapt the model to OSU patients. This enables faster AI development of effective algorithms which support clinical care.
National Institutes of Health
The largest research hospital in America, the National Institutes of Health Clinical Center and NVIDIA scientists used Clara AI to develop a domain generalization method for the segmentation of the prostate from surrounding tissue on MRI. The localized model achieved performance similar to that of a radiologist and outperformed other state-of-the-art algorithms that were trained and evaluated on data from the same domain.
University of California, San Francisco
Home to the top-ranked radiology residency program in the country, UCSF is using a Clara AI-powered scalable infrastructure that will enable the seamless creation, testing, and deployment of multiple AI algorithms across radiology, serving as a pathway for future doctors to adopt the system.
“We have an incredibly innovative group of researchers who are building clinically valuable AI tools, and need a consistent way to validate and deploy these tools into clinical workflows,” said Christopher Hess, chair of radiology, UCSF. “NVIDIA Clara will be an essential component of the medical imaging AI ecosystem that enables us to develop and deploy our own and external AI models.”
Hospitals, research institutions, and the whole medical imaging industry can get started now with Clara AI. The toolkit includes two software development toolkits: the Clara Train SDK and Clara Deploy SDK, which can be accessed from NGC, and deployed on hospital-ready infrastructure: NVIDIA T4 server and NVIDIA DGX POD.