Computer engineers and radiologists from the Malone Center for Engineering in Healthcare have teamed up with Microsoft to enhance the AI capabilities offered by Microsoft’s Azure cloud computing system.

Microsoft Azure helps data scientists and developers to build and train AI models for various applications. The platform could be particularly useful in radiology, where AI algorithms can assist human radiologists to diagnose and treat diseases.

For this project, the Hopkins team will collaborate with Microsoft’s Health AI group to implement a sequential learning algorithm for training de-identified medical imaging data uploaded to Microsoft Azure from the Picture Archiving and Communication System (PACS). The algorithm will classify small batches of MRI or CT images based on disease type.

The project will be led by Dr. Craig Jones, Assistant Research Professor of Computer Science and Co-Director of the Johns Hopkins Radiology AI Lab, and Dr. Haris Sair, Director of the Johns Hopkins Division of Neuroradiology.

The goal of their collaboration is to maximize the capabilities of Microsoft Azure to accept and process the radiology data. The researchers say their work will support the creation of prediction methods to determine the amount of data needed to attain a particular accuracy level.