Facebook’s artificial intelligence lab is partnering with New York University’s medical school to try and try to improve magnetic resonance imaging systems, according to a CNN report.
NYU’s medical school will provide the social media giant an anonymous dataset of 10,000 MRI exams to train a deep-learning algorithm to recognize anatomical arrangements in hopes that the AI will be able to be integrated into future MRI machines. The algorithm could then, hypothetically, create a portion of the image to save time, as MRI scans can often take as long as an hour to perform, according to the report.
“You could be in and out in five minutes. It would be a real game-changer,” NYU School of Medicine research in radiology vice chair Daniel Sodickson said, according to CNN.
Researchers are optimistic, and preliminary findings released last year show that AI could be used to reconstruct MRI data, according to the report.
The two parties began discussing the partnership last year as the AI team at Facebook wanted to “work on something with real-world benefits,” according to CNN. The groups plan to open-source any findings in hope that the work will be eventually expanded upon.