Sign In

Communications of the ACM

ACM News

Facebook AI, NYU Langone Create Accurate MRIs with Less Time and Data

View as: Print Mobile App Share:
attendant prepares patient for MRI scan

Artificial intelligence filled in gaps in scan data and speeded up the MRI process, researchers say.

Researchers at Facebook Inc. and NYU Langone Health say that a two-year experiment demonstrates how artificial intelligence can produce a fourfold increase in the speed of magnetic resonance imaging.

The results of the joint effort, which originated at Facebook, show that accurate MRIs can be produced more quickly as AI functions with less data, researchers say.

"Our goal is a five-minute MRI. This has the potential to be a new paradigm," says Michael Recht, professor and chair of the radiology department at NYU Langone Health and a co-author of the research, published in the American Journal of Roentgenology.

"The challenge was, 'Can we create MRI images with a fraction of the data?' There are certain benchmarks in computer vision and medical imaging that require a higher bar. Our study results show it is possible," says Nafissa Yakubova, a co-author of the research paper and a leader of the Facebook AI Research team.

From The Wall Street Journal
View Full Article – May Require Paid Registration


No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account