The team of researchers is looking to make MRIs up to 10 times faster. Increased scan times may take some of the dread out of MRIs. (Photo via Getty Images)
MRIs are important procedures used to detect brain tumors, brain injuries, infection, and dementia just to name a few conditions. But it’s a long, drawn-out process, not to mention stressful. And if you’re claustrophobic, the thought of going into that narrow tube is terrifying. This is why scientists with the Facebook Artificial Intelligence Research (FAIR) group has teamed up with the New York University School of Medicine to study if AI can speed up the MRI process.
The team is currently working on technology that could make MRIs as much as 10 times faster than it currently takes using AI. The way MRIs work is by gathering data and turning it into cross-sectional images of internal body structure, but the scan time takes longer depending on how large the scanning area is. NYU and FAIR researchers plan to speed up scans by collecting less raw data and letting trained neural networks fill in the gaps. "The key is to train artificial neural networks to recognize the underlying structure of the images in order to fill in views omitted from the accelerated scan," said researchers in a blog post about the work.
But the project isn’t without its challenges. One of the biggest issues researchers face is making sure the MRI achieves high accuracy while reducing scan times. It’s important they avoid mistakes, which could lead to a misdiagnosis. This is why the team has worked on the project for so long – NYU researchers started looking into AI to speed up MRIs in 2016.
The team will use roughly 10,000 clinical cases and 3 million MRI images that have had the patient information removed from it. They also plan to make the work open source meaning the AI models, baselines, evaluation metrics, and image data sets will be available to other researchers.
So, what does Facebook have to do with this? Nothing. The project isn’t related to their bottom-line business or its mission of connecting the world. Rather, they’re interested in extending the capabilities of AI. The company has experimented with AI with different third-party groups in the past, but this marks the first time they’ve worked on something that aims to solve a problem.
Aside from speeding up MRIs, the project could also result in more accessible diagnostic technology. If it proves to be successful, it could eventually replace X-ray and CT machines in certain cases ensuring patients can avoid being exposed to radiation. So far, the technology is nowhere near ready to be implemented in hospitals.
Have a story tip? Message me at: cabe(at)element14(dot)com