In the latest report from CNN, Facebook is looking to use its artificial intelligence lab in working alongside the medical school of New York University. This move will reportedly make MRI exams at least 10 times faster. If it proves successful, radiologists will be happy knowing that they could complete a single test in minutes.
MRI – magnetic resonance imaging – is used by doctors in taking a closer look at bones, organs, and tissues, among others. This is done without the need to expose patients to harmful radiation. The image quality, in particular, is quite helpful in spotting soft tissue damage.
The only catch, however, is that a single test can take as long as an hour. If a person has a hint of claustrophobia, he can really struggle in remaining perfectly still in a tube-like machine that long. Even typing a machine within that period can drive costs up, considering it limits the number of exams hospitals can perform in a single day.
But as far as computer scientists at Facebook are concerned, there is a workaround. This can be done by using machine learning in order to make things a lot faster than they are now. This idea encouraged NYU to provide an anonymous dataset of about 10,000 MRI exams. This basically involves as many as 3 million images of brains, knees, and livers.
Researchers will be using the data in hopes of creating an algorithm. This will be done through a method known as deep learning. From there, they hope to recognize the very arrangement of bones, ligaments, muscles, and other things that make up the human body.
By building this knowledge into the software that powers an MRI machine, the AI can successfully create a portion of the image and, thus, save time.
According to Daniel Sodickson, the vice chair for research in radiology at NYU School of Medicine, this method will help patients to be in and out in about “five minutes.” “It would be a real game-changer,” he added.
Sodickson even compared the image reconstruction technique to how a person’s brain is able to successfully make out objects in the dark. Basically, when a person looks at a scene in low light, he does not necessarily have a complete view of an object. However, he does know what the underlying structure of that object is; hence, he can quickly and very accurately fill in that missing information without missing a beat.
But, of course, like any other innovation, this idea comes with a huge challenge. Basically, researchers will have to find a way to figure out how to do that without missing even a single important detail. This could be anything from a small bone fracture to a tiny tear in a ligament. Everything must be seen and focused upon.
Nonetheless, researchers are optimistic that they can be successful at this project. As a matter of fact, preliminary findings released last year by radiologists from NYU showed the artificial intelligence to be a tool in reconstructing MRI data.
If tests are made faster and more efficient, radiologists will be able to perform a wider variety of tests. This can be likened to increasing the shutter speed of a camera in order for the turbocharged tests to be used to, say, track the beating of a patient’s heart.
It was last year when the social media giant started talking to NYU about the project. This was thanks to the AI team who all wanted to work on something with real-world benefits, even as it performs basic research. The plan was further explained by Larry Zitnick of the company’s Artificial Intelligence Research group. They are looking to open-source any findings in the hope that sharing the data will encourage others to expand upon its work.