Only the people who have gone through the MRI machines know the amount of patience that is required to wait while the complete process is over. These MRI scans are usually very long and consume a lot of time in order to be efficient. These machines can take more than an hour for the complete scan which can be difficult for some people that have difficulty in laying for that much time as some people can have few specific conditions such as Claustrophobia in the children. But recently, Facebook has joined hands with the researchers working at the NYU School of Medicine and Department of Radiology in an attempt to make these MRI scans more and more accessible.
The scientists, who are with the Facebook Artificial Intelligence Research Group, which is also called as FAIR, noted that these lengthy scans are indeed problematic and thus reached to a conclusion of using AI to ease the process. The researchers have partnered with NYU by keeping a prime goal in mind and that is using the AI to make the scanning process nearly 10 times faster, which is practically achievable.
The MRI basically uses the data gathered by it to turn the same into some cross-sectional images which are of the internal body structures of the human body that include the blood vessels and the organs. The math is very simple: the larger the area to be examined, the longer the time required by the MRI. This is exactly where the AI comes into picture.
The researchers want to achieve this by scanning as less raw data as possible while the neural networks would be trained to fill in the necessary gaps. The views that are omitted from the scan would be filled by the neural networks after knowing the complete underlying structure which is expected to make the scans much faster without hampering the accuracy.