TECH

Facebook wants to use Artificial Intelligence to accelerate magnetic resonance imaging
The social network has announced new research to see if artificial intelligence (AI) can be used to reconstruct magnetic resonance imaging (MRI) images in new and faster ways. In this case, up to 10 times faster.To help train the algorithms of the fastMRI project, the University of New York (NYU) will give away three million magnetic resonance imaging, knee, brain, and liver images from 10,000 clinical cases, properly anonymised.A NYU imaging specialist explained that although MRIs are one of the best ways to see what's going on in the patient's body, they require it to stand still in a time that can go up to an hour while the scanner collects the data. If this project succeeds, it can turn a 50-minute evaluation into a five-minute evaluation.In his blog, Facebook clarified that by using AI throughout this process it is possible to collect "less data and therefore scan images faster, preserving or even improving the vast content of MRI information."The social network explains that "the key is to train artificial neural networks to recognize the underlying structure of the images in order to fill in the images that are not revealed in the exams."This approach is similar to the way humans process sensory information. When we experience the world, our brains often receive an incomplete image that we need to transform into information," he concluded.
Given concerns that may arise regarding the privacy of data from the patients involved in the project, the partners ensured that all names were removed as well as "all other health information."


Sapo