Artificial Intelligence to Improve Resolution of Brain MRI

The method designed by researchers from the University of Malaga enhances the detection of pathologies with greater precision and clarity, without the need for complementary tests

Researchers from the ICAI group -Computer Intelligence and Image Analysis- from the University of Málaga have designed "an unprecedented method capable of improving the images of the brain obtained by magnetic resonance using artificial intelligence", explain university sources.

It is a new model that has allowed the images to go from low to high resolution, without distorting the brain structures of the patients. To do this, it uses a deep artificial neural network - a model inspired by the functioning of the human brain - that 'learns' this process.

"Deep learning is based on very extensive neuronal networks, so its capacity to learn is also very extensive, reaching the complexity and abstraction of a brain", explains the researcher Karl Thurnhofer, main author of this study, who points out that thanks to this technique it is possible to carry out identification tasks by themselves, without supervision, of which not even the human eye would be capable.

In this sense, Thurnhofer emphasizes that until now "the acquisition of quality brain images depended on the time the patient was immobilized in the scanner, with our method the image processing is done later in the computer," he says.

According to the experts, the results will allow specialists to identify more clearly and precisely brain-related pathologies such as physical injuries, cancers or language disorders, among others, since the details of the images are finer, thus avoiding having to resort to complementary tests in the event of doubtful diagnoses.