Institute of Bioimaging

Canon studies AI-based MR technology with international academic partners

March 22, 2018
by Lauren Dubinsky, Senior Reporter
Canon Medical Systems Corporation is collaborating with Kumamoto University and the University of Bordeaux to study the application of an AI-based technology in MR imaging.

The Deep Learning Reconstruction (DLR) technology could be useful in eliminating noise from newly acquired images. It may also have clinical applications because it acquires ultra-high-resolution images much faster than standard imaging methods.

Compared to a conventional smoothing filter, the DLR noise elimination method causes a small reduction in image quality. However, it minimizes signal variation in organ parenchyma, which improves image quality as well as accuracy in quantitative analysis.

Vincent Dousset, professor at the University of Bordeaux, explained that when DLR is used in conjunction with Canon’s 3T MR system it generates images comparable to those acquired with a 7T MR system. Because of that, DLR might be able to replace some high-field conventional MR studies.

Canon stated that DLR is a “major technological advance that will dramatically change how MR examinations are performed in the future.”

An opening ceremony took place earlier this month at the Institute of Bioimaging at the University of Bordeaux, which is where the collaborative research will be conducted.