Faster scanning in the tube

Being pushed into an MRI tube for an examination and lying there motionless for a while while making unpleasant noises: This is a major problem for many patients, some of whom can only tolerate the confinement with sedatives. Siemens Healthineers is helping with a technology that uses artificial intelligence to shorten scan times and improve imaging.


A woman is about to enter the MRI-tube.
Artificial intelligence reduces magnetic resonance imaging (MRI) scan time by up to 70%. © Shutterstock/Garnet

In magnetic resonance imaging (MRI), image quality is defined by the relationship between scan duration, resolution and image noise. Improving one of these components usually means compromising one of the others.

Deep Resolve, a deep learning technology for reconstructing clinical images developed by Siemens Healthineers and unveiled in 2020, aims to solve this dilemma. It offers radiology staff two options: a shortened scan time with the same resolution and reduced noise. Or: increased image quality while maintaining the same scan time.

AI algorithms use raw data

The technology works with the raw data from the scanner. This allows the AI algorithms used to be applied at the earliest stages of image reconstruction. According to the company, this early use of all available raw material has the potential to reduce the duration of scans by up to 70 percent. The so-called Simultaneous Multi Slice technology - i.e. the simultaneous readout of multiple image slices - can even speed up scan times by up to 80 percent.

One example: When examining a knee, imaging with MRI usually takes about ten minutes. "With deep-resolve algorithms, we have reduced this time to less than two minutes - with the same image quality and diagnostic value," says Arthur Kaindl, head of magnetic resonance imaging at Siemens Healthineers.

The use of Deep Resolve is not limited to specific body regions and can help with almost any diagnostic procedure using MRI. It consists of different algorithms that can be combined. Throughout the reconstruction process, the system performs an automatic mandatory data consistency check to ensure the diagnostic value and quality of the image.

Application facts

Technology field
Data Management and Analysis
Image Recognition and Understanding
Application industry
Health and Pharmaceuticals
Field of application
  • Intelligent Assistance Systems
Value-added activity
Research and Development [R&D]
Bundesministerium für Bildung und Forschung (BMBF)
AI Developers

Siemens Healthineers

Large Enterprise