Especially at lower field the introduction of Deep Resolve, our AI powered image reconstruction technology, is changing the game in MRI. Deep Resolve employs a deep neural network in an iterative process to reduce noise that is introduced by an accelerated acquisition.
The network is trained with tens of thousands of data pairs representing accelerated and non-accelerated acquisitions. Based on this training, the network within Deep Resolve is able to precisely detect noise that originates from the accelerated acquisition and is able to remove it. The originally acquired raw data is incorporated throughout the entire reconstruction process, until the output of the final reconstructed image. This ensures robust results and that all the potentially clinically relevant information is preserved. Deep Resolve enables reduced acquisition time while preserving the integrity of the underlying signal and the clinical information.
Image Gallery MAGNETOM Free.Max – the Power of Deep Resolve
Heike Weh and Thomas Illigen
(Siemens Healthineers, Erlangen, Germany)