English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Magnetic Resonance Imaging can produce detailed images of the anatomy andphysiology of the human body that can assist doctors in diagnosing and treatingpathologies such as tumours. However, MRI suffers from very long acquisitiontimes that make it susceptible to patient motion artifacts and limit itspotential to deliver dynamic treatments. Conventional approaches such asParallel Imaging and Compressed Sensing allow for an increase in MRIacquisition speed by reconstructing MR images from sub-sampled MRI dataacquired using multiple receiver coils. Recent advancements in Deep Learningcombined with Parallel Imaging and Compressed Sensing techniques have thepotential to produce high-fidelity reconstructions from highly accelerated MRIdata. In this work we present a novel Deep Learning-based Inverse Problemsolver applied to the task of Accelerated MRI Reconstruction, called theRecurrent Variational Network (RecurrentVarNet), by exploiting the propertiesof Convolutional Recurrent Neural Networks and unrolled algorithms for solvingInverse Problems. The RecurrentVarNet consists of multiple recurrent blocks,each responsible for one iteration of the unrolled variational optimizationscheme for solving the inverse problem of multi-coil Accelerated MRIReconstruction. Contrary to traditional approaches, the optimization steps areperformed in the observation domain ($k$-space) instead of the image domain.Each block of the RecurrentVarNet refines the observed $k$-space and comprisesa data consistency term and a recurrent unit which takes as input a learnedhidden state and the prediction of the previous block. Our proposed methodachieves new state of the art qualitative and quantitative reconstructionresults on 5-fold and 10-fold accelerated data from a public multi-coil braindataset, outperforming previous conventional and deep learning-basedapproaches.

Recurrent Variational Network: A Deep Learning Inverse Problem Solver applied to the task of Accelerated MRI Reconstruction