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

As aliasing artefacts are highly structural and non-local, many MRIreconstruction networks use pooling to enlarge filter coverage and incorporateglobal context. However, this inadvertently impedes fine detail recovery asdownsampling creates a resolution bottleneck. Moreover, real and imaginaryfeatures are commonly split into separate channels, discarding phaseinformation particularly important to high frequency textures. In this work, weintroduce an efficient multi-scale reconstruction network using dilatedconvolutions to preserve resolution and experiment with a complex-valuedversion using complex convolutions. Inspired by parallel dilated filters,multiple receptive fields are processed simultaneously with branches that seeboth large structural artefacts and fine local features. We also adopt denseresidual connections for feature aggregation to efficiently increase scale andthe deep cascade global architecture to reduce overfitting. The real-valuedversion of this model outperformed common reconstruction architectures as wellas a state-of-the-art multi-scale network whilst being three times moreefficient. The complex-valued network yielded better qualitative results whenmore phase information was present.

Multi-scale MRI reconstruction via dilated ensemble networks