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

Dense prediction in medical volume provides enriched guidance for clinicalanalysis. CNN backbones have met bottleneck due to lack of long-rangedependencies and global context modeling power. Recent works proposed tocombine vision transformer with CNN, due to its strong global capture abilityand learning capability. However, most works are limited to simply applyingpure transformer with several fatal flaws (i.e., lack of inductive bias, heavycomputation and little consideration for 3D data). Therefore, designing anelegant and efficient vision transformer learner for dense prediction inmedical volume is promising and challenging. In this paper, we propose a novel3D Shuffle-Mixer network of a new Local Vision Transformer-MLP paradigm formedical dense prediction. In our network, a local vision transformer block isutilized to shuffle and learn spatial context from full-view slices ofrearranged volume, a residual axial-MLP is designed to mix and captureremaining volume context in a slice-aware manner, and a MLP view aggregator isemployed to project the learned full-view rich context to the volume feature ina view-aware manner. Moreover, an Adaptive Scaled Enhanced Shortcut is proposedfor local vision transformer to enhance feature along spatial and channeldimensions adaptively, and a CrossMerge is proposed to skip-connects themulti-scale feature appropriately in the pyramid architecture. Extensiveexperiments demonstrate the proposed model outperforms other state-of-the-artmedical dense prediction methods.

3D Shuffle-Mixer: An Efficient Context-Aware Vision Learner of Transformer-MLP Paradigm for Dense Prediction in Medical Volume