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

Purpose: Long scan time in phase encoding for forming complete K-spacematrices is a critical drawback of MRI, making patients uncomfortable andwasting important time for diagnosing emergent diseases. This paper aims toreducing the scan time by actively and sequentially selecting partial phases ina short time so that a slice can be accurately reconstructed from the resultantslice-specific incomplete K-space matrix. Methods: A transformer based deepreinforcement learning framework is proposed for actively determining asequence of partial phases according to reconstruction-quality based Q-value (afunction of reward), where the reward is the improvement degree ofreconstructed image quality. The Q-value is efficiently predicted from binaryphase-indicator vectors, incomplete K-space matrices and their correspondingundersampled images with a light-weight transformer so that the sequentialinformation of phases and global relationship in images can be used. Theinverse Fourier transform is employed for efficiently computing theundersampled images and hence gaining the rewards of selecting phases. Results:Experimental results on the fastMRI dataset with original K-space dataaccessible demonstrate the efficiency and accuracy superiorities of proposedmethod. Compared with the state-of-the-art reinforcement learning based methodproposed by Pineda et al., the proposed method is roughly 150 times faster andachieves significant improvement in reconstruction accuracy. Conclusions: Wehave proposed a light-weight transformer based deep reinforcement learningframework for generating high-quality slice-specific trajectory consisting of asmall number of phases. The proposed method, called TITLE (Transformer InvolvedTrajectory LEarning), has remarkable superiority in phase-encode selectionefficiency and image reconstruction accuracy.

Active Phase-Encode Selection for Slice-Specific Fast MR Scanning Using a Transformer-Based Deep Reinforcement Learning Framework