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

Parallel imaging is a commonly used technique to accelerate magneticresonance imaging (MRI) data acquisition. Mathematically, parallel MRIreconstruction can be formulated as an inverse problem relating the sparselysampled k-space measurements to the desired MRI image. Despite the success ofmany existing reconstruction algorithms, it remains a challenge to reliablyreconstruct a high-quality image from highly reduced k-space measurements.Recently, implicit neural representation has emerged as a powerful paradigm toexploit the internal information and the physics of partially acquired data togenerate the desired object. In this study, we introduced IMJENSE, ascan-specific implicit neural representation-based method for improvingparallel MRI reconstruction. Specifically, the underlying MRI image and coilsensitivities were modeled as continuous functions of spatial coordinates,parameterized by neural networks and polynomials, respectively. The weights inthe networks and coefficients in the polynomials were simultaneously learneddirectly from sparsely acquired k-space measurements, without fully sampledground truth data for training. Benefiting from the powerful continuousrepresentation and joint estimation of the MRI image and coil sensitivities,IMJENSE outperforms conventional image or k-space domain reconstructionalgorithms. With extremely limited calibration data, IMJENSE is more stablethan supervised calibrationless and calibration-based deep-learning methods.Results show that IMJENSE robustly reconstructs the images acquired at5$\mathbf{\times}$ and 6$\mathbf{\times}$ accelerations with only 4 or 8calibration lines in 2D Cartesian acquisitions, corresponding to 22.0% and19.5% undersampling rates. The high-quality results and scanning specificitymake the proposed method hold the potential for further accelerating the dataacquisition of parallel MRI.

IMJENSE: Scan-specific Implicit Representation for Joint Coil Sensitivity and Image Estimation in Parallel MRI