Optimization and Evaluation of k-Space Undersampling Strategies for Accelerated Deuterium CSI

Biomedical Engineering

Bret Ashby

Abstract

This project focuses on improving the efficiency of Deuterium Chemical Shift Imaging (CSI) by optimizing k-space undersampling strategies and reconstruction performance. Undersampling mask strategies including pseudorandom, a priori (FLASH, T1W, T2W), and a central weighted variable density mask were evaluated across multiple acceleration factors.

Reconstruction accuracy was quantified using normalized Root Mean Square Error (NRMSE) and Structural Similarity Index (SSIM) to identify mask strategies that balance acquisition speed and image fidelity.

Video

Research poster

Faculty mentor

Portrait of Vikram Kodibagkar

Vikram Kodibagkar

Professor

School of Biological and Health Systems Engineering

[email protected]