Meeting Banner
Abstract #2376

An Efficient Scheme of Trajectory Optimization for Both Parallel Imaging and Compressed Sensing

Enhao Gong1, Feng Huang2, Kui Ying3, Xuening Liu4, George Randy Duensing5

1Electrical Engineering, Stanford University, Stanford, CA, United States; 2Philips Healthcare, Shanghai, China; 3Department of Engineering Physics, Tsinghua University, Beijing, China; 4Department of Automation, Tsinghua University, Beijing, China; 5Philips Healthcare, Gainesville, FL, United States

Undersampling of k-space is a widely adopted approach for fast imaging. Instead of using a fixed sampling trajectory, trajectory optimization has been proposed for both Parallel Imaging and Compressed Sensing to achieve significantly improved reconstruction. Here we present an efficient scheme for clinically applicable trajectory optimization by using one scan in the exam as references and fast pseudo-reconstruction. Experiments on in-vivo datasets illustrated the proposed scheme can results in great improvement of reconstruction using Parallel Imaging and Compressed Sensing.

Abstract PDF Poster PDF

Keywords

accelerated acceleration accurately achieve acquisition adaptive adjacent adoption affected among annealing applicability applicable application applied approaches artifact artifacts audience automation balance best better calibrated china chose chosen clinical clinically close coil compressed computation computational computed contrasts correlated correlation costs criterion datasets density domain efficient electrical engineering entire error errors exam expand expensive extrapolation fact fast features field fixed flowchart fully function generated global gong good head hence heuristic highly illustrated implemented implementing improved improvement in vivo initial initialization instead intensity iteration iterations iterative laptop like limit limited locations long magnitude make matrix minimize minimizing moreover noise nonlinear norm objective operator optimal optimization optimize optimized optimizing parallel peaks performance prevent proposed pseudo quantified random randy reconstructed reconstruction recover recovery reduces reduction regularization repeatedly retrospectively sampled sampling scanned scheme search sense sensing series several shanghai significantly simulated since solutions space spatially spirit statistical steps stimulated strategy suppressing system takes target trajectory true variable view widely