Meeting Banner
Abstract #2652

Elora: Enforcing Low Rank for Parallel MR Reconstruction

Jun Liu1, Axel Loewe1, Michael O. Zenge2, Alban Lefebvre1, Edgar Mueller2, Mariappan S. Nadar1

1Imaging and Computer Vision, Siemens Corporation, Corporate Technology, Princeton, NJ, United States; 2MR Application & Workflow Development, Siemens AG, Healthcare Sector, Erlangen, Bavaria, Germany

Parallel imaging exploits the difference in sensitivities between individual coil elements in a receive array to reduce the number of gradient encodings required for imaging. SENSE and GRAPPA are two representative approaches. In this abstract, a new approach, called Elora is proposed. Elora implicitly uses coil sensitivities by estimating a low rank subspace from the calibration data, and then works by enforcing the low rank constraint on the sliding blocks of the k-space data.

Keywords

abstract achieved acquisition among application applying approaches approximates approximation array arrow available avoid best block blocks blue bottom calibration called cardiac caution channels clinical coil coils commercially comparable computation computed computer computing concepts constraint contains converted corporate corporation correlations decomposition denote denotes determining development diagonal directly disclaimer domain easily encodings enforcing entries errors estimating every example explicitly extracting extracts favor field firstly form formulated formulation full generated good gradient healthy illustrated illustrates implicitly improved include individual initialized inverse involves iteration iterative kernel knowledge larger lead makes matrix middle might minimal missing much multiplication needed needs normalized note onto operator optimization paper parallel patches plot positive preliminary problem procedure profiles projected proposed rank rather reasonably receive reconstruction reduce representative required respectively rest restricted sampled sampling scanner sector segmented sense sensitivities setting singular sliding smaller solving space spanned specifically specified square step structures subspace technology usually utilizes vector vectors vision visually volunteer white