Meeting Banner
Abstract #1804

Accelerated Real-Time MR Thermometry Using a New Compressed Sensing Framework of Nonlinear Filter and K-T FOCUSS

Feiyu Chen1, Xiaoying Cai1, Xinwei Shi1, Shuo Chen2, Enhao Gong3, Kui Ying2, Shi Wang2

1Department of Biomedical Engineering, Tsinghua University, Beijing, China; 2Department of Engineering Physics, Tsinghua University, Beijing, China; 3Electrical Engineering, Stanford University, Stanford, CA, United States

Phase information is significant in temperature mapping using proton resonance frequency shift (PRFS) method. Acceleration methods can be applied to reconstruction in order to shorten the imaging duration and accomplish real-time temperature mapping. A method to improve the accuracy of phase reconstruction in dynamic scans is proposed in our research. Compressed Sensing with nonlinear filters such as median filter and k-t FOCUSS are combined in our method. Phantom experiments have demonstrated that the proposed method is a promising tool for real-time temperature monitoring using PRFS method.

Abstract PDF Poster PDF

Keywords

accelerated acceleration accomplish according accuracy accurate accurately acquisition additionally advantages among applied artificially audience batch best biomedical black blocks blue boundary brightened calculating central china clinical combination combined combining composition compressed conducted correction correlations curve curves density designed detect displays divided domain duration dynamic effective electrical enables encoding engineering enhance error especially estimation experiment exploit exploiting fact fidelity filter filters final flowchart focal frames framework frequency fully generally gong gradient green highly implementation implemented improve in vivo inaccurate indicate initial initialization latest listed location make mapping maps median might minimize mixed monitoring motion needed nonlinear outer padding periodic phantom physics pixel potential practice precise prediction preliminary process processes promising proposed proton real reconstruct reconstructed reconstruction reduction represents require researchers residue resolution ring robust sampled scanner sensing shorten side since solver space sparse sparsity specifically step steps subject superior system table target temperature temporal term theory thermometry tool trajectory trans transform variable water whole