Improving Compressed Sensing Initialization and Convergence Using an Efficient Auto-Calibrating Parallel Imaging Method

Keywords

accelerated acceleration accuracy accurate address among applications approaches auto brain calibrating calibration central channel channels clinical coil coils collected combine combined complexity compressed computation compute computer condition considerable convergence converges converted dataset datasets datum decouple decoupling density determined developed direct directly domain efficient eigenvector electrical engineering error evaluate evaluated exploit faster feasibility final finally find fitting full furthermore generate generation healthy highly hybrid improved improving increasing initial initialization intensive inversion investigated iteration iterations iterative knee linear materials matrix modified much multiplications necessitate needed neighbors next optimal original parallel park pattern patterns potential preliminary process processed processing promising proportion proposed proton pseudo radiology random reapplied reconstruct reconstruction reconstructions reduce reduced reduces require requires requiring resolution sampling scanners science sensing sensitivity significantly simulate slightly solution solutions solver source sources space sparsity speed spirit square step steps still subsequent successfully syntheses synthesis synthetic target thousand thread transform transformed typically undesirable uniform upper volunteers zero