Mahmoud Mostapha1, Boris Mailhe1, Simon Arberet1, Dominik Nickel2, and Mariappan S. Nadar 1
1Digital Technology and Innovation, Siemens Healthineers, Princeton, NJ, United States, 2Magnetic Resonance, Siemens Healthineers, Erlangen, Germany
1Digital Technology and Innovation, Siemens Healthineers, Princeton, NJ, United States, 2Magnetic Resonance, Siemens Healthineers, Erlangen, Germany
Predictions from the proposed end-to-end system achieved a PSNR of 33.93 dB and SSIM of 0.840 like those obtained using precomputed CSMs. However, we observed more artifacts with precomputed CSMs. DS-CSME system required less time (~0.2s) to estimate CSMs than ESPIRiT (~1s).
DS-CSME:
a deep learning solution for coils sensitivity estimation, allowing end-to-end
learning framework for accelerated parallel magnetic resonance imaging
reconstruction.
An example comparing
the fully sampled target to predictions obtained using precomputed CSMs and those
obtained by the end-to-end system with DC-CSME. At ~5× acceleration,
predictions with precomputed CSMs show more artifacts.