Deep J-Sense: Accelerated MRI Reconstruction via Unrolled Alternating Optimization
This work addresses robustness issues in accelerated MRI reconstruction for medical imaging applications, representing an incremental improvement over prior methods.
The paper tackles the problem of degraded performance in accelerated MRI reconstruction when coil sensitivity estimates are poor or scan parameters differ from training conditions, by introducing Deep J-Sense, which refines both magnetization and coil sensitivity maps, resulting in increased reconstruction performance and robustness to varying acceleration factors and calibration region sizes on the knee fastMRI dataset.
Accelerated multi-coil magnetic resonance imaging reconstruction has seen a substantial recent improvement combining compressed sensing with deep learning. However, most of these methods rely on estimates of the coil sensitivity profiles, or on calibration data for estimating model parameters. Prior work has shown that these methods degrade in performance when the quality of these estimators are poor or when the scan parameters differ from the training conditions. Here we introduce Deep J-Sense as a deep learning approach that builds on unrolled alternating minimization and increases robustness: our algorithm refines both the magnetization (image) kernel and the coil sensitivity maps. Experimental results on a subset of the knee fastMRI dataset show that this increases reconstruction performance and provides a significant degree of robustness to varying acceleration factors and calibration region sizes.