MRI Image Reconstruction Using Distributed Compressed Sensing

The patent segments the MR signal acquisition process by using multiple receiver coils (e.g., 8 coils) to simultaneously capture signals from different spatial regions. Each coil receives signals from a specific subset of the imaged area, enabling parallel acquisition that accelerates the scanning process while the subsequent processing steps mitigate the harmful artifacts through distributed signal processing across multiple channels.

The patent applies compressed sensing techniques that change the sampling parameters by acquiring signals at rates lower than the traditional Nyquist rate. By using iterative reconstruction algorithms with sparsity constraints and incorporating coil sensitivity profiles, the system achieves accelerated acquisition (improved productivity) while reducing aliasing artifacts through intelligent parameter optimization rather than simple rate increases.

The patent implements iterative reconstruction algorithms that incorporate feedback loops. The algorithm repeatedly refines the image reconstruction by comparing the reconstructed signal with the actual measurements, adjusting the solution progressively. This feedback mechanism enables the system to achieve both accelerated acquisition (by using fewer samples) and maintained reconstruction accuracy (through iterative refinement), resolving the trade-off between productivity and precision.

The patent creates a multi-functional reconstruction system that simultaneously handles multiple objectives: it performs compressed sensing for acceleration, incorporates parallel imaging for further speedup, applies sparsity constraints for accuracy, and uses coil sensitivity profiles for artifact reduction. This universal approach integrates multiple techniques into a single framework that achieves both high productivity and high precision that cannot be obtained by any single method alone.