Focus Session 5: MR Techniques and Methods: Rapid Image Acquisition and Reconstruction (2019)-Rapid Imaging in CMR: Current and Emerging Techniques

Video Transcription

Video Summary

The transcript reviews methods for accelerating cardiac MRI. It first explains <strong>parallel imaging</strong>, the most widely used clinical approach, which undersamples data below Nyquist and uses information from multiple receiver coils to reconstruct images. It is fast and practical, but limited by <strong>noise amplification</strong> and fold-over artifacts, usually allowing modest acceleration factors. Next, it covers <strong>compressed sensing</strong>, which exploits the fact that MRI images are sparse or compressible in transform domains. By using incoherent undersampling and iterative reconstruction with data consistency plus regularization, it reduces artifacts and enables higher-resolution imaging in both dynamic and volumetric applications. The talk then introduces <strong>simultaneous multi-slice imaging</strong>, which excites multiple slices at once to increase coverage and shorten scan time. Finally, it discusses <strong>deep learning</strong>, including unrolled networks and k-space interpolation, as a promising but still evolving approach.

Keywords

cardiac MRI

parallel imaging

compressed sensing

simultaneous multi-slice imaging

deep learning