Aliasing Artefact Suppression in Compressed Sensing MRI for Random Phase-Encode Undersampling.

Area of Science:

• Magnetic Resonance Imaging (MRI)
• Image Reconstruction
• Compressed Sensing (CS)

Background:

• Compressed sensing (CS) MRI commonly uses random phase-encode undersampling.
• This 1-D randomness causes coherent aliasing artifacts, degrading image quality.
• A novel reconstruction scheme is proposed to mitigate these artifacts.

Purpose of the Study:

• To develop a new reconstruction scheme for compressed sensing MRI.
• To reduce 1-D undersampling-induced aliasing artifacts in Cartesian k-space trajectories.
• To improve the overall image quality of CS-reconstructed MRI scans.

Main Methods:

• The proposed method involves a two-step reconstruction process.
• Step one: Transforms 2-D image reconstruction into parallel 1-D signal reconstruction utilizing phase direction incoherence.
• Step two: Performs a 2-D CS reconstruction using new k-space data, exploiting correlations in 1-D reconstructed signals.

Main Results:

• The method was evaluated on cardiac cine, brain, foot, and angiogram images.
• Faithful MR image reconstruction was demonstrated at reduction factors up to 10.
• The proposed method outperformed the conventional CS method in reconstruction fidelity.

Conclusions:

• The novel method achieves more accurate reconstruction results than conventional approaches.
• It demonstrates a 2-5 dB gain in peak Signal-to-Noise Ratio (SNR) and higher structural similarity.
• The technique effectively suppresses coherent artifacts, enhancing MRI image quality.