Motion Correction with Combination of Disentangled Cycle-GAN and k-space Subsampling for the gadoxetic acid-enhanced liver MRI

Motivation: The gadoxetic acid-enhanced liver MRI is often accompanied by significant motion artifacts due to drug side effects. Goal(s): Motion correction by integrating Disentangled Cycle-GAN with the k-space Subsampling (DCGAN-kS) method. Approach: Convert motion correction to the image domain transfer problem resolved by DCGAN with the aid of the k-space subsampling strategy for reducing features and simplifying the domain transfer problem. Results: The method can effectively remove artifacts for the arterial phase imaging of the gadoxetic acid-enhanced liver MRI. Impact: The proposed scheme outperforms the other state-of-the-art methods for motion correction in the gadoxetic acid-enhanced liver MRI, which could enhance the image quality and reduce failed scanning.