Fig. 3.

Examples of deep learning-based motion artifact reduction. The motion artifacts in the acquired images (upper row) were reduced (lower row). The residual components are shown in the middle row (adapted from Fig. 10 of Tamada et al.⁵⁰).