Multimodal magnetic resonance imaging synthesis via cross-modal feature fusion and hierarchical feature extraction

The proposed framework achieves high structural fidelity, fine-detail preservation, and robustness to motion artifacts in multimodal MRI synthesis. These findings confirm its effectiveness for generating reliable synthetic modalities, while future work will extend validation to additional datasets to further assess generalization and clinical applicability.