Uni-COAL: A unified framework for cross-modality synthesis and super-resolution of MR images

Zhiyun Song, Zengxin Qi, Xin Wang, Xiangyu Zhao, Zhenrong Shen, Sheng Wang, Manman Fei, Zhe Wang, Di Zang, Dongdong Chen, Linlin Yao, Mengjun Liu, Qian Wang, Xuehai Wu, Lichi Zhang

Abstract: Highlights•Propose a unified framework that flexibly performs cross-modality synthesis (CMS), super-resolution (SR), and their combination (CMSR) of MR images.•Introduce an alias-free generator based on the Shannon-Nyquist framework for improved detail restoration.•Leverage the Segment Anything Model (SAM) for accurate anatomical structure preservation during synthesis.•Demonstrate superior performance across multiple CMS, SR, and CMSR tasks on diverse MRI datasets.