分布式超快核磁共振成像的数学理论与算法

The deployment of low-field based distributed magnetic resonance imaging (MRI) system is an important measure to implement the strategy of inclusive health care. However, the low signal-to-noise ratio and slow scanning speed of this new model hinder its practical applications. Therefore, it is highly desired to study innovative theories of ultra-fast speed magnetic resonance imaging (MRI) and design corresponding algorithms as soon as possible..Focusing on the application requirements of low-field based ultra-fast MRI, this project aims at the innovative model of distributed MRI, and studies the innovative mathematical theories of inverse problems to provide theoretical bases for distributed ultra-fast MRI. In particular, this project proposes to study the learned regularization theories and mathematical methods for ultra-fast MRI, combined with the physical imaging mechanisms, to realize high-quality MRI image reconstruction with low-field and high acceleration ratio. This project will also explore the correlations between multi-contrast images and develop mathematical methods for single pulse sequence guided multi-contrast imaging based on the anatomical structure-consistency. Meanwhile, the software system of distributed ultra-fast MRI will be developed and integrated with domestic MR equipment to realize pilot deployment, so as to provide technical-supports for the medical-health of Chinese people. .All in all, the implementation of this project will provide new knowledge and technologies for a deeper understanding of the learned regularization and the development of fast MRI, which is of important scientific significance and application value.

基于低场系统的分布式核磁共振成像模式是实行普惠医疗战略的重要举措，但是这一创新模式下磁共振信号的低信噪比和扫描速度慢严重阻碍了其可行性，开展低场下超快磁共振成像的理论与算法研究成为迫切需求。本项目面向低场与超快磁共振成像的应用需求，研究创新的反问题求解数学理论，为分布式超快磁共振成像提供理论基础。具体而言，项目将结合物理成像机制，研究模型数据双驱动的可学习正则化理论及超快磁共振成像数学方法，实现低场高加速倍数下的高质量磁共振图像重建；研究多对比度磁共振图像的相关性，并基于解剖结构不变性建模，发展单序列引导的多对比度快速磁共振成像数学方法；同时研发分布式超快磁共振成像软件系统，并与国产磁共振设备集成，实现试点部署，为普惠我国民众的医疗健康提供技术支撑。本项目的实施将为深入理解可学习正则化成像理论，以及发展快速磁共振成像方法提供新知识和新技术，因此具有重要的科学意义和应用价值。

基于低场系统的分布式核磁共振成像模式是实行普惠医疗战略的重要举措，但是这一创新模式下磁共振信号的低信噪比和扫描速度慢严重阻碍了其可行性，开展低场下超快磁共振成像的理论与算法研究成为迫切需求。本项目面向低场与超快磁共振成像的应用需求，研究创新的反问题求解数学理论，为分布式超快磁共振成像提供理论基础。具体而言，项目将结合物理成像机制，研究模型数据双驱动的可学习正则化理论及超快磁共振成像数学方法，实现低场高加速倍数下的高质量磁共振图像重建；研究多对比度磁共振图像的相关性，并基于解剖结构不变性建模，发展单序列引导的多对比度快速磁共振成像数学方法。本项目已发表SCI期刊论文9篇，CCF-B类会议论文2篇，5项研究成果在国际医学磁共振学会年会ISMRM上进行了展示(3项oral power pitch)，申请发明专利5项，培养博士生3名，硕士生5名。