AI-Empowered Fast Magnetic Resonance Imaging

As a non-invasive imaging modality, magnetic resonance imaging (MRI) has been widely utilized for examining the structure and functional status of human tissues and organs. Due to the physical constraints of the scanning devices, the data acquisition time is prolonged, which limits the spatial and temporal resolution of MRI. Moreover, subjects are typically required to remain still or hold their breath inside the scanner for certain imaging bodies. Therefore, the slow acquisition of MR images causes inconvenience for both patients and the medical system. Various methods to accelerate the speed of magnetic resonance imaging have been proposed to address this. Furthermore, with the rapid advancement of artificial intelligence, fast MRI methods based on deep learning have been a significant development. Leveraging the powerful knowledge extraction capabilities of deep neural networks, the quality of MR images is continuously improving. According to the roles of different imaging modalities, this chapter first introduces the roles of four MRI modalities (structural magnetic resonance imaging, quantitative magnetic resonance imaging, diffusion magnetic resonance imaging, and functional magnetic resonance imaging). Then, it provides a detailed overview of the research and progress of artificial intelligence in these four magnetic resonance imaging modalities. Finally, it summarizes and discusses possible future directions.