压缩传感理论在磁共振成像技术中的应用

The Application of Compressed Sensing in Magnetic Resonance Imaging

这篇文章主要介绍了压缩传感(Compressed Sensing,CS)理论在磁共振成像(Magnetic Resonance Imaging,MRI)技术中的应用,这一技术可简称为CS-MRI。通常情况下,MRI的采集速度相对比较慢,从而限制了它的进一步发展,而CS-MRI可以在只有很少的MRI采集信号的情况下精确地重建出组织影像,从而大大提高采集效率,而且这种方法还具有很强的去噪能力。因此,可以利用CS-MRI在相同的空间分辨率下获得更快的成像速度,或者在同样的时间分辨率下获得更精细的组织影像。为了实现CS-MRI,必须对传统的信号采集方式和数据处理方式进行修改。本文首先从总体上概况了CS-MRI的理论基础,然后分别从稀疏变换,不相干欠采样和非线性重建算法三个方面具体阐述了它的的具体实现方法,最后对CS-MRI的超强的去噪能力进行了解释。

The paper mainly introduces the application of Compressed Sensing （CS） in Magnetic Resonance Imaging （MRI）, which is called CS-MRI for short. In general, the acquisition speed is relatedly slow, which limits its development to some extent. However, CS-MRI has the ability to reconstructing tissue image with few MRI data, which can enhance MRrs acquisition efficiency, as well as the strong ability to defending noise. Therefore, we can achieve either faster imaging speed under the same space resolution, or higher image quality under the same time resolution. To apply CS-MRI successfully, new methods of signal sampling and data processing should be developed. Firstly, the basic theory of CS-MRI is demonstrated. Then, we decompose the process to three particular parts： sparse transform, incoherent under-sampling, and non-linear reconstruction algorithm. In the end, why CS-MRI is able to denoise is explained.