一种有效的低场磁共振系统的水脂分离成像方法

Efficient Water Fat Separation Algorithm for Low Field Magnetic Resonance Imaging System

提出了一种基于Dixon算法的磁共振水脂分离成像改进方法.该方法分别采集人体组织中的水和脂肪同相位和反相位的磁共振图像,利用化学位移效应所产生的相位差分离出水和脂肪的图像.采用了改进的区域增长算法进行相位校正,为了在低场、较高噪声时获得更好的稳定性,为感兴趣区域和背景噪声分别开辟两类不同的多个堆栈,并设定感兴趣区域的增长优先级总是高于背景噪声,使得由背景噪声引起的误差传播被有效抑制.实验结果证明,改进算法在低场高噪声条件下比原算法更稳定,并在0.3 T低场永磁型磁共振系统上对水模、膝关节与人脑等部位获得了满意的水脂分离图像.

An efficient method of water fat separation for the magnetic resonance imaging （MRI） is presented. Based on the Dixon algorithm, where in-phase and opposed phase images are acquired by spin echo （SE） sequence. From these images the water-only and fat-only images can be extracted following the chemical shift principle. To obtain better performance under high noise condition on the low field MRI system, two different types of multiple stacks are created respectively for object region and background region. Pixels in background regions always have an inferior priority during region growing, so the errors due to background noises are restrained effectively. The modified method is more robust, and experiments on 0. 3 T system show the satisfying results.