基于背景信息的磁共振成像尖峰噪声消除

磁共振成像扫描过程中如果在K空间中出现尖峰噪声，将会严重影响图像质量. 该文提出了一种利用磁共振图像背景信息来定位并修复尖峰噪声的方法. 首先将K空间数据沿着频率编码k_x和相位编码k_y 2个方向分别进行一维傅立叶变换，然后在获得的2个中间域数据中，分别截取出背景区域， 进行一维反傅立叶变换得到与背景对应的频率域数据，通过对背景频率域数据的分析，获得尖峰噪声的k_x和k_y坐标，最后对异常点进行修复. 该算法可以很好地消除各种模式的尖峰噪声，与已有的算法相比，能够更好地处理K空间中连续多个相位编码行上出现尖峰噪声的情况，有效地消除图像伪影.

Spike Noise Removal for Magnetic Resonance Imaging Based on the Information in the Background Region

In magnetic resonance imaging, spike noise in the K-space will degrade the quality of the reconstructed images severely. We proposed a new approach to locate and remove spike noises based on the information in the background region. Firstly, one-dimensional discrete Fourier transform (DFT) is applied to K-space data along the k_x and k_y direction respectively to acquire two half transformed datasets. Then background in these two datasets are extracted and analyzed to locate and remove the spike noise. By experiments on simulated and real data, the proposed approach was proven to be able to remove certain kinds of spike noises effectively. Compared with the previous algorithms, this approach has the unique advantage in that it can deal with the cases in which consecutive rows of K-space data are contaminated with spike noises.