基于正则化的SPACE-RIP并行磁共振成像重建算法

采用并行磁共振成像可以提高成像速度和图像分辨率,但这是以牺牲重建图像的信噪比为代价的。为此,该文提出了一种基于正则化的并行磁共振成像重建算法,降低由于阵列线圈的几何相关而造成的信噪比损失。以预扫描图像作为最后重建图像的先验信息,用L曲线方法求解最佳正则化参数。实验表明,该方法能较大限度地减弱噪声对重建结果的影响,当SPACE-RIP并行成像技术的加速因子达到4时,仍可得到高质量的高分辨率重建影像。重建后的图像质量良好,对临床诊断具有较高实用价值。

Regularized SPACE-RIP parallel magnetic resonance imaging reconstruction method

Parallel magnetic resonance imaging provides reduced magnetic resonance image acquisition time and increased spatiotemporal resolution,but it suffers from degradation of the signal-to-noise ratio(SNR)in the final reconstructed image.A regularization-based reconstruction algorithm was developed to reduce the SNR loss due to the phased array coils geometric correlation.Prescaned images are used as a priori information for the final reconstructed image.The optimum regularization parameters are determined using the L-curve technique.The results indicate that the algorithm considerably reduces the noise influence on the reconstructed image.While maintaining the high spatiotemporal resolution for SPACE-RIP parallel imaging acceleration factors up to 4.The high quality reconstructed images show that the algorithm is indeed valuable for diagnosis.