Kim Kio, Habas Piotr, Rajagopalan Vidya, Scott Julia, Corbett-Detig James, Rousseau Francois, Glenn Orit, Barkovich James, Studholme Colin
Department of Radiology and Biomedical Imaging, University of California San Francisco, CA 94143, USA.
Annu Int Conf IEEE Eng Med Biol Soc. 2010;2010:879-82. doi: 10.1109/IEMBS.2010.5627876.
The slice intersection motion correction (SIMC) method is a powerful tool to compensate for motion that occurs during in utero acquisition of the multislice magnetic resonance (MR) images of the human fetal brain. The SIMC method makes use of the slice intersection intensity profiles of orthogonally planned slice pairs to simultaneously correct for the relative motion occurring between all the acquired slices. This approach is based on the assumption that the bias field is consistent between slices. However, for some clinical studies where there is a strong bias field combined with significant fetal motion relative to the coils, this assumption is broken and the resulting motion estimate and the reconstruction to a 3D volume can both contain errors. In this work, we propose a method to correct for the relative differences in bias field between all slice pairs. For this, we define the energy function as the mean square difference of the intersection profiles, that is then minimized with respect to the bias field parameters of the slices. A non iterative method which considers the relative bias between each slice simultaneously is used to efficiently remove inconsistencies. The method, when tested on synthetic simulations and actual clinical imaging studies where bias was an issue, brought a significant improvement to the final reconstructed image.
切片交叉运动校正（SIMC）方法是一种强大的工具，用于补偿在子宫内获取人类胎儿脑多切片磁共振（MR）图像期间发生的运动。SIMC方法利用正交规划切片对的切片交叉强度轮廓来同时校正所有采集切片之间发生的相对运动。该方法基于切片之间偏置场一致的假设。然而，对于一些临床研究，当存在强偏置场且胎儿相对于线圈有显著运动时，该假设不成立，由此产生的运动估计和三维体积重建都可能包含误差。在这项工作中，我们提出了一种校正所有切片对之间偏置场相对差异的方法。为此，我们将能量函数定义为交叉轮廓的均方差，然后针对切片的偏置场参数将其最小化。使用一种同时考虑每个切片之间相对偏差的非迭代方法来有效消除不一致性。该方法在合成模拟和存在偏置问题的实际临床成像研究中进行测试时，显著改善了最终重建图像。
