Freire Luis C, Gouveia Ana R, Godinho Fernando M
Escola Superior de Tecnologia da Saúde de Lisboa, Instituto Politécnico de Lisboa, 1990-096, Portugal.
Annu Int Conf IEEE Eng Med Biol Soc. 2010;2010:5624-7. doi: 10.1109/IEMBS.2010.5628038.
In this work, we present a neural network (NN) based method designed for 3D rigid-body registration of FMRI time series, which relies on a limited number of Fourier coefficients of the images to be aligned. These coefficients, which are comprised in a small cubic neighborhood located at the first octant of a 3D Fourier space (including the DC component), are then fed into six NN during the learning stage. Each NN yields the estimates of a registration parameter. The proposed method was assessed for 3D rigid-body transformations, using DC neighborhoods of different sizes. The mean absolute registration errors are of approximately 0.030 mm in translations and 0.030 deg in rotations, for the typical motion amplitudes encountered in FMRI studies. The construction of the training set and the learning stage are fast requiring, respectively, 90 s and 1 to 12 s, depending on the number of input and hidden units of the NN. We believe that NN-based approaches to the problem of FMRI registration can be of great interest in the future. For instance, NN relying on limited K-space data (possibly in navigation echoes) can be a valid solution to the problem of prospective (in frame) FMRI registration.
在这项工作中，我们提出了一种基于神经网络（NN）的方法，该方法专为功能磁共振成像（fMRI）时间序列的三维刚体配准而设计，它依赖于待对齐图像的有限数量的傅里叶系数。这些系数包含在位于三维傅里叶空间第一卦限（包括直流分量）的一个小立方邻域内，然后在学习阶段被输入到六个神经网络中。每个神经网络产生一个配准参数的估计值。使用不同大小的直流邻域对所提出的方法进行了三维刚体变换评估。对于fMRI研究中遇到的典型运动幅度，平移的平均绝对配准误差约为0.030毫米，旋转的平均绝对配准误差约为0.030度。训练集的构建和学习阶段速度很快，分别需要90秒和1到12秒，这取决于神经网络的输入和隐藏单元数量。我们相信，基于神经网络的fMRI配准问题方法在未来可能会非常有意义。例如，依赖有限K空间数据（可能在导航回波中）的神经网络可以成为前瞻性（帧内）fMRI配准问题的有效解决方案。
