Model-Based Deep Learning PET Image Reconstruction Using Forward–Backward Splitting Expectation–Maximization
Model-Based Deep Learning PET Image Reconstruction Using Forward–Backward Splitting Expectation–Maximization
复制标题
DOI：
10.1109/trpms.2020.3004408
复制
发表时间：
2020-06-23
影响因子：
4.4
通讯作者：
Reader AJ
中科院分区：
文献类型：
--
作者：
Mehranian A;Reader AJ
我们提出了一种前向-后向分裂算法，将深度学习集成到最大后验概率（MAP）正电子发射断层扫描（PET）图像重建中。MAP重建分为正则化，期望最大化（EM）和加权融合。对于正则化，考虑使用Bowsher先验（使用马尔可夫随机场）或残差学习单元（使用卷积神经网络）。对于后者，我们提出的前向-后向分裂EM（FBSEM），用有序子集（OS）加速，被展开成一个递归神经网络，其中网络参数（包括正则化强度）在所有状态之间共享，并在PET重建期间学习。我们的网络使用仅PET（FBSEM-p）和PET-MR（FBSEM-pm）数据集进行训练和评估，用于低剂量模拟和短时间体内脑成像。将其与OSEM、Bowsher MAPEM以及在相同的仅PET（Unet-p）或PET-MR（Unet-pm）数据集上训练的重建后U-Net去噪进行了比较。对于模拟，FBSEM-p（m）和Unet-p（m）网实现了相当的性能，平均分别为14.4%和13.4%的归一化均方根误差（NRMSE）;并且都优于OSEM和MAPEM方法（分别为20.7%和17.7%的NRMSE）。对于体内数据集，FBSEM-p（m）、Unet-p（m）、MAPEM和OSEM方法在不同脑区中的平均均方根误差分别为3.9%、5.7%、5.9%和7.8%。总之，所研究的U-Net去噪方法实现了与FBSEM网络的代表性实现相当的性能。
We propose a forward-backward splitting algorithm to integrate deep learning into maximum-a-posteriori (MAP) positron emission tomography (PET) image reconstruction. The MAP reconstruction is split into regularization, expectation-maximization (EM), and a weighted fusion. For regularization, the use of either a Bowsher prior (using Markov-random fields) or a residual learning unit (using convolutional-neural networks) were considered. For the latter, our proposed forward-backward splitting EM (FBSEM), accelerated with ordered subsets (OS), was unrolled into a recurrent-neural network in which network parameters (including regularization strength) are shared across all states and learned during PET reconstruction. Our network was trained and evaluated using PET-only (FBSEM-p) and PET-MR (FBSEM-pm) datasets for low-dose simulations and short-duration in-vivo brain imaging. It was compared to OSEM, Bowsher MAPEM, and a post-reconstruction U-Net denoising trained on the same PET-only (Unet-p) or PET-MR (Unet-pm) datasets. For simulations, FBSEM-p(m) and Unet-p(m) nets achieved a comparable performance, on average, 14.4% and 13.4% normalized root-mean square error (NRMSE), respectively; and both outperformed OSEM and MAPEM methods (with 20.7% and 17.7% NRMSE, respectively). For in-vivo datasets, FBSEM-p(m), Unet-p(m), MAPEM, and OSEM methods achieved average root-sum-of-squared errors of 3.9%, 5.7%, 5.9%, and 7.8% in different brain regions, respectively. In conclusion, the studied U-Net denoising method achieved a comparable performance to a representative implementation of the FBSEM net.