Deep Convolutional Neural Network for Inverse Problems in Imaging
Deep Convolutional Neural Network for Inverse Problems in Imaging
复制标题
DOI：
10.1109/tip.2017.2713099
复制
发表时间：
2017-09-01
影响因子：
10.6
通讯作者：
Unser, Michael
中科院分区：
文献类型：
--
作者：
Jin, Kyong Hwan;McCann, Michael T.;Unser, Michael
本文提出了一种新的基于深度卷积神经网络（CNN）的病态逆问题求解算法。在过去的几十年里，正则化迭代算法已经成为求解病态逆问题的标准方法。这些方法产生了出色的结果，但由于前向和伴随算子的高计算成本以及超参数选择的困难等因素，在实践中部署可能具有挑战性。本文的出发点是观察到当正演模型的法算子（H* H，其中H*为正演成像算子H的伴随算子）为卷积时，展开迭代方法具有CNN（滤波后逐点非线性）的形式。基于这一观察结果，我们建议使用直接反演，然后使用CNN来解决正卷积反问题。直接反演封装了系统的物理模型，但在问题不恰当的情况下会导致伪命题；CNN将多分辨率分解和残差学习相结合，在保持图像结构的同时学习去除这些伪影。我们展示了所提出的网络在合成幻象的平行束x射线计算机断层扫描以及真实实验图的稀疏视图重建（低至50视图）中的性能。对于更逼真的幻影，所提出的网络优于总变差正则化迭代重建，并且在GPU上重构512 x 512图像需要不到一秒的时间。
In this paper, we propose a novel deep convolutional neural network (CNN)-based algorithm for solving ill-posed inverse problems. Regularized iterative algorithms have emerged as the standard approach to ill-posed inverse problems in the past few decades. These methods produce excellent results, but can be challenging to deploy in practice due to factors including the high computational cost of the forward and adjoint operators and the difficulty of hyperparameter selection. The starting point of this paper is the observation that unrolled iterative methods have the form of a CNN (filtering followed by pointwise non-linearity) when the normal operator (H* H, where H* is the adjoint of the forward imaging operator, H) of the forward model is a convolution. Based on this observation, we propose using direct inversion followed by a CNN to solve normal-convolutional inverse problems. The direct inversion encapsulates the physical model of the system, but leads to artifacts when the problem is ill posed; the CNN combines multiresolution decomposition and residual learning in order to learn to remove these artifacts while preserving image structure. We demonstrate the performance of the proposed network in sparse-view reconstruction (down to 50 views) on parallel beam X-ray computed tomography in synthetic phantoms as well as in real experimental sinograms. The proposed network outperforms total variation-regularized iterative reconstruction for the more realistic phantoms and requires less than a second to reconstruct a 512 x 512 image on the GPU.