Accelerated MRI With Un-Trained Neural Networks
Accelerated MRI With Un-Trained Neural Networks
复制标题
DOI：
10.1109/tci.2021.3097596
复制
发表时间：
2020-07
影响因子：
5.4
通讯作者：
Mohammad Zalbagi Darestani;Reinhard Heckel
中科院分区：
文献类型：
--
作者：
Mohammad Zalbagi Darestani;Reinhard Heckel
卷积神经网络（CNN）对于图像重建问题非常有效。通常，CNN是在大量训练图像上训练的。然而，最近，未经训练的CNN（如深度图像先验和深度解码器）在图像重建问题（如去噪和修复）方面取得了优异的性能，而无需使用任何训练数据-除了利用一些样本进行超参数调整。出于这一发展的动机，我们解决了在加速MRI与未经训练的神经网络中出现的重建问题。我们提出了一种基于深度解码器变体的高度优化的未经训练的恢复方法，并表明它在重建性能方面明显优于其他未经训练的方法，特别是基于稀疏性的经典压缩感知方法和未经训练的神经网络的朴素应用。我们还比较了训练方法在理想设置中的性能（重建精度和计算成本），特别是在fastMRI数据集上，其中训练和测试数据来自相同的分布。在这里，我们发现我们的未经训练的算法实现了与基线训练的神经网络相似的性能，但最先进的训练网络优于未经训练的网络。最后，我们对训练和测试分布略有不同的非理想设置进行了比较，发现我们的未训练方法与最先进的加速MRI重建方法具有相似的性能。
Convolutional Neural Networks (CNNs) are highly effective for image reconstruction problems. Typically, CNNs are trained on large amounts of training images. Recently, however, un-trained CNNs such as the Deep Image Prior and Deep Decoder have achieved excellent performance for image reconstruction problems such as denoising and inpainting, without using any training data—except leveraging a few samples for hyper-parameter tuning. Motivated by this development, we address the reconstruction problem arising in accelerated MRI with un-trained neural networks. We propose a highly optimized un-trained recovery approach based on a variation of the Deep Decoder and show that it significantly outperforms other un-trained methods, in particular sparsity-based classical compressed sensing methods and naive applications of un-trained neural networks in terms of reconstruction performance. We also compare performance (both in terms of reconstruction accuracy and computational cost) in an ideal setup for trained methods, specifically on the fastMRI dataset, where the training and test data come from the same distribution. Here, we find that our un-trained algorithm achieves similar performance to a baseline trained neural network, but a state-of-the-art trained network outperforms the un-trained one. Finally, we perform a comparison on a non-ideal setup where the train and test distributions are slightly different, and find that our un-trained method achieves similar performance to a state-of-the-art accelerated MRI reconstruction method.