Abstract:<h3> Purpose</h3><p>Deep‐neural‐network‐based image reconstruction has demonstrated promising performance in medical imaging for under‐sampled and low‐dose scenarios. However, it requires large amount of memory and extensive time for the training. It is especially challenging to train the reconstruction networks for three‐dimensional computed tomography (CT) because of the high resolution of CT images. The purpose of this work is to reduce the memory and time consumption of the training of the reconstruction networks for CT to make it practical for current hardware, while maintaining the quality of the reconstructed images.</p><h3> Methods</h3><p>We unrolled the proximal gradient descent algorithm for iterative image reconstruction to finite iterations and replaced the terms related to the penalty function with trainable convolutional neural networks (CNN). The network was trained greedily iteration by iteration in the image‐domain on patches, which requires reasonable amount of memory and time on mainstream graphics processing unit (GPU). To overcome the local‐minimum problem caused by greedy learning, we used deep UNet as the CNN and incorporated separable quadratic surrogate with ordered subsets for data fidelity, so that the solution could escape from easy local minimums and achieve better image quality.</p><h3> Results</h3><p>The proposed method achieved comparable image quality with state‐of‐the‐art neural network for CT image reconstruction on 2D sparse‐view and limited‐angle problems on the low‐dose CT challenge dataset. The difference in root‐mean‐square‐error (RMSE) and structural similarity index (SSIM) was within [−0.23,0.47]HU and [0,0.001] respectively with 95% confidence level. For 3D image reconstruction with ordinary‐size CT volume, the proposed method only needed 2 GB graphics processing unit (GPU) memory and 0.45 seconds per training iteration as minimum requirement, whereas existing methods may require 417 GB and 31 minutes. The proposed method achieved improved performance compared to total‐variation‐ and dictionary‐learning‐based iterative reconstruction for both 2D and 3D problems.</p><h3> Conclusions</h3><p>We proposed a training‐time computationally efficient neural network for CT image reconstruction. The proposed method achieved comparable image quality with state‐of‐the‐art neural network for CT reconstruction, with significantly reduced memory and time requirement during training. The proposed method is applicable to 3D image reconstruction problems such as cone‐beam CT and tomosynthesis on mainstream GPUs.</p><p>This article is protected by copyright. All rights reserved.</p>

Hybrid U-Net and Swin-transformer network for limited-angle cardiac computed tomography

Deep Fusion Network Based Sparse View CT Reconstructions for Clinical Diagnostic Scanners

Single-view Cone Beam CT Reconstruction with Swin Transformer Based Deep Learning.

TF-Unet:An Automatic Cardiac MRI Image Segmentation Method

Deep-Learning-Based Few-Angle Cardiac SPECT Reconstruction Using Transformer

Transformer-based Dual-domain Network for Few-view Dedicated Cardiac SPECT Image Reconstructions

Comparison of CNN-based and Transformer-based Approaches for Sparse-view CT Reconstruction

Dual-Domain Reconstruction Network Incorporating Multi-Level Wavelet Transform and Recurrent Convolution for Sparse View Computed Tomography Imaging

CoST-UNet: Convolution and swin transformer based deep learning architecture for cardiac segmentation

Improving coronary artery imaging in single source CT with cardiac motion correction using attention and spatial transformer based neural networks

Low-dose CT image reconstruction based on a deep neural network

Re-UNet: A Novel Multi-scale Reverse U-shaped Network Architecture for Low-dose CT Image Reconstruction

Hybrid-Domain Integrative Transformer Iterative Network for Spectral CT Imaging

A Deep Learning Reconstruction Method for Sparsely Spaced Multidetector Helical CT Using Cascaded Dual-Domain CNN

Image-based Scatter Correction for Cone-Beam CT Using Flip Swin Transformer U-shape Network

Deep Spatial-Temporal Convolutional Neural Networks for Medical Image Restoration

Computationally efficient deep neural network for computed tomography image reconstruction

Cardiac MRI segmentation of the atria based on UU-NET

Re-UNet: a novel multi-scale reverse U-shape network architecture for low-dose CT image reconstruction

A model-guided deep network for limited-angle computed tomography

Synthesizing CT from Ultrashort Echo-Time MR Images via Convolutional Neural Networks