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>

A deep-learning-based workflow for reconstructing and segmenting challenging sets of time-resolved X-ray micro-computed tomography data

Deep 3D reconstruction of synchrotron X-ray computed tomography for intact lungs

Machine-Learning-based Algorithms for Automated Image Segmentation Techniques of Transmission X-ray Microscopy (TXM)

Deep learning based workflow for accelerated industrial X-ray Computed Tomography

Deep Learning for Material Decomposition in Photon-Counting CT

Algorithm-driven Advances for Scientific CT Instruments: From Model-based to Deep Learning-based Approaches

Synthetic Data Generation for Automatic Segmentation of X-ray Computed Tomography Reconstructions of Complex Microstructures

μ-Net: A Deep Learning-Based Architecture for μ-CT Segmentation

Deep learning of multi-resolution X-Ray micro-CT images for multi-scale modelling

A tomographic workflow to enable deep learning for X-ray based foreign object detection

A workflow for segmenting soil and plant X-ray CT images with deep learning in Googles Colaboratory

A deep learning reconstruction framework for X-ray computed tomography with incomplete data

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

Deep Learning-Based Reconstruction of Interventional Tools from Four X-Ray Projections for Tomographic Interventional Guidance

Deep learning to overcome Zernike phase-contrast nanoCT artifacts for automated micro-nano porosity segmentation in bone

Computationally efficient deep neural network for computed tomography image reconstruction

A Self-Supervised Approach to Reconstruction in Sparse X-Ray Computed Tomography

High-performance deep learning segmentation for non-destructive testing of X-ray tomography

TomoGRAF: A Robust and Generalizable Reconstruction Network for Single-View Computed Tomography

Automated Segmentation of Computed Tomography Images with Submanifold Sparse Convolutional Networks

Limited Angle Tomography for Transmission X-Ray Microscopy Using Deep Learning