One-step inverse generation network for sparse-view dual-energy CT reconstruction and material imaging.

OBJECTIVE: Sparse-view dual-energy spectral computed tomography (DECT) imaging is a challenging inverse problem. Due to the incompleteness of the collected data, the presence of streak artifacts can result in the degradation of reconstructed spectral images. The subsequent material decomposition task in DECT can further lead to the amplification of artifacts and noise. APPROACH: To address this problem, we propose a novel one-step inverse generation network (OIGN) for sparse-view dual- energy CT imaging, which can achieve simultaneous imaging of spectral images and materials. The entire OIGN consists of five sub-networks that form four modules, including the pre-reconstruction module, the pre-decomposition module, and the following residual filtering module and residual decomposition module. The residual feedback mechanism is introduced to synchronize the optimization of spectral CT images and materials. MAIN RESULTS: Numerical simulation experiments show that the OIGN has better performance on both reconstruction and material decomposition than other state-of-the-art spectral CT imaging algorithms. OIGN also demonstrates high imaging efficiency by completing two high-quality imaging tasks in just 50 seconds. Additionally, anti-noise testing is conducted to evaluate the robustness of OIGN. Significance. These findings have great potential in high-quality multi-task spectral CT imaging in clinical diagnosis.

MISD-IR: material-image subspace decomposition-based iterative reconstruction with spectrum estimation for dual-energy computed tomography.

Background: Dual-energy computed tomography (DECT) is a promising technique, which can provide unique capability for material quantification. The iterative reconstruction of material maps requires spectral information and its accuracy is affected by spectral mismatch. Simultaneously estimating the spectra and reconstructing material maps avoids extra workload on spectrum estimation and the negative impact of spectral mismatch. However, existing methods are not satisfactory in image detail preservation, edge retention, and convergence rate. The purpose of this paper was to mine the similarity between the reconstructed images and the material images to improve the imaging quality, and to design an effective iteration strategy to improve the convergence efficiency. Methods: The material-image subspace decomposition-based iterative reconstruction (MISD-IR) with spectrum estimation was proposed for DECT. MISD-IR is an optimized model combining spectral estimation and material reconstruction with fast convergence speed and promising noise suppression capability. We proposed to reconstruct the material maps based on extended simultaneous algebraic reconstruction techniques and estimation of the spectrum with model spectral. To stabilize the iteration and alleviate the influence of errors, we introduced a weighted proximal operator based on the block coordinate descending algorithm (WP-BCD). Furthermore, the reconstructed computed tomography (CT) images were introduced to suppress the noise based on subspace decomposition, which relies on non-local regularization to prevent noise accumulation. Results: In numerical experiments, the results of MISD-IR were closer to the ground truth compared with other methods. In real scanning data experiments, the results of MISD-IR showed sharper edges and details. Compared with other one-step iterative methods in the experiment, the running time of MISD-IR was reduced by 75%. Conclusions: The proposed MISD-IR can achieve accurate material decomposition (MD) without known energy spectrum in advance, and has good retention of image edges and details. Compared with other one-step iterative methods, it has high convergence efficiency.

A dual-energy CT reconstruction method based on anchor network from dual quarter scans.

Compared with conventional single-energy computed tomography (CT), dual-energy CT (DECT) provides better material differentiation but most DECT imaging systems require dual full-angle projection data at different X-ray spectra. Relaxing the requirement of data acquisition is an attractive research to promote the applications of DECT in wide range areas and reduce the radiation dose as low as reasonably achievable. In this work, we design a novel DECT imaging scheme with dual quarter scans and propose an efficient method to reconstruct the desired DECT images from the dual limited-angle projection data. We first study the characteristics of limited-angle artifacts under dual quarter scans scheme, and find that the negative and positive artifacts of DECT images are complementarily distributed in image domain because the corresponding X-rays of high- and low-energy scans are symmetric. Inspired by this finding, a fusion CT image is generated by integrating the limited-angle DECT images of dual quarter scans. This strategy enhances the true image information and suppresses the limited-angle artifacts, thereby restoring the image edges and inner structures. Utilizing the capability of neural network in the modeling of nonlinear problem, a novel Anchor network with single-entry double-out architecture is designed in this work to yield the desired DECT images from the generated fusion CT image. Experimental results on the simulated and real data verify the effectiveness of the proposed method. This work enables DECT on imaging configurations with half-scan and largely reduces scanning angles and radiation doses.

Direct Multi-Material Reconstruction via Iterative Proximal Adaptive Descent for Spectral CT Imaging.

Spectral computed tomography (spectral CT) is a promising medical imaging technology because of its ability to provide information on material characterization and quantification. However, with an increasing number of basis materials, the nonlinearity of measurements causes difficulty in decomposition. In addition, noise amplification and beam hardening further reduce image quality. Thus, improving the accuracy of material decomposition while suppressing noise is pivotal for spectral CT imaging. This paper proposes a one-step multi-material reconstruction model as well as an iterative proximal adaptive decent method. In this approach, a proximal step and a descent step with adaptive step size are designed under the forward-backward splitting framework. The convergence analysis of the algorithm is further discussed according to the convexity of the optimization objective function. For simulation experiments with different noise levels, the peak signal-to-noise ratio (PSNR) obtained by the proposed method increases approximately 23 dB, 14 dB, and 4 dB compared to those of other algorithms. Magnified areas of thorax data further demonstrated that the proposed method has a better ability to preserve details in tissues, bones, and lungs. Numerical experiments verify that the proposed method efficiently reconstructed the material maps, and reduced noise and beam hardening artifacts compared with the state-of-the-art methods.

Low-dose spectral reconstruction with global, local, and nonlocal priors based on subspace decomposition.

Background: Multienergy computed tomography (MECT) is a promising imaging modality for material decomposition, lesion detection, and other clinical applications. However, there is an urgent need to design efficient and accurate algorithms to solve the inverse problems related to spectral reconstruction and improve image quality, especially under low-dose and incomplete datasets. The key issue for MECT reconstruction is how to efficiently describe the interchannel and intrachannel priors in multichannel images. Methods: In this model, in order to correlate the similarities of interchannel images and regularize the multichannel images, the global, local, and nonlocal priors are jointly integrated into the low-dose MECT reconstruction model. First, the subspace decomposition method first employs the global low-rankness to map the original MECT images to the low-dimensional eigenimages. Then, nonlocal self-similarity of the eigenimages is cascaded into the optimization model. Additionally, the L0 quasi-norm on gradient images is incorporated into the proposed method to further enhance the local sparsity of intrachannel images. The alternating direction method is applied to solve the optimization model in an iterative scheme. Results: Simulation, preclinical, and real datasets were applied to validate the effectiveness of the proposed method. From the simulation dataset, the new method was found to reduce the root-mean-square error (RMSE) by 42.31% compared with the latest research fourth-order nonlocal tensor decomposition MECT reconstruction (FONT-SIR) method under 160 projection views. The calculation time of an iteration for the proposed method was 23.07% of the FONT-SIR method. The results of material decomposition in real mouse data further confirmed the accuracy of the proposed method for different materials. Conclusions: We developed a method in which the global, local, and nonlocal priors are jointly used to develop the reconstruction model for low-dose MECT, where the global low-rankness and nonlocal prior are cascaded by subspace decomposition and block-matching, and the L0 sparsity is applied to express the local prior. The results of the experiments demonstrate that the proposed method based on subspace improves computational efficiency and has advantages in noise suppression and structure preservation over competing algorithms.

Image-spectral decomposition extended-learning assisted by sparsity for multi-energy computed tomography reconstruction.

Background: Multi-energy computed tomography (CT) provides multiple channel-wise reconstructed images, and they can be used for material identification and k-edge imaging. Nonetheless, the projection datasets are frequently corrupted by various noises (e.g., electronic, Poisson) in the acquisition process, resulting in lower signal-noise-ratio (SNR) measurements. Multi-energy CT images have local sparsity, nonlocal self-similarity in spatial dimension, and correlation in spectral dimension. Methods: In this paper, we propose an image-spectral decomposition extended-learning assisted by sparsity (IDEAS) method to fully exploit these intrinsic priors for multi-energy CT image reconstruction. Particularly, a nonlocal low-rank Tucker decomposition (TD) is employed to utilize the correlation and nonlocal self-similarity priors. Moreover, considering the advantages of multi-task tensor dictionary learning (TDL) in sparse representation, an adaptive spatial dictionary and an adaptive spectral dictionary are trained during the iterative reconstruction process. Furthermore, a weighted total variation (TV) regularization term is employed to encourage local sparsity. Results: Numerical simulation, physical phantom, and preclinical mouse experiments are performed to validate the proposed IDEAS algorithm. Specifically, in the simulation experiments, the proposed IDEAS reconstructed high-quality images that are very close to the references. For example, the root mean square error (RMSE) of IDEAS image in energy bin 1 is as low as 0.0672, while the RMSE of other methods are higher than 0.0843. Besides, the structural similarity (SSIM) of IDEAS reconstructed image in energy bin 1 is greater than 0.98. For material decomposition, the RMSE of IDEAS bone component is as low as 0.0152, and other methods are higher than 0.0199. In addition, the computational cost of IDEAS is as low as 98.8 s for one iteration, and the competing tensor decomposition method is higher than 327 s. Conclusions: To further improve the quality of the reconstructed multi-energy CT images, multiple prior regularizations are introduced to the multi-energy CT reconstructed model, leading to an IDEAS method. Both qualitative and quantitative evaluation of our results confirm the outstanding performance of the proposed algorithm compared to the state-of-the-arts.

Super-resolution image reconstruction from sparsity regularization and deep residual-learned priors.

BACKGROUND: Computed tomography (CT) plays an important role in the field of non-destructive testing. However, conventional CT images often have blurred edge and unclear texture, which is not conducive to the follow-up medical diagnosis and industrial testing work. OBJECTIVE: This study aims to generate high-resolution CT images using a new CT super-resolution reconstruction method combining with the sparsity regularization and deep learning prior. METHODS: The new method reconstructs CT images through a reconstruction model incorporating image gradient L0-norm minimization and deep image priors using a plug-and-play super-resolution framework. The deep learning priors are learned from a deep residual network and then plugged into the proposed new framework, and alternating direction method of multipliers is utilized to optimize the iterative solution of the model. RESULTS: The simulation data analysis results show that the new method improves the signal-to-noise ratio (PSNR) by 7% and the modulation transfer function (MTF) curves show that the value of MTF50 increases by 0.02 factors compared with the result of deep plug-and-play super-resolution. Additionally, the real CT image data analysis results show that the new method improves the PSNR by 5.1% and MTF50 by 0.11 factors. CONCLUSION: Both simulation and real data experiments prove that the proposed new CT super-resolution method using deep learning priors can reconstruct CT images with lower noise and better detail recovery. This method is flexible, effective and extensive for low-resolution CT image super-resolution.

Projection domain processing for low-dose CT reconstruction based on subspace identification.

PURPOSE: Low-dose computed tomography (LDCT) has promising potential for dose reduction in medical applications, while suffering from low image quality caused by noise. Therefore, it is in urgent need for developing new algorithms to obtain high-quality images for LDCT. METHODS: This study tries to exploit the sparse and low-rank properties of images and proposes a new algorithm based on subspace identification. The collection of transmission data is sparsely represented by singular value decomposition and the eigen-images are then denoised by block-matching frames. Then, the projection is regularized by the correlation information under the frame of prior image compressed sensing (PICCS). With the application of a typical analytical algorithm on the processed projection, the target images are obtained. Both numerical simulations and real data verifications are carried out to test the proposed algorithm. The numerical simulations data is obtained based on real clinical scanning three-dimensional data and the real data is obtained by scanning experimental head phantom. RESULTS: In simulation experiment, using new algorithm boots the means of PSNR and SSIM by 1 dB and 0.05, respectively, compared with BM3D under the Gaussian noise with variance 0.04. Meanwhile, on the real data, the proposed algorithm exhibits superiority over compared algorithms in terms of noise suppression, detail preservation and computational overhead. The means of PSNR and SSIM are improved by 1.84 dB and 0.1, respectively, compared with BM3D under the Gaussian noise with variance 0.04. CONCLUSION: This study demonstrates the feasibility and advantages of a new algorithm based on subspace identification for LDCT. It exploits the similarity among three-dimensional data to improve the image quality in a concise way and shows a promising potential on future clinical diagnosis.

Total variation combining nonlocal means filtration for image reconstruction in X-ray computed tomography.

BACKGROUND: Image reconstruction for realistic medical images under incomplete observation is still one of the core tasks for computed tomography (CT). However, the stair-case artifacts of Total variation (TV) based ones have restricted the usage of the reconstructed images. OBJECTIVE: This work aims to propose and test an accurate and efficient algorithm to improve reconstruction quality under the idea of synergy between local and nonlocal regularizations. METHODS: The total variation combining the nonlocal means filtration is proposed and the alternating direction method of multipliers is utilized to develop an efficient algorithm. The first order approximation of linear expansion at intermediate point is applied to overcome the computation of the huge CT system matrix. RESULTS: The proposed method improves root mean squared error by 25.6% compared to the recent block-matching sparsity regularization (BMSR) on simulation dataset of 19 views. The structure similarities of image of the new method is higher than 0.95, while that of BMSR is about 0.92. Moreover, on real rabbit dataset of 20 views, the peak signal-to-noise ratio (PSNR) of the new method is 36.84, while using other methods PSNR are lower than 35.81. CONCLUSIONS: The proposed method shows advantages on noise suppression and detail preservations over the competing algorithms used in CT image reconstruction.

Framelet tensor sparsity with block matching for spectral CT reconstruction.

PURPOSE: Spectral computed tomography (CT) based on the photon-counting detection system has the capability to produce energy-discriminative attenuation maps of objects with a single scan. However, the insufficiency of photons collected into the narrow energy bins results in high quantum noise levels causing low image quality. This work aims to improve spectral CT image quality by developing a novel regularization based on framelet tensor prior. METHODS: First, similar patches are extracted from highly correlated interchannel images in spectral and spatial domains, and stacked to form a third-order tensor after vectorization along the energy channels. Second, the framelet tensor nuclear norm (FTNN) is introduced and applied to construct the regularization to exploit the sparsity embedded in nonlocal similarity of spectral images, and thus the reconstruction problem is modeled as a constrained optimization. Third, an iterative algorithm is proposed by utilizing the alternating direction method of multipliers framework in which efficient solvers are developed for each subproblem. RESULTS: Both numerical simulations and real data verifications were performed to evaluate and validate the proposed FTNN-based method. Compared to the analytic, TV-based, and the state-of-the-art tensor-based methods, the proposed method achieves higher numerical accuracy on both reconstructed CT images and decomposed material maps in the mouse data indicating the capability in noise suppression and detail preservation of the proposed method. CONCLUSIONS: A framelet tensor sparsity-based iterative algorithm is proposed for spectral reconstruction. The qualitative and quantitative comparisons show a promising improvement of image quality, indicating its promising potential in spectral CT imaging.

One half-scan dual-energy CT imaging using the Dual-domain Dual-way Estimated Network (DoDa-Net) model.

BACKGROUND: Compared with single-energy computed tomography (CT), dual-energy CT (DECT) can distinguish materials better. However, most DECT reconstruction theories require two full-scan projection datasets of different energies, and this requirement is hard to meet, especially for cases where a physical blockage disables a full circular rotation. Thus, it is critical to relax the requirements of data acquisition to promote the application of DECT. METHODS: A flexible one half-scan DECT scheme is proposed, which acquires two projection datasets on two-quarter arcs (one for each energy). The limited-angle problem of the one half-scan DECT scheme can be solved by a reconstruction method. Thus, a dual-domain dual-way estimation network called DoDa-Net is proposed by utilizing the ability of deep learning in non-linear mapping. Specifically, the dual-way mapping Generative Adversarial Network (DM-GAN) was first designed to mine the relationship between two different energy projection data. Two half-scan projection datasets were obtained, the data of which was twice that of the original projection dataset. Furthermore, the data transformation from the projection domain to the image domain was realized by the total variation (TV)-based method. In addition, the image processing network (Im-Net) was employed to optimize the image domain data. RESULTS: The proposed method was applied to a digital phantom and real anthropomorphic head phantom data to verify its effectiveness. The reconstruction results of the real data are encouraging and prove the proposed method's ability to suppress noise while preserving image details. Also, the experiments conducted on simulated data show that the proposed method obtains the closest results to the ground truth among the comparison methods. For low- and high-energy reconstruction, the peak signal-to-noise ratio (PSNR) of the proposed method is as high as 40.3899 and 40.5573 dB, while the PSNR of other methods is lower than 36.5200 dB. Compared with FBP, TV, and other GAN-based methods, the proposed method reduces root mean square error (RMSE) by, respectively, 0.0124, 0.0037, and 0.0016 for low-energy reconstruction, and 0.0102, 0.0028, and 0.0015 for high-energy reconstruction. CONCLUSIONS: The developed DoDa-Net model for the proposed one half-scan DECT scheme consists of two stages. In stage one, DM-GAN is used to realize the dual map of projection data. In stage two, the TV-based method is employed to transform the data from the projection domain to the image domain. Furthermore, the reconstructed image is processed by the Im-Net. According to the experimental results of qualitative and quantitative evaluation, the proposed method has advantages in detail preservation, indicating the potential of the proposed method in one half-scan DECT reconstruction.

Multi-segment spectral reconstruction via zero-value set prior.

Most scanning schemes of multi-energy computed tomography (MECT) require multiple sets of full-scan measurements under different x-ray spectra, which limits the application of MECT with incomplete scan. To handle this problem, a flexible MECT scanning strategy is proposed in this paper, which divides one half scan into three curves. Also, a novel MECT reconstruction algorithm is developed to relax the requirement of data acquisition of MECT. For MECT, gradient images of CT images at different energies ideally share the same position of zero-value set (Pos-OS) for the same object. Based on this observation, the characteristics of limited-angle artifacts is first explored, and it is found that the limited-angle artifacts in the image domain are closely related to the angle trajectory of the scan. Inspired by this discovery, the Pos-OS of the gradient image from the fusion CT image is extracted, and it is incorporated as prior knowledge into the TV minimization model in the form of equality constraints. The alternating direction method is exploited to solve the improved optimization model iteratively. Based on this, the proposed algorithm is derived to eliminate the limited angle artifacts in the image domain.The experimental results show that the proposed method achieves higher reconstruction quality under the designed scanning configuration than other methods in the literature.

Dual residual convolutional neural network (DRCNN) for low-dose CT imaging.

The excessive radiation doses in the application of computed tomography (CT) technology pose a threat to the health of patients. However, applying a low radiation dose in CT can result in severe artifacts and noise in the captured images, thus affecting the diagnosis. Therefore, in this study, we investigate a dual residual convolution neural network (DRCNN) for low-dose CT (LDCT) imaging, whereby the CT images are reconstructed directly from the sinogram by integrating analytical domain transformations, thus reducing the loss of projection information. With this new framework, feature extraction is performed simultaneously on both the sinogram-domain sub-net and the image-domain sub-net, which utilize the residual shortcut networks and play a complementary role in suppressing the projection noise and reducing image error. This new DRCNN approach helps not only decrease the sinogram noise but also preserve significant structural information. The experimental results of simulated and real projection data demonstrate that our DRCNN achieve superior performance over other state-of-art methods in terms of visual inspection and quantitative metrics. For example, comparing with RED-CNN and DP-ResNet, the value of PSNR using our DRCNN is improved by nearly 3 dB and 1 dB, respectively.

An effective sinogram inpainting for complementary limited-angle dual-energy computed tomography imaging using generative adversarial networks.

Dual-energy computed tomography (DECT) provides more anatomical and functional information for image diagnosis. Presently, the popular DECT imaging systems need to scan at least full angle (i.e., 360°). In this study, we propose a DECT using complementary limited-angle scan (DECT-CL) technology to reduce the radiation dose and compress the spatial distribution of the imaging system. The dual-energy total scan is 180°, where the low- and high-energy scan range is the first 90° and last 90°, respectively. We describe this dual limited-angle problem as a complementary limited-angle problem, which is challenging to obtain high-quality images using traditional reconstruction algorithms. Furthermore, a complementary-sinogram-inpainting generative adversarial networks (CSI-GAN) with a sinogram loss is proposed to inpainting sinogram to suppress the singularity of truncated sinogram. The sinogram loss focuses on the data distribution of the generated sinogram while approaching the target sinogram. We use the simultaneous algebraic reconstruction technique namely, a total variable (SART-TV) algorithms for image reconstruction. Then, taking reconstructed CT images of pleural and cranial cavity slices as examples, we evaluate the performance of our method and numerically compare different methods based on root mean square error (RMSE), peak signal-to-noise ratio (PSNR) and structural similarity (SSIM). Compared with traditional algorithms, the proposed network shows advantages in numerical terms. Compared with Patch-GAN, the proposed network can also reduce the RMSE of the reconstruction results by an average of 40% and increase the PSNR by an average of 26%. In conclusion, both qualitative and quantitative comparison and analysis demonstrate that our proposed method achieves a good artifact suppression effect and can suitably solve the complementary limited-angle problem.

Multi-energy CT reconstruction using tensor nonlocal similarity and spatial sparsity regularization.

BACKGROUND: Multi-energy computed tomography (MECT) based on a photon-counting detector is an emerging imaging modality that collects projections at several energy bins with a single scan. However, the limited number of photons collected into the divided, narrow energy bins results in high quantum noise levels in reconstructed images. This study aims to improve MECT image quality by minimizing noise levels while retaining image details. METHODS: A novel MECT reconstruction method was proposed by exploiting the nonlocal tensor similarity among interchannel images and spatial sparsity in single-channel images. Similar patches were initially extracted from the interchannel images in spectral and spatial domains, then stacked into a new three-order tensor. Intrinsic tensor sparsity regularization that combined the Tuker and canonical polyadic (CP) low-rank decomposition techniques were applied to exploit the nonlocal similarity of the formulated tensor. Spatial sparsity in single-channel images was modeled by total variation (TV) regularization that utilizes the compressibility of gradient image. A new MECT reconstruction model was established by simultaneously incorporating the intrinsic tensor sparsity and TV regularizations. The iterative alternating minimization method was utilized to solve the reconstruction model based on a flexible framework. RESULTS: The proposed method was applied to the digital phantom and real mouse data to assess its feasibility and reliability. The reconstruction and decomposition results in the mouse data were encouraging and demonstrated the ability of the proposed method in noise suppression while preserving image details, not observed with other methods. Imaging data from the digital phantom illustrated this method as achieving the best intuitive reconstruction and decomposition results among all compared methods. They reduced the root mean square error (RMSE) by 89.75%, 50.75%, and 36.54% on the reconstructed images compared with analytic, TV-based, and tensor-based methods, respectively. This phenomenon was also observed with decomposition results, where the RMSE was also reduced by 97.96%, 67.74%, 72.05%, respectively. CONCLUSIONS: In this study, we proposed a reconstruction method for photon counting detector-based MECT, using the intrinsic tensor sparsity and TV regularizations. Improvements in noise suppression and detail preservation in the digital phantom and real mouse data were validated by the qualitative and quantitative evaluations on the reconstruction and decomposition results, verifying the potential of the proposed method in MECT reconstruction.

Generative adversarial network-based sinogram super-resolution for computed tomography imaging.

Compared with the conventional 1×1 acquisition mode of projection in computed tomography (CT) image reconstruction, the 2×2 acquisition mode improves the collection efficiency of the projection and reduces the x-ray exposure time. However, the collected projection based on the 2×2 acquisition mode has low resolution (LR) and the reconstructed image quality is poor, thus limiting the use of this mode in CT imaging systems. In this study, a novel sinogram-super-resolution (SR) generative adversarial network model is proposed to obtain high-resolution (HR) sinograms from LR sinograms, thereby improving the reconstruction image quality under the 2×2 acquisition mode. The proposed generator is based on the residual network for LR sinogram feature extraction and SR sinogram generation. A relativistic discriminator is designed to render the network capable of obtaining more realistic SR sinograms. Moreover, we combine the cycle consistency loss, sinogram domain loss, and reconstruction image domain loss in the total loss function to supervise SR sinogram generation. Then, a trained model can be obtained by inputting the paired LR/HR sinograms into the network. Finally, the classic filtered-back-projection reconstruction algorithm is used for CT image reconstruction based on the generated SR sinogram. The qualitative and quantitative results of evaluations on digital and real data illustrate that the proposed model not only obtains clean SR sinograms from noisy LR sinograms but also outperforms its counterparts.

Unpaired Low-Dose CT Denoising Network Based on Cycle-Consistent Generative Adversarial Network with Prior Image Information.

The widespread application of X-ray computed tomography (CT) in clinical diagnosis has led to increasing public concern regarding excessive radiation dose administered to patients. However, reducing the radiation dose will inevitably cause server noise and affect radiologists' judgment and confidence. Hence, progressive low-dose CT (LDCT) image reconstruction methods must be developed to improve image quality. Over the past two years, deep learning-based approaches have shown impressive performance in noise reduction for LDCT images. Most existing deep learning-based approaches usually require the paired training dataset which the LDCT images correspond to the normal-dose CT (NDCT) images one-to-one, but the acquisition of well-paired datasets requires multiple scans, resulting the increase of radiation dose. Therefore, well-paired datasets are not readily available. To resolve this problem, this paper proposes an unpaired LDCT image denoising network based on cycle generative adversarial networks (CycleGAN) with prior image information which does not require a one-to-one training dataset. In this method, cyclic loss, an important trick in unpaired image-to-image translation, promises to map the distribution from LDCT to NDCT by using unpaired training data. Furthermore, to guarantee the accurate correspondence of the image content between the output and NDCT, the prior information obtained from the result preprocessed using the LDCT image is integrated into the network to supervise the generation of content. Given the map of distribution through the cyclic loss and the supervision of content through the prior image loss, our proposed method can not only reduce the image noise but also retain critical information. Real-data experiments were carried out to test the performance of the proposed method. The peak signal-to-noise ratio (PSNR) improves by more than 3 dB, and the structural similarity (SSIM) increases when compared with the original CycleGAN without prior information. The real LDCT data experiment demonstrates the superiority of the proposed method according to both visual inspection and quantitative evaluation.

Promising Generative Adversarial Network Based Sinogram Inpainting Method for Ultra-Limited-Angle Computed Tomography Imaging.

Limited-angle computed tomography (CT) image reconstruction is a challenging problem in the field of CT imaging. In some special applications, limited by the geometric space and mechanical structure of the imaging system, projections can only be collected with a scanning range of less than 90°. We call this kind of serious limited-angle problem the ultra-limited-angle problem, which is difficult to effectively alleviate by traditional iterative reconstruction algorithms. With the development of deep learning, the generative adversarial network (GAN) performs well in image inpainting tasks and can add effective image information to restore missing parts of an image. In this study, given the characteristic of GAN to generate missing information, the sinogram-inpainting-GAN (SI-GAN) is proposed to restore missing sinogram data to suppress the singularity of the truncated sinogram for ultra-limited-angle reconstruction. We propose the U-Net generator and patch-design discriminator in SI-GAN to make the network suitable for standard medical CT images. Furthermore, we propose a joint projection domain and image domain loss function, in which the weighted image domain loss can be added by the back-projection operation. Then, by inputting a paired limited-angle/180° sinogram into the network for training, we can obtain the trained model, which has extracted the continuity feature of sinogram data. Finally, the classic CT reconstruction method is used to reconstruct the images after obtaining the estimated sinograms. The simulation studies and actual data experiments indicate that the proposed method performed well to reduce the serious artifacts caused by ultra-limited-angle scanning.

Image domain dual material decomposition for dual-energy CT using butterfly network.

PURPOSE: Dual-energy CT (DECT) has been increasingly used in imaging applications because of its capability for material differentiation. However, material decomposition suffers from magnified noise from two CT images of independent scans, leading to severe degradation of image quality. Existing algorithms exhibit suboptimal decomposition performance because they fail to fully depict the mapping relationship between DECT images and basis materials under noisy conditions. Convolutional neural network exhibits great promise in the modeling of data coupling and has recently become an important technique in medical imaging application. Inspired by its impressive potential, we developed a new Butterfly network to perform the image domain dual material decomposition. METHODS: The Butterfly network is derived from the model of image domain DECT decomposition by exploring the geometric relationship between the mapping functions of the data model and network components. The network is designed as the double-entry double-out crossover architecture based on the decomposition formulation. It enters a pair of dual-energy images as inputs and defines the ground true decomposed images as each label. The crossover architecture, which plays an important role in material decomposition, is designed to implement the information exchange between the two material generation pathways in the network. The proposed network is further applied on the digital phantom and clinical data to evaluate its performance. RESULTS: The qualitative and quantitative evaluations of the material decomposition of digital phantoms and clinical data indicate that the proposed network outperforms its counterparts. For the digital phantom, the proposed network reduces the standard deviation (SD) of noise in tissue, bone, and mixture regions by an average of 95.75% and 86.58% compared with the direct matrix inversion and the conventional iterative method, respectively. The line profiles and image biases of the decomposition results of digital phantom indicate that the proposed network provides the decomposition results closest to the ground truth. The proposed network reduces the SD of the noise in decomposed images of clinical head data by over 90% and 80% compared with the direct matrix inversion and conventional iterative method, respectively. As the modulation transfer function decreases to 50%, the proposed network increases the spatial resolution by average factors of 1.34 and 1.17 compared with the direct matrix inversion and conventional iterative methods, respectively. The proposed network is further applied to the clinical abdomen data. Among the three methods, the proposed method received the highest score from six radiologists in the visual inspection of noise suppression in the clinical data. CONCLUSIONS: We develop a model-based Butterfly network to perform image domain material decomposition for DECT. The decomposition results of digital phantom validate its capability of decomposing two basis materials from DECT images. The proposed approach also leads to higher decomposition quality in noise suppression on clinical datasets as compared with those using conventional schemes.

Efficient solving algorithm for determining the exact sampling condition of limited-angle computed tomography reconstruction.

Total variation (TV) regularization-based iterative reconstruction algorithms have an impressive potential to solve limited-angle computed tomography with insufficient sampling projections. The analysis of exact reconstruction sampling conditions for a TV-minimization reconstruction model can determine the minimum number of scanning angle and minimize the scanning range. However, the large-scale matrix operations caused by increased testing phantom size are the computation bottleneck in determining the exact reconstruction sampling conditions in practice. When the size of the testing phantom increases to a certain scale, it is very difficult to analyze quantitatively the exact reconstruction sampling condition using existing methods. In this paper, we propose a fast and efficient algorithm to determine the exact reconstruction sampling condition for large phantoms. Specifically, the sampling condition of a TV minimization model is modeled as a convex optimization problem, which is derived from the sufficient and necessary condition of solution uniqueness for the L1 minimization model. An effective alternating direction minimization algorithm is developed to optimize the objective function by alternatively solving two sub-problems split from the convex problem. The Cholesky decomposition method is used in solving the first sub-problem to reduce computational complexity. Experimental results show that the proposed method can efficiently solve the verification problem of the accurate reconstruction sampling condition. Furthermore, we obtain the lower bounds of scanning angle range for the exact reconstruction of a specific phantom with the larger size.