HEAL: High-Frequency Enhanced and Attention-Guided Learning Network for Sparse-View CT Reconstruction.

X-ray computed tomography (CT) imaging technology has become an indispensable diagnostic tool in clinical examination. However, it poses a risk of ionizing radiation, making the reduction of radiation dose one of the current research hotspots in CT imaging. Sparse-view imaging, as one of the main methods for reducing radiation dose, has made significant progress in recent years. In particular, sparse-view reconstruction methods based on deep learning have shown promising results. Nevertheless, efficiently recovering image details under ultra-sparse conditions remains a challenge. To address this challenge, this paper proposes a high-frequency enhanced and attention-guided learning Network (HEAL). HEAL includes three optimization strategies to achieve detail enhancement: Firstly, we introduce a dual-domain progressive enhancement module, which leverages fidelity constraints within each domain and consistency constraints across domains to effectively narrow the solution space. Secondly, we incorporate both channel and spatial attention mechanisms to improve the network's feature-scaling process. Finally, we propose a high-frequency component enhancement regularization term that integrates residual learning with direction-weighted total variation, utilizing directional cues to effectively distinguish between noise and textures. The HEAL network is trained, validated and tested under different ultra-sparse configurations of 60 views and 30 views, demonstrating its advantages in reconstruction accuracy and detail enhancement.

Correction of scatter in CBCT with a new grid design.

OBJECTIVE: This study aims at developing a simple and rapid Compton scatter correction approach for cone-beam CT (CBCT) imaging. Approach: In this work, a new Compton scatter estimation model is established based on two distinct CBCT scans: one measures the full primary and scatter signals without anti-scatter grid (ASG), and the other measures a portion of primary and scatter signals with ASG. To accelerate the entire data acquisition speed, a half anti-scatter grid (h-ASG) that covers half of the full detector surface is proposed. As a result, the distribution of scattered X-ray photons could be estimated from a single CBCT scan. Physical phantom experiments are conducted to validate the performance of the newly proposed scatter correction approach. Main Results: Results demonstrate that the proposed half grid approach can quickly and precisely estimate the distribution of scattered X-ray photons from only one single CBCT scan, resulting in a significant reduction of shading artifacts. In addition, it is found that the h-ASG approach is less sensitive to the grid transmission fractions, grid ratio and object size, indicating a robust performance of the new method. Significance: In the future, the Compton scatter artifacts can be quickly corrected using a half grid in CBCT imaging.

Dynamic X-ray imaging with screen-printed perovskite CMOS array.

High performance X-ray detector with ultra-high spatial and temporal resolution are crucial for biomedical imaging. This study reports a dynamic direct-conversion CMOS X-ray detector assembled with screen-printed CsPbBr3, whose mobility-lifetime product is 5.2 × 10-4 cm2 V-1 and X-ray sensitivity is 1.6 × 104 µC Gyair-1 cm-2. Samples larger than 5 cm[Formula: see text]10 cm can be rapidly imaged by scanning this detector at a speed of 300 frames per second along the vertical and horizontal directions. In comparison to traditional indirect-conversion CMOS X-ray detector, this perovskite CMOS detector offers high spatial resolution (5.0 lp mm-1) X-ray radiographic imaging capability at low radiation dose (260 nGy). Moreover, 3D tomographic images of a biological specimen are also successfully reconstructed. These results highlight the perovskite CMOS detector's potential in high-resolution, large-area, low-dose dynamic biomedical X-ray and CT imaging, as well as in non-destructive X-ray testing and security scanning.

Transferring U-Net between low-dose CT denoising tasks: a validation study with varied spatial resolutions.

Background: Recently, deep learning techniques have been widely used in low-dose computed tomography (LDCT) imaging applications for quickly generating high quality computed tomography (CT) images at lower radiation dose levels. The purpose of this study is to validate the reproducibility of the denoising performance of a given network that has been trained in advance across varied LDCT image datasets that are acquired from different imaging systems with different spatial resolutions. Methods: Specifically, LDCT images with comparable noise levels but having different spatial resolutions were prepared to train the U-Net. The number of CT images used for the network training, validation and test was 2,400, 300 and 300, respectively. Afterwards, self- and cross-validations among six selected spatial resolutions (62.5, 125, 250, 375, 500, 625 µm) were studied and compared side by side. The residual variance, peak signal to noise ratio (PSNR), normalized root mean square error (NRMSE) and structural similarity (SSIM) were measured and compared. In addition, network retraining on a small number of image set was performed to fine tune the performance of transfer learning among LDCT tasks with varied spatial resolutions. Results: Results demonstrated that the U-Net trained upon LDCT images having a certain spatial resolution can effectively reduce the noise of the other LDCT images having different spatial resolutions. Regardless, results showed that image artifacts would be generated during the above cross validations. For instance, noticeable residual artifacts were presented at the margin and central areas of the object as the resolution inconsistency increased. The retraining results showed that the artifacts caused by the resolution mismatch can be greatly reduced by utilizing about only 20% of the original training data size. This quantitative improvement led to a reduction in the NRMSE from 0.1898 to 0.1263 and an increase in the SSIM from 0.7558 to 0.8036. Conclusions: In conclusion, artifacts would be generated when transferring the U-Net to a LDCT denoising task with different spatial resolution. To maintain the denoising performance, it is recommended to retrain the U-Net with a small amount of datasets having the same target spatial resolution.

Performance evaluation of quantitative material decomposition in slow kVp switching dual-energy CT.

BACKGROUND: Slow kVp switching technique is an important approach to realize dual-energy CT (DECT) imaging, but its performance has not been thoroughly investigated yet. OBJECTIVE: This study aims at comparing and evaluating the DECT imaging performance of different slow kVp switching protocols, and thus helps determining the optimal system settings. METHODS: To investigate the impact of energy separation, two different beam filtration schemes are compared: the stationary beam filtration and dynamic beam filtration. Moreover, uniform tube voltage modulation and weighted tube voltage modulation are compared along with various modulation frequencies. A model-based direct decomposition algorithm is employed to generate the water and iodine material bases. Both numerical and physical experiments are conducted to verify the slow kVp switching DECT imaging performance. RESULTS: Numerical and experimental results demonstrate that the material decomposition is less sensitive to beam filtration, voltage modulation type and modulation frequency. As a result, robust material-specific quantitative decomposition can be achieved in slow kVp switching DECT imaging. CONCLUSIONS: Quantitative DECT imaging can be implemented with slow kVp switching under a variety of system settings.

Super Resolution Dual-Energy Cone-Beam CT Imaging With Dual-Layer Flat-Panel Detector.

In flat-panel detector (FPD) based cone-beam computed tomography (CBCT) imaging, the native receptor array is usually binned into a smaller matrix size. By doing so, the signal readout speed could be increased by 4-9 times at the expense of a spatial resolution loss of 50%-67%. Clearly, such manipulation poses a key bottleneck in generating high spatial and high temporal resolution CBCT images at the same time. In addition, the conventional FPD is also difficult in generating dual-energy CBCT images. In this paper, we propose an innovative super resolution dual-energy CBCT imaging method, named as suRi, based on dual-layer FPD (DL-FPD) to overcome these aforementioned difficulties at once. With suRi, specifically, a 1D or 2D sub-pixel (half pixel in this study) shifted binning is applied instead of the conventionally aligned binning to double the spatial sampling rate during the dual-energy data acquisition. As a result, the suRi approach provides a new strategy to enable high spatial resolution CBCT imaging while at high readout speed. Moreover, a penalized likelihood material decomposition algorithm is developed to directly reconstruct the high resolution bases from these dual-energy CBCT projections containing sub-pixel shifts. Numerical and physical experiments are performed to validate this newly developed suRi method with phantoms and biological specimen. Results demonstrate that suRi can significantly improve the spatial resolution of the CBCT image. We believe this developed suRi method would greatly enhance the imaging performance of the DL-FPD based dual-energy CBCT systems in future.

Super-resolution dual-layer CBCT imaging with model-guided deep learning.

Objective.This study aims at investigating a novel super resolution CBCT imaging approach with a dual-layer flat panel detector (DL-FPD).Approach.With DL-FPD, the low-energy and high-energy projections acquired from the top and bottom detector layers contain over-sampled spatial information, from which super-resolution CT images can be reconstructed. A simple mathematical model is proposed to explain the signal formation procedure in DL-FPD, and a dedicated recurrent neural network, named suRi-Net, is developed based upon the above imaging model to nonlinearly retrieve the high-resolution dual-energy information. Physical benchtop experiments are conducted to validate the performance of this newly developed super-resolution CBCT imaging method.Main Results.The results demonstrate that the proposed suRi-Net can accurately retrieve high spatial resolution information from the low-energy and high-energy projections of low spatial resolution. Quantitatively, the spatial resolution of the reconstructed CBCT images from the top and bottom detector layers is increased by about 45% and 54%, respectively.Significance.In the future, suRi-Net will provide a new approach to perform high spatial resolution dual-energy imaging in DL-FPD-based CBCT systems.

A pixelated liquid perovskite array for high-sensitivity and high-resolution X-ray imaging scintillation screens.

Scintillators with high spatial resolution at a low radiation dose rate are desirable for X-ray medical imaging. A low radiation dose rate can be achieved using a sufficiently thick scintillator layer to absorb the incident X-ray energy completely, however, often at the expense of low spatial resolution due to the issue of optical crosstalk of scintillation light. Therefore, to achieve high sensitivity combined with high-resolution imaging, a thick scintillator with perfect light guiding properties is in high demand. Herein, a new strategy is developed to address this issue by embedding liquid scintillators into lead-containing fiber-optical plates (FOPs, n = 1.5) via the siphon effect. The liquid scintillator is composed of perovskite quantum dots (QDs)/2,5-diphenyloxazole (PPO) and the non-polar high-refractive index (n = 1.66) solvent α-bremnaphthalene. Benefiting from the pixelated and thickness-adjustable scintillators, the proposed CsPbBr3 QDs/PPO liquid scintillator-based X-ray detector achieves a detection limit of 79.1 µGy s-1 and a spatial resolution of 4.6 lp mm-1. In addition, it displays excellent tolerance against radiation (>34 h) and shows outstanding stability under ambient conditions (>160 h). This strategy could also be applied to other liquid scintillators (such as CsPbCl3 QDs and Mn:CsPbCl3 QDs). The combination of high sensitivity, high spatial resolution and stability, easy fabrication and maintenance, and a reusable substrate matrix makes these liquid scintillators a promising candidate for practical X-ray medical imaging applications.

CsPbBr3-DMSO merged perovskite micro-bricks for efficient X-ray detection.

Inorganic perovskite wafers with good stability and adjustable sizes are promising in X-ray detection but the high synthetic temperature is a hindrance. Herein, dimethyl sulfoxide (DMSO) is used to prepare the CsPbBr3 micro-bricks powder at room temperature. The CsPbBr3 powder has a cubic shape with few crystal defects, small charge trap density, and high crystallinity. A trace amount of DMSO attaches to the surface of the CsPbBr3 micro-bricks via Pb-O bonding, forming the CsPbBr3-DMSO adduct. During hot isostatic processing, the released DMSO vapor merges the CsPbBr3 micro-bricks, producing a compact and dense CsPbBr3 wafer with minimized grain boundaries and excellent charge transport properties. The CsPbBr3 wafer shows a large mobility-lifetime (µτ) product of 5.16 × 10-4 cm2·V-1, high sensitivity of 14,430 µC·Gyair-1·cm-2, low detection limit of 564 nGyair·s-1, as well as robust stability in X-ray detection. The results reveal a novel strategy with immense practical potential pertaining to high-contrast X-ray detection. Electronic Supplementary Material: Supplementary material (further details of the characterization, SEM images, AFM images, KPFM images, schematic illustration, XRD patterns, XPS spectra, FTIR spectra, UPS spectra, and stability tests) is available in the online version of this article at 10.1007/s12274-023-5487-3.

Attention-based dual-branch deep network for sparse-view computed tomography image reconstruction.

Background: The widespread application of X-ray computed tomography (CT) imaging in medical screening makes radiation safety a major concern for public health. Sparse-view CT is a promising solution to reduce the radiation dose. However, the reconstructed CT images obtained using sparse-view CT may suffer severe streaking artifacts and structural information loss. Methods: In this study, a novel attention-based dual-branch network (ADB-Net) is proposed to solve the ill-posed problem of sparse-view CT image reconstruction. In this network, downsampled sinogram input is processed through 2 parallel branches (CT branch and signogram branch) of the ADB-Net to independently extract the distinct, high-level feature maps. These feature maps are fused in a specified attention module from 3 perspectives (channel, plane, and spatial) to allow complementary optimizations that can mitigate the streaking artifacts and the structure loss in sparse-view CT imaging. Results: Numerical simulations, an anthropomorphic thorax phantom, and in vivo preclinical experiments were conducted to verify the sparse-view CT imaging performance of the ADB-Net. The proposed network achieved a root-mean-square error (RMSE) of 20.6160, a structural similarity (SSIM) of 0.9257, and a peak signal-to-noise ratio (PSNR) of 38.8246 on numerical data. The visualization results demonstrate that this newly developed network can consistently remove the streaking artifacts while maintaining the fine structures. Conclusions: The proposed attention-based dual-branch deep network, ADB-Net, provides a promising alternative to reconstruct high-quality sparse-view CT images for low-dose CT imaging.

Automatic cerebral computed tomography venographic imaging based on the prior knowledge of cerebral blood circulation.

BACKGROUND AND PURPOSE: Current clinical computed tomography venographic (cCTV) images present limited cerebral venous profiles. Therefore, this study aimed to develop an automatic cerebral CTV imaging technique using computed tomographic perfusion (CTP) images in a cohort of patients with stroke. MATERIALS AND METHODS: We retrospectively evaluated 10 (intracerebral hemorrhage) and 2 (acute ischemic stroke) patients who underwent institutional CTP imaging. CTV images were processed with the proposed CTV (pCTV) technique, and pCTV and cCTV images were then independently evaluated by two experienced neuroradiologists blinded to all clinical information using a novel scoring method that considered overall image quality, venous visibility, and arterial mis-segmentation. Venous visibility was separately evaluated for the dural sinus, superficial vein, and deep vein. Then, statistical analysis was performed to determine whether the pCTV technique was superior to the cCTV technique. RESULTS: In total, 14 sets of pCTV images were generated and compared with cCTV images. The overall image quality and venous visibility scores of pCTV images were significantly higher than those of cCTV images (all values of p<0.05), especially for the dural sinus (median [25th, 75th percentiles], 14.00 [13.63, 15.50] vs. 7.50 [7.00, 10.88]), and superficial vein (9.00 [8.88, 10.00] vs. 3.25 [1.63, 8.25]), while the difference in arterial mis-segmentation was not statistically significant (p= 0.164). CONCLUSIONS: This study proposed an automatic cerebral CTV imaging technique to eliminate residual bone and soft tissues, minimize the impact of the cerebral arterial system, and present a relatively comprehensive cerebral venous system, which would help physicians assess cerebral venous outflow profiles after stroke and seek imaging markers associated with clinical outcomes.

Energy resolving dark-field imaging with dual phase grating interferometer.

X-ray dark-filed imaging is a powerful approach to quantify the dimension of micro-structures of the object. Often, a series of dark-filed signals have to be measured under various correlation lengths. For instance, this is often achieved by adjusting the sample positions by multiple times in Talbot-Lau interferometer. Moreover, such multiple measurements can also be collected via adjustments of the inter-space between the phase gratings in dual phase grating interferometer. In this study, the energy resolving capability of the dual phase grating interferometer is explored with the aim to accelerate the data acquisition speed of dark-filed imaging. To do so, both theoretical analyses and numerical simulations are investigated. Specifically, the responses of the dual phase grating interferometer at varied X-ray beam energies are studied. Compared with the mechanical position translation approach, the combination of such energy resolving capability helps to greatly shorten the total dark-field imaging time in dual phase grating interferometer.

PbI2 -DMSO Assisted In Situ Growth of Perovskite Wafers for Sensitive Direct X-Ray Detection.

Although perovskite wafers with a scalable size and thickness are suitable for direct X-ray detection, polycrystalline perovskite wafers have drawbacks such as the high defect density, defective grain boundaries, and low crystallinity. Herein, PbI2 -DMSO powders are introduced into the MAPbI3 wafer to facilitate crystal growth. The PbI2 powders absorb a certain amount of DMSO to form the PbI2 -DMSO powders and PbI2 -DMSO is converted back into PbI2 under heating while releasing DMSO vapor. During isostatic pressing of the MAPbI3 wafer with the PbI2 -DMSO solid additive, the released DMSO vapor facilitates in situ growth in the MAPbI3 wafer with enhanced crystallinity and reduced defect density. A dense and compact MAPbI3 wafer with a high mobility-lifetime (µτ) product of 8.70 × 10-4 cm2 V-1 is produced. The MAPbI3 -based direct X-ray detector fabricated for demonstration shows a high sensitivity of 1.58 × 104 µC Gyair-1  cm-2 and a low detection limit of 410 nGyair s-1 .

Feasibility study of three-material decomposition in dual-energy cone-beam CT imaging with deep learning.

Objective.In this work, a dedicated end-to-end deep convolutional neural network, named as Triple-CBCT, is proposed to demonstrate the feasibility of reconstructing three different material distribution volumes from the dual-energy CBCT projection data.Approach.In Triple-CBCT, the features of the sinogram and the CT image are independently extracted and cascaded via a customized domain transform network module. This Triple-CBCT network was trained by numerically synthesized dual-energy CBCT data, and was tested with experimental dual-energy CBCT data of the Iodine-CaCl2solution and pig leg specimen scanned on an in-house benchtop system.Main results.Results show that the information stored in both the sinogram and CT image domains can be used together to improve the decomposition quality of multiple materials (water, iodine, CaCl2or bone) from the dual-energy projections. In addition, both the numerical and experimental results demonstrate that the Triple-CBCT is able to generate high-fidelity dual-energy CBCT basis images.Significance.An innovative end-to-end network that joints the sinogram and CT image domain information is developed to facilitate high quality automatic decomposition from the dual-energy CBCT scans.

Generalized deep iterative reconstruction for sparse-view CT imaging.

Sparse-view CT is a promising approach for reducing the x-ray radiation dose in clinical CT imaging. However, the CT images reconstructed from the conventional filtered backprojection algorithm suffer from severe streaking artifacts. Iterative reconstruction algorithms have been widely adopted to mitigate these streaking artifacts, but they may prolong the CT imaging time due to the intense data-specific computations. Recently, a model-driven deep learning CT image reconstruction method, which unrolls the iterative optimization procedures into a deep neural network, has shown exciting prospects for improving image quality and shortening the reconstruction time. In this work, we explore a generalized unrolling scheme for such an iterative model to further enhance its performance on sparse-view CT imaging. By using it, the iteration parameters, regularizer term, data-fidelity term and even the mathematical operations are all assumed to be learned and optimized via network training. Results from the numerical and experimental sparse-view CT imaging demonstrate that the newly proposed network with the maximum generalization provides the best reconstruction performance.

DIRECT-Net: A unified mutual-domain material decomposition network for quantitative dual-energy CT imaging.

PURPOSE: The purpose of this paper is to present an end-to-end deep convolutional neural network to improve the dual-energy CT (DECT) material decomposition performance. METHODS: In this study, we proposes a unified mutual-domain (sinogram domain and CT domain) material decomposition network (DIRECT-Net) for DECT imaging. By design, the DIRECT-Net has immediate access to mutual-domain data, and utilizes stacked convolution neural network layers for noise reduction and material decomposition. The training data are numerically generated following the fundamental DECT imaging physics. Numerical simulation of the XCAT digital phantom, experiments of a biological specimen, a calcium chloride phantom and an iodine solution phantom are carried out to evaluate the performance of DIRECT-Net. Comparisons are performed with different DECT decomposition algorithms. RESULTS: Results demonstrate that the proposed DIRECT-Net can generate water and bone basis images with less artifacts compared to the other decomposition methods. Additionally, the quantification errors of the calcium chloride (75-375 mg/cm3 ) and the iodine (2-20 mg/cm3 ) are less than 4%. CONCLUSIONS: An end-to-end material decomposition network is proposed for quantitative DECT imaging. The qualitative and quantitative results demonstrate that this new DIRECT-Net has promising benefits in improving the DECT image quality.

Performance evaluation of dual-energy CT and differential phase contrast CT in quantitative imaging applications.

PURPOSE: The purpose of this study is to evaluate and compare the quantitative material decomposition performance of the dual-energy CT (DECT) and differential phase contrast CT (DPCT) via numerical observer studies. METHODS: The electron density ( ρ e $\rho _{{\rm e}}$ ) and the effective atomic number ( Z eff $Z_{{\rm eff}}$ ) are selected as the decomposition bases. The image domain based decomposition algorithms with certain noise suppression are used to extract the ρ e $\rho _{{\rm e}}$ and Z eff $\text{Z}_{{\rm eff}}$ information under three different spatial resolutions (0.3 mm, 0.1 mm, and 0.03 mm). The contrast-to-noise-ratio (CNR) and the numerical human observer model which is sensitive to the noise textures are investigated to compare the quantitative imaging performance of DECT and DPCT under varied radiation dose levels. RESULTS: The model observer results show that the DECT is superior to DPCT at 0.3 mm spatial resolution (300 mm object size); the DECT and DPCT show similar quantitative imaging performance at 0.1 mm spatial resolution (100 mm object size); and the DPCT outperforms the DECT by approximately 1.5 times for the 0.3 mm sized imaging target at 0.03 mm spatial resolution (30 mm object size). CONCLUSIONS: In conclusion, the DECT would be recommended to obtain ρ e $\rho _{{\rm e}}$ and Z eff $Z_{{\rm eff}}$ for the low spatial resolution quantitative imaging applications such as the diagnostic CT imaging. Whereas, the DPCT would be recommended for ultra high spatial resolution imaging tasks of small objects such as the micro-CT imaging. This study provides a reference to determine the most appropriate quantitative X-ray CT imaging method for a certain radiation dose level.

Sequential Growth of 2D/3D Double-Layer Perovskite Films with Superior X-Ray Detection Performance.

Perovskite materials in different dimensions show great potential in direct X-ray detection, but each with limitations stemming from its own intrinsic properties. Particularly, the sensitivity of two-dimensional (2D) perovskites is limited by poor carrier transport while ion migration in three-dimensional (3D) perovskites causes the baseline drifting problem. To circumvent these limitations, herein a double-layer perovskite film is developed with properly aligned energy level, where 2D (PEA)2 MA3 Pb4 I13 (PEA=2-phenylethylammonium, MA=methylammonium) is cascaded with vertically crystallized 3D MAPbI3 . In this new design paradigm, the 3D layer ensures fast carrier transport while the 2D layer mitigates ion migration, thus offering a high sensitivity and a greatly stabilized baseline. Besides, the 2D layer increases the film resistivity and enlarges the energy barrier for hole injection without compromising carrier extraction. Consequently, the double-layer perovskite detector delivers a high sensitivity (1.95 × 104 µC Gyair -1 cm-2 ) and a low detection limit (480 nGyair s-1 ). Also demonstrated is the X-ray imaging capacity using a circuit board as the object. This work opens up a new avenue for enhancing X-ray detection performance via cascade assembly of various perovskites with complementary properties.

Ultrathin and Ultrasensitive Direct X-ray Detector Based on Heterojunction Phototransistors.

Most contemporary X-ray detectors adopt device structures with non/low-gain energy conversion, such that a fairly thick X-ray photoconductor or scintillator is required to generate sufficient X-ray-induced charges, and thus numerous merits for thin devices, such as mechanical flexibility and high spatial resolution, have to be compromised. This dilemma is overcome by adopting a new high-gain device concept of a heterojunction X-ray phototransistor. In contrast to conventional detectors, X-ray phototransistors allow both electrical gating and photodoping for effective carrier-density modulation, leading to high photoconductive gain and low noise. As a result, ultrahigh sensitivities of over 105  µC Gyair -1  cm-2 with low detection limit are achieved by just using an ≈50 nm thin photoconductor. The employment of ultrathin photoconductors also endows the detectors with superior flexibility and high imaging resolution. This concept offers great promise in realizing well-balanced detection performance, mechanical flexibility, integration, and cost for next-generation X-ray detectors.

Angular sensitivity of an x-ray differential phase contrast imaging system with real and virtual source images.

In this work, a novel, to the best of our knowledge, approach based on an x-ray thin lens imaging theory is proposed to predict the angular sensitivity responses of dual-phase-grating differential phase contrast (DPC) interferometers. Experimental validations have been performed to demonstrate the high accuracy of theoretical predictions using two different setups: one with real source images and the other with virtual source images. This new sensitivity calculation method is helpful to optimize the DPC imaging performance of a dual-phase-grating system.