Computationally efficient 4D spectral-spatial EPR imaging.

HighlightsWhat were the main findings?CZT and GAGG crystals exhibited higher CNR values than CsI.HAp’s CNR values were high, as expected.What are the implications of the main findings?CZT and GAGG crystals could provide an excellent alternative to CsI.HAp’s CNR values enable it to be distinguished from other types of microcalcifications.Microcalcifications (HAp, CaCO3, and CaC2O4) in breast tissue may indicate malignancy. Early-stage breast cancer diagnosis may benefit from the clinical application of dual-energy techniques. Dual-energy cone-beam computed tomography (CBCT) could strongly contribute to an accurate diagnosis, especially in dense breasts. This study focused on photon-counting detector alternatives to the standard cesium iodide (CsI) that CBCT currently relies on and investigated potential advantages over the employed CsI scintillators. Denser detector materials with a higher effective atomic number than CsI could improve image quality. A micro-CBCT was simulated in GATE using seven different detector configurations (CsI, bismuth germanate (BGO), lutetium oxyorthosilicate (LSO), lutetium–yttrium oxyorthosilicate (LYSO), gadolinium aluminum gallium garnet (GAGG), lanthanum bromide (LaBr3), and cadmium zinc telluride (CZT)) and four breast tissue phantoms containing microcalcifications of both type I and type II. The dual-energy methodology was applied to planar and tomographic acquisition data. Tomographic data were reconstructed using filtered backprojection (FBP) and the ordered-subsets expectation-maximization (OSEM) algorithm. Image quality was measured using contrast-to-noise ratio (CNR) values. Both monoenergetic and polyenergetic models were considered. CZT and GAGG crystals presented higher CNR values than CsI. HAp microcalcifications exhibited the highest CNR values, which, when accompanied by OSEM, could be distinguished for classification. Detector configurations based on CZT or GAGG crystals could be adequate alternatives to CsI in dual-energy CBCT.

Local Analysis of Iterative Reconstruction from Discrete Generalized Radon Transform Data in the Plane

To compare the image quality of high resolution CT scan of temporal bone reconstructed using novel adaptive statistical iterative reconstruction-v technique with conventional filter back projection reconstruction technique

To compare the impact of different reconstruction algorithms on the image quality of 60 kVp head and neck CT angiography(CTA) using subjective and objective metrics, with a focus on vessel edge sharpness. This prospective study enrolled 45 patients who underwent ultra-low-voltage (60 kVp) head and neck CTA. Image datasets were reconstructed with filtered back-projection (FBP), ClearView (CV) and ClearInfinity (CI) algorithms at low (30%), medium (50%), and high (70%) strengths. Image quality was assessed subjectively and objectively via the Kruskal‒Wallis test for multiple comparisons. Objective parameters, including edge rise slope (ERS) and edge rise distance (ERD), were analyzed via the Friedman test of multiple comparisons statistics. Subjective assessments favored the CI50 reconstruction algorithm, demonstrating superior or satisfactory results compared to the other algorithms, with significantly better vessel delineation, edge definition and diagnostic confidence (all P < 0.05). Objective analysis revealed that the CV50 and CV70 algorithms significantly reduced ERS and/or elevated ERD (both P < 0.05). However, the CI50 algorithm maintained comparable vessel edge sharpness (P > 0.05) across all evaluated head and neck vascular segments when compared with the FBP algorithm. The CI50 reconstruction algorithm optimizes image quality in 60 kVp head and neck CTA. It provides vessel edge sharpness comparable to FBP while offering superior vessel delineation, edge definition, and diagnostic confidence compared to FBP and CV algorithm. These findings suggest that CI50 has the potential to improve diagnostic accuracy in low-dose vascular imaging.

Space constraints in active surface wave surveys often lead to a limited aperture length and close source-to-receiver spacing, hindering mode separation and amplifying near-field effects. These factors distort the phase velocity estimates, reducing the accuracy of dispersion imaging and subsequent shear wave velocity estimation. Traditional wavefield transformation techniques, including ω-c transform, ω-k transform, τ-p transform, HRLRT, and frequency-domain beamforming techniques, assume planar wavefronts and fail to account for near-field effects. On the other hand, the cylindrical frequency-domain beamforming method can only partially eliminate near-field effects. It struggles with mode separation, especially in cases where multiple modes are closely spaced due to shifts in energy peaks and finite-aperture limitations. Therefore, addressing both near-field effects and mode separation is crucial for improving dispersion imaging. To overcome these challenges, we propose a High-Resolution Hankel function-based Radon Transform (HRHRT). The proposed method incorporates cylindrical weights to compensate for radiation damping and employs a tailored forward projection operator to address model incompatibility effects. Within a weighted preconditioned conjugate gradient least-squares framework, it effectively mitigates near-field effects while enhancing mode separation by reducing energy smearing and refining spectral power around modal peaks. The algorithm is rigorously tested on published synthetic and real-field datasets across diverse subsurface conditions. Synthetic results demonstrate the superiority of HRHRT over traditional wavefield transformation techniques by suppressing near-field effects at lower frequencies more effectively. Furthermore, inversion results show that HRHRT consistently yields the lowest mean absolute percentage error in shear wave velocity estimation compared to other methods. Application to real datasets confirms its ability to enhance the resolution of closely spaced modes. Consequently, HRHRT significantly improves the accuracy of predicted Vs profiles from MASW tests, making itself as a valuable advancement in surface wave dispersion imaging.

Introduction: This paper introduces a novel hybrid approach to computed tomography (CT) image reconstruction, designed to enhance medical imaging techniques. The study meticulously compares the performance of this innovative method with established algorithms, including back projection, simultaneous algebraic reconstruction (SAR), and simultaneous algebraic reconstruction iteration (SART) coupled with a total variation minimization algorithm. The evaluation utilizes the NIH-AAPM-Mayo Clinic CT Grand Challenge dataset, ensuring robust and relevant results. Two key performance metrics are taken for comparison: the Structural Similarity Index Measure (SSIM) and the Peak Signal-to-Noise Ratio (PSNR). Objectives: To enhance the quality of filtered back projection (FBP) images in low-dose imaging, Methods: A hybrid model is proposed and compared with SART variants. It combines FBP with a modified CNN featuring three 2D convolution layers (32, 64, and 128 filters of size 3x3) and a ReLU activation function, with an input shape of 150x150 in a batch size of 8. Three 2D max-pooling layers with 2x2 kernel sizes are included. The output is flattened and passed through a dense layer (128 units, ReLU activation), followed by a dropout layer (0.5) to reduce overfitting. The final dense layer uses Softmax for activation. The model is compiled with categorical cross-entropy the loss function and the ADAM as optimizer. Training occurs on a machine with an NVIDIA GEFORCE RTX GPU (6 GB memory). Results: The hybrid algorithm achieved an SSIM value of 0.7916, indicating superior structural fidelity in reconstructed images. Additionally, it demonstrated a PSNR of 19.0424 dB, confirming its effectiveness in producing higher-quality images. Conclusions: The findings have crucial implications for medical imaging, promoting safer practices to reduce radiation exposure by enhancing image quality. This balance between quality and safety highlights the significance of the research, which could revolutionize diagnostic methods in healthcare. Overall, it represents a pivotal advancement in hybrid CT reconstruction, paving the way for further innovations in medical imaging technologies.

Adamkiewicz artery visualisation using iterative model-based reconstruction in low-dose computed tomograpyhy (CT) angiography.

Image Quality Variation with Gantry Rotation Time and Reconstruction Algorithm in Ultra-high-resolution CT.

BackgroundPulsed Electron paramagnetic resonance (EPR) imaging (EPRI) is an advanced oxygen imaging modality for precision radiotherapy, typically acquires high signal-to-noise ratio (SNR) data by averaging the repeatedly collected projections at the corresponding angle to suppress the random noise. This scan mode is the reason for the slow scan speed. The present mitigation is to reduce the repetition times (termed 'shots') for each projection, which leads to noisy projections.ObjectiveAlthough the directional total variation (DTV) algorithm could reconstruct the image from these noisy projections, it may appear staircase artifacts. To solve this problem, we further propose a novel high order DTV (HODTV) algorithm for fast 3D pulsed EPRI.MethodsThe HODTV model has introduced the regularization of high order derivatives, in which the objective term and the high order derivate regularization aim for data fidelity and detail recovery, respectively. Then, we derive its Chambolle-Pock (CP) solving algorithm and verify the correctness. To evaluate the HODTV algorithm, both qualitative and quantitative results are performed with real-world data.ResultsCompared with the filtered back projection (FBP), total variation (TV), and DTV algorithms, the results demonstrate that our method can achieve higher accurate reconstruction. In specific cases, our algorithm only requires 100 shots of scan acquisitions in 6 seconds, whereas the FBP algorithm needs 2000 shots of scan acquisitions taking 120 seconds.ConclusionsThe practical development of clinical imaging workflow, including but not limited to fast 3D pulsed EPRI, may make use of our work.

ABSTRACT Ground roll (GR) is the predominant coherent noise observed in seismic data, often generated by low-velocity Rayleigh waves travelling near the surface. It severely degrades seismic data quality and complicates subsurface imaging. Traditional filtering methods often struggle when GR frequencies overlap with reflection event frequencies, making it challenging to attenuate the noise while preserving essential signals. This study applies a commercially available model-based surface wave attenuation technique to address the challenge of ground roll attenuation. A 1D viscoelastic model derived from well-log data or preliminary seismic analysis is used to generate synthetic seismograms using the matrix propagator method. Dispersion spectra for both the original and synthetic shot gathers are computed using the Linear Radon Transform. A genetic algorithm (GA) workflow is used to optimise the model by minimising the mismatch between the original and synthetic dispersion spectra. The synthetic data generated from the optimised model is then adaptively subtracted from the original data. This process is repeated multiple times to account for lateral variations not captured in the 1D model. The method’s effectiveness is demonstrated on two datasets: 2D noisy data with masked reflections and high-resolution 3D data with hidden coal seam reflections. In both cases, GR is effectively attenuated, revealing previously hidden features and enhancing imaging quality.

Denoising Diffusion Probabilistic Models (DDPM) have emerged as a promising generative framework for sample synthesis, yet their limitations in detail preservation hinder practical applications in computed tomography (CT) image reconstruction. To address these technical constraints and enhance reconstruction quality from compromised CT projection data, this study proposes the Projection Hybrid Inverse Reconstruction Framework (PHIRF) - a novel paradigm integrating conventional reconstruction methodologies with DDPM architecture. The framework implements a dual-phase approach: Initially, conventional CT reconstruction algorithms (e.g., Filtered back projection(FBP), Algebraic Reconstruction Technique(ART), Maximum-Likelihood Expectation Maximization (ML-EM)) are employed to generate preliminary reconstructions from incomplete projections, establishing low-dimensional feature representations. These features are subsequently parameterized and embedded as conditional constraints in the reverse diffusion process of DDPM, thereby guiding the generative model to synthesize enhanced tomographic images with improved structural fidelity. Comprehensive evaluations were conducted on three representative ill-posed projection scenarios: limited-angle projections, sparse-view acquisitions, and low-dose measurements. Experimental results demonstrate that PHIRF achieves state-of-the-art performance across all compromised data conditions, particularly in preserving fine anatomical details and suppressing reconstruction artifacts. Quantitative metrics and visual assessments confirm the framework's consistent superiority over existing deep learning-based reconstruction approaches, substantiating its adaptability to diverse projection degradation patterns. This hybrid architecture establishes a new paradigm for combining physical prior knowledge with data-driven generative models in medical image reconstruction tasks.

Abstract Motivated by recent work on the Sachdev-Ye-Kitaev (SYK) model, we consider the effect of Radon or X-ray transformations, on the Laplace eigenfunctions in hyperbolic Bolyai-Lobachevsky space. We show that the Radon map from this space to Lorentzian-signature de Sitter space is easier to interpret if we use the Poincaré disc model and eigenfunctions rather than the upper-half-plane model. In particular, this version of the transform reveals the geometric origin of the boundary conditions imposed on the eigenfunctions that are involved in calculating the SYK four-point function.

In 99mTc myocardial perfusion SPECT, extra-cardiac accumulation from organs such as the liver or gastrointestinal tract may overlap with the inferior wall, causing artifacts that interfere with image interpretation. This study aimed to quantitatively evaluate the effectiveness of a novel image reconstruction method, the masking process on unsmoothed images (MUS method; CardioMUSk®, PDRadiopharma Inc., Tokyo, Japan), in reducing the influence of extra-cardiac accumulation using both phantom and clinical images. This retrospective study included 200 patients (400 scans) who underwent a one-day stress-rest protocol using 99mTc-sestamibi (MIBI) with pharmacologic stress administered first. Image reconstruction was performed using filtered back projection (FBP) and ordered subset expectation maximization with resolution recovery (OS-EM-RR), both with and without the MUS method. First, visual classification of extra-cardiac accumulation patterns relative to the inferior wall was performed, and the separation capability of each reconstruction method was assessed. Next, phantom experiments were conducted to investigate the effects of extra-cardiac accumulation volume, proximity, and concentration on contrast in the inferior wall. Furthermore, quantitative comparison of relative contrast between the inferior wall and the lateral and septal walls was performed using clinical data. The MUS method reduced the proportion of visually unseparated cases from 15.5% to 3.5% compared with the conventional method. In phantom studies, larger extra-cardiac accumulation and closer proximity to the myocardium resulted in greater degradation of inferior wall contrast. When a distance of 2cm was maintained between extra-cardiac accumulation and the myocardium, the effect was substantially reduced. In clinical images, the MUS method significantly improved relative contrast in the inferolateral/inferior wall at the mid-ventricular level (Wilcoxonp = 0.030) and in the inferoseptal/inferiorwall at the basal level (Wilcoxonp < 0.001), while no significant improvement was observed in the basal inferolateral/inferior wall region (Wilcoxonp = 0.605). The MUS method demonstrated enhanced separation of extra-cardiac accumulation and improved contrast in the inferior myocardial wall compared with conventional methods. It was particularly effective in cases where extra-cardiac accumulation overlapped or closely contacted the myocardium, indicating its potential clinical utility in 99mTc myocardial perfusion SPECT.

Diabetic retinopathy is a serious ocular complication that poses a significant threat to patients' vision and overall health. Early detection and accurate grading are essential to prevent vision loss. Current automatic grading methods rely heavily on deep learning applied to retinal fundus images, but the complex, irregular patterns of lesions in these images, which vary in shape and distribution, make it difficult to capture the subtle changes. This study introduces RadFuse, a multi-representation deep learning framework that integrates non-linear RadEx-transformed sinogram images with traditional fundus images to enhance diabetic retinopathy detection and grading. Our RadEx transformation, an optimized non-linear extension of the Radon transform, generates sinogram representations to capture complex retinal lesion patterns. By leveraging both spatial and transformed domain information, RadFuse enriches the feature set available to deep learning models, improving the differentiation of severity levels. We conducted extensive experiments on two benchmark datasets, APTOS-2019 and DDR, using three convolutional neural networks (CNNs): ResNeXt-50, MobileNetV2, and VGG19. RadFuse showed significant improvements over fundus-image-only models across all three CNN architectures and outperformed state-of-the-art methods on both datasets. For severity grading across five stages, RadFuse achieved a quadratic weighted kappa of 93.24%, an accuracy of 87.07%, and an F1-score of 87.17%. In binary classification between healthy and diabetic retinopathy cases, the method reached an accuracy of 99.09%, precision of 98.58%, and recall of 99.64%, surpassing previously established models. These results demonstrate RadFuse's capacity to capture complex non-linear features, advancing diabetic retinopathy classification and promoting the integration of advanced mathematical transforms in medical image analysis. The source code will be available at https://github.com/Farida-Ali/RadEx-Transform/tree/main .

Standard computed tomography (CT) reconstruction algorithms such as filtered back projection (FBP) and Feldkamp-Davis-Kress (FDK) require many views for producing high-quality reconstructions, which can slow image acquisition and increase cost in non-destructive evaluation (NDE) applications. Over the past 20 years, a variety of methods have been developed for computing high-quality CT reconstructions from sparse views. However, the problem of how to select the best views for CT reconstruction remains open. In this paper, we present a novel view covariance loss (VCL) function that measures the joint information of a set of views by approximating the normalized mean squared error (NMSE) of the reconstruction. We present fast algorithms for computing the VCL along with an algorithm for selecting a subset of views that approximately minimizes its value. Our experiments on simulated and measured data indicate that for a fixed number of views our proposed view covariance loss selection (VCLS) algorithm results in reconstructions with lower NRMSE, fewer artifacts, and greater accuracy than current alternative approaches.

Photoacoustic computed tomography (PACT) synergizes optical absorption contrast with ultrasonic resolution for noninvasive biomedical imaging yet faces limitations in signal-to-noise ratio (SNR), resolution, and contrast. This study introduces a coherent-excitation PACT system integrating interferometric optical excitation and Frequency-Compensated Filtered Back Projection (FC-FBP) reconstruction. The proposed method utilizes phase-locked dual-pulse interferometric excitation to amplify photoacoustic emissions; for the isolated chicken heart, the resolution is improved by 7.9% compared to the single-pulse protocol. The FC-FBP algorithm compensates for frequency-dependent acoustic attenuation via depth-adaptive Gaussian filtering, enhancing the projected signal in the target area while suppressing speckle artifacts. Through experimental validation, we confirm that the coherent-excitation scheme enables simultaneous optimization of optical fluence distribution and acoustic coherence; hence, it can be used to resolve previously indistinguishable hemoglobin oxygenation gradients in murine tumor models. This advancement establishes a high-sensitivity PACT framework, showing potential for real-time intraoperative imaging and dynamic metabolic monitoring in clinical applications.

Improving image quality and diagnostic performance using deep learning image reconstruction in 100-kVp CT enterography for patients with wide-range body mass index.

This paper explores industrial computed tomography (CT) for detection of anomalies such as voids and gaps in power connectors for automotive application. Such defects can lead to leakage paths which can cause electrical malfunction or short circuits. A primary challenge in CT analysis arises from the proximity of dense metal parts next to plastic material, which induces strong imaging artifacts that hinder the evaluation of such defects. This study evaluates three imaging methods - standard Filtered Back Projection reconstruction (FBP), Quantum Reconstruction Technique (QRT), and dual-energy combination (DE) - to mitigate these artifacts and improve defect visualization. Each method's efficacy is assessed by analyzing the accuracy in detecting and characterizing gaps and voids. Results are compared to a target preparation and cross-section analysis with optical microscopy, providing a benchmark for reliability and accuracy. Full experimental findings and comparisons are detailed in the complete paper.

On the Application of Radon Transformation for the Synthesis of Thinned Planar Antenna Arrays