The first line of non-destructive inspection of cargo often relies on single or double-view X-ray radiography, which is fast but lacks depth resolution and is prone to object occlusion. In contrast, conventional X-ray computed tomography (CT) allows 3D imaging but typically relies on mechanically rotating gantries, which limits throughput and increases system complexity. Recently, multi-source fixed-gantry X-ray systems have been proposed as a promising acquisition geometry to combine high imaging speed with volumetric imaging, while reducing mechanical complexity. The precision of the reconstructed images stemming from these systems as a function of the acquisition setup has however hardly been explored. This paper proposes a flexible framework for optimal experiment design of a rectangular multi-source X-ray cargo scanning system. The proposed framework allows the experimenter to calculate the highest attainable imaging precision, as quantified by the Cram & eacute;r-Rao lower bound (CRLB), as a function of the X-ray system's geometric settings, which facilitates optimal experiment design. To illustrate this potential, several system configurations with differently positioned and oriented sources are evaluated and compared in terms of the CRLB-based A-optimality criterion.

Porosity formation in polymer selective laser sintering (SLS) significantly impacts part quality, as unwanted pores compromise mechanical properties. In-process control of porosity is a promising solution, but existing in-situ monitoring systems lack the necessary speed to meaningfully intervene in small localized areas, and the specificity to guide targeted process adjustments. This study addresses these limitations by introducing an in-situ monitoring system that predicts control-actionable porosity targets at high speeds (2 kHz). The system combines multimodal visible and short-wave infrared imaging, complementing information about the powder bed condition and thermal behavior. A neural network then makes predictions based on learned correlations between the image data and the porosity targets. The neural network incorporates both contextual and temporal information through a specialized architecture that leverages adaptive weight networks and a temporal sliding window. The proposed targets are extracted post-print using x-ray computed tomography analysis and serve as ground truth targets to train the neural network. Based on pore concentrations and spatial uniformity measurements, the targets have a clear physical justification, and allow any printed part to be used as training data. Crucially, by distinguishing between insufficient melting and overheating, these targets resolve the directionality of the required laser power adjustment, enabling non-ambiguous control actions. Experimental results on cylindrical test parts demonstrate good correlations between the porosity targets and expected pore formations. Furthermore, the proposed in-situ system shows strong predictive performance, with normalized mean absolute errors of 14% for predicting relative pore concentrations, and a correlation coefficient of 0.68 for predicting the underlying cause. The combination of a high-speed system and the prediction of control-actionable porosity targets pave the way for effective process control in SLS, improving part quality.

Process-induced defects in the Selective Laser Sintering (SLS) of polymers continue to be common, leading to undesirable mechanical properties, discontinuation of print jobs, and generally complicating part reproducibility. Among these defects, part deformations, such as warping and local geometrical inaccuracies, remain one of the most critical. Non-homogeneous build chamber temperatures, geometry-induced thermal bottlenecks, and the overall temperature sensitivity of the SLS process make deformation defects notoriously difficult to avoid. Lately, in-situ monitoring systems have been gaining attention for their ability to detect process-induced defects by actively monitoring the process. However, existing systems often lack sufficient information and processing speed for meaningful intervention in the printing process. In this work, we advance the state-of-the-art of inprocess SLS deformation detection by employing a high-speed multifaceted data-fusion approach. The proposed system combines a high-speed dynamic region-of-interest camera (2 kHz) in the visual spectrum, which observes powder bed related inconsistencies, with a short-wave infrared camera, to observe thermal-related effects. Local geometry effects are accounted for by continuously fusing the camera observations with geometrical features precomputed from the 3D design file. Furthermore, a temporal sliding window approach ensures defect predictions are not only determined from single observations, but also from longer-evolving process conditions. A custom GPU-accelerated data-driven algorithm further ensures high-speed and accurate inference. The system was validated on two Polyamide 12 printed objects deliberately designed to produce deformation defects. Ground truth deformation targets extracted using XCT served both as training targets for the data-driven algorithm and as validation targets for a separate testing dataset. The experiments demonstrated normalized mean absolute error values of just 12% for detecting deformation magnitudes. The results show promise for the system to be leveraged for high-speed intra-layer closed-loop control of deformation defects, ultimately improving part quality and reproducibility in the SLS process.

This study presents a method for estimating the mass fractions of chemical elements in an object, which is modeled as a homogeneous mixture, using a single X-ray radiograph and its surface mesh. The method assumes that the chemical elements in the mixture and their mass attenuation coefficients are known. A stochastic gradient descent algorithm is used to iteratively minimize the error between a scanned radiograph and a simulated polychromatic radiograph, enabling the estimation of mass fractions. The CAD-ASTRA toolbox is employed to compute the path-lengths and simulate polychromatic X-ray radiographs for test objects under varying noise conditions [1].

Edge Illumination X-ray Phase Contrast Imaging (EI-XPCI) is a powerful, nondestructive imaging technique known for its enhanced sensitivity compared to conventional X-ray imaging techniques. However, EI-XPCI is prone to increased noise due to the presence of gratings that absorb X-rays, resulting in lower flux compared to conventional imaging. Noise propagates from the X-ray projections and affects the estimation of the three contrasts: phase contrast, dark field, and attenuation. Therefore, effective noise-suppression techniques are required. Despite advances in conventional X-ray image denoising methods, strategies specifically tailored for EI-XPCI remain underdeveloped. In this study, we aimed to enhance EI-XPCI image quality by adapting the Kernel Basis Network for Image Restoration (KBNet), a trained image denoising model. Our novel approach involves simultaneously inputting multiple phase step images for denoising, leveraging the inherent correlations between these images. Using our EI-XPCI simulation framework, we generated phase steps and subsequently denoised them using KBNet. The performance of KBNet was assessed in terms of the Peak Signal-to-Noise Ratio (PSNR) and the Structural Similarity Index Measure (SSIM). Our findings demonstrate that KBNet, with its multi-image input approach, outperforms state-of-the-art denoising techniques while preserving image details, offering a promising solution for noise reduction in EI-XPCI.

Accurate 3D mesh registration is essential in many industrial applications of X-ray imaging, as it allows quality assessment and inspection of manufactured objects. Conventional methods rely mainly on time-consuming and expensive X-ray computed tomography (X-CT) or ancillary camera systems. Instead, we propose a novel approach for efficient 3D multi-mesh registration in few-view industrial X-ray imaging scenarios. Our approach harnesses the capabilities of CAD-ASTRA, an X-ray mesh projector, compatible with the ASTRA toolbox and popular GPU libraries such as CuPy and PyTorch, for the simulation of X-ray projec tions from a known object surface mesh. As a differentiable program, CAD-ASTRA allows iterative improvement of the objects’ position in space by back-propagation of a differentiable measure of the projection error. The potential of this approach is demonstrated through tests on simultaneous multiple object registration in a poly-chromatic imaging, even in cases where the spectral characteristics of the imaging system are unknown. Results from a diverse set of real experiments highlight the efficacy of mesh registration, achieving successful registrations even when only two projections at a 10 ^∘ angle relative to the scanning system center are available. The mesh projector facilitates resource-efficient registration in industrial applications with few viewpoints, thereby reducing the demand for resources and eliminating the need for X-CT reconstruction.

Additive Manufacturing (AM) has emerged as a manufacturing process that allows the direct production of samples from digital models. To ensure that quality standards are met in all manufactured samples of a batch, X-ray computed tomography (X-CT) is often used combined with automated anomaly detection. For the latter, deep learning (DL) anomaly detection techniques are increasingly, as they can be trained to be robust to the material being analysed and resilient towards poor image quality. Unfortunately, most recent and popular DL models have been developed for 2D image processing, thereby disregarding valuable volumetric information. This study revisits recent supervised (UNet, UNet++, UNet 3+, MSS-UNet) and unsupervised (VAE, ceVAE, gmVAE, vqVAE) DL models for porosity analysis of AM samples from X-CT images and extends them to accept 3D input data with a 3D-patch pipeline for lower computational requirements, improved efficiency and generalisability. The supervised models were trained using the Focal Tversky loss to address class imbalance that arises from the low porosity in the training datasets. The output of the unsupervised models is post-processed to reduce misclassifications caused by their inability to adequately represent the object surface. The findings were cross-validated in a 5-fold fashion and include: a performance benchmark of the DL models, an evaluation of the post-processing algorithm, an evaluation of the effect of training supervised models with the output of unsupervised models. In a final performance benchmark on a test set with poor image quality, the best performing supervised model was UNet++ with an average precision of 0.751 $\pm$ 0.030, while the best unsupervised model was the post-processed ceVAE with 0.830 $\pm$ 0.003. The VAE/ceVAE models demonstrated superior capabilities, particularly when leveraging post-processing techniques.

Accurate and fast simulation of X-ray projection data from mesh models has many applications in academia and industry, ranging from 3D X-ray computed tomography (XCT) reconstruction algorithms to radiograph-based object inspection and quality control. While software tools for the simulation of X-ray projection data from mesh models are available, they lack either performance, public availability, flexibility to implement non-standard scanning geometries, or easy integration with existing 3D XCT software. In this paper, we propose CAD-ASTRA, a highly versatile toolbox for fast simulation of X-ray projection data from mesh models. While fully functional as standalone software, it is also compatible with the ASTRA toolbox, an open-source toolbox for flexible tomographic reconstruction. CAD-ASTRA provides three specialized GPU projectors based on state-of-the-art algorithms for 3D rendering, implemented using the NVIDIA CUDA Toolkit and the OptiX engine. First, it enables X-ray phase contrast simulations by modeling refraction through ray tracing. Second, it allows the back-propagation of projective errors to mesh vertices, enabling immediate application in mesh reconstruction, deep learning, and other optimization routines. Finally, CAD-ASTRA allows simulation of polychromatic X-ray projections from heterogeneous objects with a source of finite focal spot size. Use cases on a CAD-based inspection task, a phase contrast experiment, a combined mesh-volumetric data projection, and a mesh reconstruction demonstrate the wide applicability of CAD-ASTRA.

In selective laser melting (SLM), the 3D-printed metal objects can deform, causing dimensional inaccuracies and potentially damaging the printer's recoater. While layerwise monitoring systems are available to detect these deformations, higher-speed detection would enable a controller to intervene to reduce this defect. In this work we propose the first such in-situ monitoring system that detects part deformations through the analysis of melt pool images. The system uses high-speed cameras with frame rates up to 20,000 frames per second in the visual and short-wave infrared spectrum to record the melt pool. An adaptive weight network is then used to predict deformation magnitudes based on the image data and contextual printing information such as part geometry and hatching patterns. The addition of this contextual information allows the model to normalize for within-layer printing variations when interpreting the image data. Our model was validated on three print ob-jects that display deformations resulting from different physical sources. Our model's predictions were compared to deformation measurements taken from X-ray CT scans and showed both a strong correlation with the ground truth, with average relative errors between 17 % and 22%, and generalized well to untrained deformation types. The prediction model is also able to operate at very high speeds (> 20 kHz) to facilitate high-speed intra-layer control loops.

A static, multi-source X-ray Computed Tomography (CT) system facilitates rapid multi-view X-ray radiography, significantly improving the efficiency of cargo scanning. However, reconstructing images from sparse-view X-ray data in cargo scanning is challenging, particularly when conventional deep learning reconstruction techniques are hampered by a scarcity of training data. This work proposes the application of Deep Image Prior (DIP), which does not require training data, to reduce undersampling reconstruction artifacts arising from sparse-view and restricted opening angle acquisition in X-ray CT systems tailored for large-scale cargo scanning in harbors.(1) The work particularly targets a rectangular multi-source X-ray CT system, featuring up to 40 equidistantly distributed static X-ray sources with a 30-degree opening angle.(2) Our study demonstrates that DIP improves the quality of of sparse-view cargo CT in terms of PSNR and SSIM compared to traditional reconstruction methods.

The majority of the recent iterative approaches in 4DCT not only rely on nested iterations, thereby increasing computational complexity and constraining potential acceleration, but also fail to provide a theoretical proof of convergence for their proposed iterative schemes. On the other hand, the latest MATLAB and Python image processing toolboxes lack the implementation of analytic adjoints of affine motion operators for 3D object volumes, which does not allow gradient methods using exact derivatives towards affine motion parameters. In this work, we propose the Simultaneous Affine Motion-Compensated Image Reconstruction Technique (SAMCIRT)- an efficient iterative reconstruction scheme that combines image reconstruction and affine motion estimation in a single update step, based on the analytic adjoints of the motion operators then exact partial derivatives with respect to both the reconstruction and the affine motion parameters. Moreover, we prove the separated Lipschitz continuity of the objective function and its associated functions, including the gradient, which supports the convergence of our proposed iterative scheme, despite the non-convexity of the objective function with respect to the affine motion parameters. Results from simulation and real experiments show that our method outperforms the state-of-the-art CT reconstruction with affine motion correction methods in computational feasibility and projection distance. In particular, this allows accurate reconstruction for a real, nonstationary diamond, showing a novel application of 4DCT.

The manufacturing of metal parts via powder-bed fusion is often still facing quality issues due to microstructural porosity. Minimizing this porosity remains a priority and requires the optimization of printing process parameters. While the analysis of printed parts using X-ray computed tomography can localize and identify the pore types (e.g. keyhole or lack-of-fusion pores), these pore types can be difficult to identify across printer settings and print materials. Therefore, there is a need for a material and process agnostic approach. This work presents such an approach by considering a set of geometric pore features that do not differ considerably across print scenarios. These features are then leveraged for supervised pore type classification. The distributions of pore features were analyzed in different materials and under varying laser parameters, showing that they behave in a generic way. For classification, it is observed that they outperform other features leveraged in the state-of-the-art for pore classification in a single material, reaching up to 93.0% accuracy. Additionally, accuracies up to 90.2% for cross-material classification were observed by training on pores of one material and validating on another. These results pave the way to a general-purpose pore classification method usable across materials and process conditions.

Additive manufacturing (AM) is increasingly gaining interest as a low-waste production technique, capable of producing objects using a computer-aided design file. It is particularly interesting for rapid prototyping of parts and manufacturing objects that have complex shapes. However, as in the case of AM through selective laser melting (SLM), manufactured objects may contain defects that can seriously alter their properties. These defects may appear as pores, which can be detected by X-ray computed tomography (X-CT) in a non-destructive manner. CT images can simply be segmented by thresholding or through more advanced techniques such as discrete X-CT reconstruction or machine learning techniques. Nevertheless, these techniques are vulnerable to image reconstruction artefacts. In this work, we evaluate the performance of state-of-the-art, deeply supervised 3D deep learning networks (UNet++, UNet 3+ and UNet-MSS) in terms of segmentation performance of pores from X-ray CT images. The networks have been trained on a real CT dataset, with (noisy) labels produced from both conventional thresholding of the CT images as well as more advanced discrete polychromatic reconstructions. Furthermore, the performance of the networks was evaluated on a test dataset with severe CT artefacts. Pore segmentation from real CT images, which include noise and reconstruction artefacts, revealed that the best performing network was UNet++ with an average Sorensen-Dice score of 0.869 +/- 0.006.

To create an accurate 3D reconstruction of the vascular trees, it is necessary to know the exact geometrical parameters of the angiographic imaging system. Many previous studies used vascular structures to estimate the system’s exact geometry. However, utilizing interventional devices and their relative features may be less challenging, as they are unique in different views. We present a semi-automatic self-calibration approach considering the markers attached to the interventional instruments to estimate the accurate geometry of a biplane X-ray angiography system for neuroradiologic use. A novel approach is proposed to detect and segment the markers using machine learning classification, a combination of support vector machine and boosted tree. Then, these markers are considered as reference points to optimize the acquisition geometry iteratively. The method is evaluated on four clinical datasets and three pairs of phantom angiograms. The mean and standard deviation of backprojection error for the catheter or guidewire before and after self-calibration are $$7.13\pm 6.47$$  mm and $$0.10\pm 0.06$$  mm, respectively. The mean and standard deviation of the 3D root-mean-square error (RMSE) for some markers in the phantom reduced from $$0.51\pm 0.11$$ to $$0.31\pm 0.08$$  mm. A semi-automatic approach to estimate the accurate geometry of the C-arm system was presented. Results show the reduction in the 2D backprojection error as well as the 3D RMSE after using our proposed self-calibration technique. This approach is essential for 3D reconstruction of the vascular trees or post-processing techniques of angiography systems that rely on accurate geometry parameters.

X-ray computed tomography (X-CT) plays an important role in non-destructive quality inspection and process evaluation in metal additive manufacturing, as several types of defects such as keyhole and lack of fusion pores can be observed in these 3D images as local changes in material density. Segmentation of these defects often relies on threshold methods applied to the reconstructed attenuation values of the 3D image voxels. However, the segmentation accuracy is affected by unavoidable X-CT reconstruction features such as partial volume effects, voxel noise and imaging artefacts. These effects create false positives, difficulties in threshold value selection and unclear or jagged defect edges. In this paper, we present a new X-CT defect segmentation method based on preprocessing the X-CT image with a 3D total variation denoising method. By comparing the changes in the histogram, threshold selection can be significantly better, and the resulting segmentation is of much higher quality. We derive the optimal algorithm parameter settings and demonstrate robustness for deviating settings. The technique is presented on simulated data sets, compared between low- and high-quality X-CT scans, and evaluated with optical microscopy after destructive tests.

Beam hardening and scattering effects can seriously degrade image quality in polychromatic X-ray CT imaging. In recent years, polychromatic image reconstruction techniques and scatter estimation using Monte Carlo simulation have been developed to compensate for beam hardening and scattering CT artifacts, respectively. Both techniques require knowledge of the X-ray tube energy spectrum. In this work, Monte Carlo simulations were used to calculate the X-ray energy spectrum of FleXCT, a novel prototype industrial micro-CT scanner, enabling beam hardening and scatter reduction for CT experiments. Both source and detector were completely modeled by Monte Carlo simulation. In order to validate the energy spectra obtained via Monte Carlo simulation, they were compared with energy spectra obtained via a second method. Here, energy spectra were calculated from empirical measurements using a step wedge sample, in combination with the Maximum Likelihood Expectation Maximization (MLEM) method. Good correlation was achieved between both approaches, confirming the correct modeling of the FleXCT system by Monte Carlo simulation. After validation of the modeled FleXCT system through comparing the X-ray spectra for different tube voltages inside the detector, we calculated the X-ray spectrum of the FleXCT X-ray tube, independent of the flat panel detector response, which is a prerequisite for beam hardening and scattering CT artifacts.

High density materials, such as metals, strongly scatter Xray photons during X-ray Computed Tomography (CT) scans, which is detrimental to the quality of the reconstructed images. In this study, a scatter compensation method for X-ray CT, based on Monte-Carlo (MC) simulations from the object's CAD model, is presented and employed in conjunction with polychromatic reconstructions. The estimation of the scatter contributions is accelerated by 1) reducing the number of simulated projections accordingly to the Nyquist theorem, 2) noise reduction 3) angular interpolation. The method was applied to enhance CT images of a steel object produced via Additive Manufacturing, whose CAD model is known. Results show that, in conjunction with polychromatic reconstruction, our method can efficiently reduce beam hardening, scattering artifacts and increase the contrast of defects within the object.

Cone-beam computed tomography (CBCT) is a widely used technique for diagnostic or monitoring purposes. Compared to the traditional CT, a CBCT is more affected by scatter artifacts because of the large volume being irradiated by the beam. The research divulged in this paper is about the assessment of the influence that metallic implants may have on degrading image quality of CBCT due to scattered radiation. The evaluation method is based on Monte-Carlo (MC) simulations of the physical processes that X-ray photons undergo in typical CBCT setups, in presence and absence of highly scattering metallic implants (coils used for treatment of aneurysms and pacemakers). The results show that the scattered radiation caused by metallic objects and reaching the detector produces slight degradation of CBCT image quality and, moreover, it is demonstrated that the intrinsic absorption and beam-hardening effect of these implants have bigger impact on the overall image fidelity.

Dynamic MRI is a technique of acquiring a series of images continuously to follow the physiological changes over time. However, such fast imaging results in low resolution images. In this work, abdominal deformation model computed from dynamic low resolution images have been applied to high resolution image, acquired previously, to generate dynamic high resolution MRI. Dynamic low resolution images were simulated into different breathing phases (inhale and exhale). Then, the image registration between breathing time points was performed using the B-spline SyN deformable model and using cross-correlation as a similarity metric. The deformation model between different breathing phases were estimated from highly undersampled data. This deformation model was then applied to the high resolution images to obtain high resolution images of different breathing phases. The results indicated that the deformation model could be computed from relatively very low resolution images.