Deep X-ray Research Articles

Correction: Detecting and classifying hidden defects in additively manufactured parts using deep learning and X-ray computed tomography

Correction: Detecting and classifying hidden defects in additively manufactured parts using deep learning and X-ray computed tomography

Three-Dimensional Label-Free Observing of the Self-Assembled Nanoparticles inside a Single Cell at Nanoscale Resolution.

Understanding the intracellular behavior of nanoparticles (NPs) plays a key role in optimizing the self-assembly performance of nanomedicine. However, conducting the 3D, label-free, quantitative observation of self-assembled NPs within intact single cells remains a substantial challenge in complicated intracellular environments. Here, we propose a deep learning combined synchrotron radiation hard X-ray nanotomography approach to visualize the self-assembled ultrasmall iron oxide (USIO) NPs in a single cell. The method allows us to explore comprehensive information on NPs, such as their distribution, morphology, location, and interaction with cell organelles, and provides quantitative analysis of the heterogeneous size and morphologies of USIO NPs under diverse conditions. This label-free, in situ method provides a tool for precise characterization of intracellular self-assembled NPs to improve the evaluation and design of a bioresponsive nanomedicine.

The SRG/eROSITA All-Sky Survey

Aims. Characterising galaxy cluster populations from a catalogue of sources selected in astronomical surveys requires knowledge of sample incompleteness, known as the selection function. The first All-Sky Survey (eRASS1) by eROSITA on board Spectrum Roentgen Gamma (SRG) has enabled the collection of large samples of galaxy clusters detected in the soft X-ray band over the western Galactic hemisphere. The driving goal consists in constraining cosmological parameters, which puts stringent requirements on the accuracy and flexibility of explainable selection function models. Methods. We used a large set of mock observations of the eRASS1 survey and we processed simulated data identically to the real eRASS1 events. We matched detected sources to simulated clusters and we associated detections to intrinsic cluster properties. We trained a series of models to build selection functions depending only on observable surface brightness data. We developed a second series of models relying on global cluster characteristics such as X-ray luminosity, flux, and the expected instrumental count rate as well as on morphological properties. We validated our models using our simulations and we ranked them according to selected performance metrics. We validated the models with datasets of clusters detected in X-rays and via the Sunyaev–Zeldovich effect. We present the complete Bayesian population modelling framework developed for this purpose. Results. Our results reveal the surface brightness characteristics most relevant to cluster selection in the eRASS1 sample, in particular the ambiguous role of central surface brightness at the scale of the instrument resolution. We have produced a series of user-friendly selection function models and demonstrated their validity and their limitations. Our selection function for bright sources reproduces the catalogue matches with external datasets well. We discuss potential inconsistencies in the selection models at a low signal-to-noise revealed by comparison with a deep X-ray sample acquired by eROSITA during its performance verification phase. Conclusions. Detailed modelling of the eRASS1 galaxy cluster selection function is made possible by reformulating selection into a classification problem. Our models are used in the first eRASS1 cosmological analysis and in sample studies of eRASS1 cluster and groups. These models are crucial for science with eROSITA cluster samples and our new methods pave the way for further investigation of faint cluster selection effects.

Detecting and classifying hidden defects in additively manufactured parts using deep learning and X-ray computed tomography

Detecting and classifying hidden defects in additively manufactured parts using deep learning and X-ray computed tomography

Advancing X-ray Inspection with Deep Learning De-noising Technology

X-ray inspection plays a pivotal role in the food and non-destructive testing industries, ensuring optimal products quality and safety. To improve contaminants and defects detection, it is crucial to reduce the amount of noise in the acquired images. Addressing this, Hamamatsu Photonics has developed a new de-noising technology based on deep learning algorithms and an innovative X-ray simulation method.

Incremental convolutional transformer for baggage threat detection

Detecting cluttered and overlapping contraband items from baggage scans is one of the most challenging tasks, even for human experts. Recently, considerable literature has grown up around the theme of deep learning-based X-ray screening for localizing contraband data. However, the existing threat detection systems are still vulnerable to high occlusion, clutter, and concealment. Furthermore, they require exhaustive training routines on large-scale and well-annotated data in order to produce accurate results. To overcome the above-mentioned limitations, this paper presents a novel convolutional transformer system that recognizes different overlapping instances of prohibited objects in complex baggage X-ray scans via a distillation-driven incremental instance segmentation scheme. Furthermore, unlike its competitors, the proposed framework allows an incremental integration of new item instances while avoiding costly training routines. In addition to this, the proposed framework also outperforms state-of-the-art approaches by achieving a mean average precision score of 0.7896, 0.5974, and 0.7569 on publicly available GDXray, SIXray, and OPIXray datasets for detecting concealed and cluttered baggage threats.

Non-destructive larval infestation detection in pear fruits using deep learning and x-ray CT generated radiographs

Non-destructive larval infestation detection in pear fruits using deep learning and x-ray CT generated radiographs

The Magellan M2FS Spectroscopic Survey of High-redshift Galaxies: The Brightest Lyman-break Galaxies at z ∼ 6

We present a study of a sample of 45 spectroscopically confirmed, UV luminous galaxies at z ∼ 6. They were selected as bright Lyman-break galaxies (LBGs) using deep multiband optical images in more than 2 deg2 of the sky, and subsequently identified via their strong Lyα emission. The majority of these LBGs span an absolute UV magnitude range from −22.0 to −20.5 mag with Lyα equivalent width (EW) between ∼10 and ∼200 Å, representing the most luminous galaxies at z ∼ 6 in terms of both UV continuum emission and Lyα line emission. We model the spectral energy distributions of 10 LBGs that have deep infrared observations from Hubble Space Telescope, JWST, and/or Spitzer, and find that they have a wide range of stellar masses and ages. They also have high star formation rates ranging from a few tens to a few hundreds of solar mass per year. Five of the LBGs have JWST or HST images, and four of them show compact morphology in these images, including one that is roughly consistent with a point source, suggesting that UV luminous galaxies at this redshift are generally compact. The fraction of our photometrically selected LBGs with strong Lyα emission (EW > 25 Å) is about 0.2, which is consistent with previous results and supports a moderate evolution of the intergalactic medium opacity at the end of cosmic reionization. Using deep X-ray images, we do not find evidence of strong active galactic nucleus (AGN) activity in these galaxies, but our constraint is loose, and we are not able to rule out the possibility of any weak AGN activity.

Roaring to Softly Whispering: X-Ray Emission after ∼3.7 yr at the Location of the Transient AT2018cow and Implications for Accretion-powered Scenarios* *Based on observations obtained with XMM-Newton, an ESA science mission with instruments and contributions directly funded by ESA Member States and NASA.

We present the first deep X-ray observations of luminous fast blue optical transient (LFBOT) AT 2018cow at ∼3.7 yr since discovery, together with the reanalysis of the observation at δ t ∼ 220 days. X-ray emission is significantly detected at a location consistent with AT 2018cow. The very soft X-ray spectrum and sustained luminosity are distinct from the spectral and temporal behavior of the LFBOT in the first ∼100 days and would possibly signal the emergence of a new emission component, although a robust association with AT 2018cow can only be claimed at δ t ∼ 220 days, while at δ t ∼ 1350 days contamination of the host galaxy cannot be excluded. We interpret these findings in the context of the late-time panchromatic emission from AT 2018cow, which includes the detection of persistent, slowly fading UV emission with ν L ν ≈ 1039 erg s−1. Similar to previous works (and in analogy with arguments for ultraluminous X-ray sources), these late-time observations are consistent with thin disks around intermediate-mass black holes (with M • ≈ 103–104 M ☉) accreting at sub-Eddington rates. However, differently from previous studies, we find that smaller-mass black holes with M • ≈ 10–100 M ☉ accreting at ≳the Eddington rate cannot be ruled out and provide a natural explanation for the inferred compact size (R out ≈ 40 R ☉) of the accretion disk years after the optical flare. Most importantly, irrespective of the accretor mass, our study lends support to the hypothesis that LFBOTs are accretion-powered phenomena and that, specifically, LFBOTs constitute electromagnetic manifestations of super-Eddington accreting systems that evolve to ≲Eddington over a ≈100-day timescale.

Mammography with deep learning for breast cancer detection.

X-ray mammography is currently considered the golden standard method for breast cancer screening, however, it has limitations in terms of sensitivity and specificity. With the rapid advancements in deep learning techniques, it is possible to customize mammography for each patient, providing more accurate information for risk assessment, prognosis, and treatment planning. This paper aims to study the recent achievements of deep learning-based mammography for breast cancer detection and classification. This review paper highlights the potential of deep learning-assisted X-ray mammography in improving the accuracy of breast cancer screening. While the potential benefits are clear, it is essential to address the challenges associated with implementing this technology in clinical settings. Future research should focus on refining deep learning algorithms, ensuring data privacy, improving model interpretability, and establishing generalizability to successfully integrate deep learning-assisted mammography into routine breast cancer screening programs. It is hoped that the research findings will assist investigators, engineers, and clinicians in developing more effective breast imaging tools that provide accurate diagnosis, sensitivity, and specificity for breast cancer.

Feeding and feedback processes in the Spiderweb proto-intracluster medium

Context. We present a detailed analysis of the thermal, diffuse emission of the proto-intracluster medium (proto-ICM) detected in the halo of the Spiderweb Galaxy at z = 2.16, within a radius of ∼150 kpc. Aims. Our main goal is to derive the thermodynamic profiles of the proto-ICM, establish the potential presence of a cool core and constrain the classical mass deposition rate (MDR) that may feed the nuclear and the star formation (SF) activity, and estimate the available energy budget of the ongoing feedback process. Methods. We combined deep X-ray data from Chandra and millimeter observations of the Sunyaev–Zeldovich (SZ) effect obtained by the Atacama Large Millimeter/submillimeter Array (ALMA). Results. Thanks to independent measurements of the pressure profile from the ALMA SZ observation and the electron density profile from the available X-ray data, we derived, for the first time, the temperature profile in the ICM of a z > 2 protocluster. It reveals the presence of a strong cool core (comparable to local ones) that may host a significant mass deposition flow, consistent with the measured local SF values. We also find mild evidence of an asymmetry in the X-ray surface brightness distribution, which may be tentatively associated with a cavity carved into the proto-ICM by the radio jets. In this case, the estimated average feedback power would be in excess of ∼1043 erg s−1. Alternatively, the asymmetry may be due to the young dynamical status of the halo. Conclusions. The cooling time of baryons in the core of the Spiderweb protocluster is estimated to be ∼0.1 Gyr, implying that the baryon cycle in the first stages of protocluster formation is characterized by a high-duty cycle and a very active environment. In the case of the Spiderweb protocluster, we are witnessing the presence of a strongly peaked core that is possibly hosting a cooling flow with a MDR up to 250–1000 M⊙ yr−1, responsible for feeding both the central supermassive black hole (SMBH) and the high star formation rate (SFR) observed in the Spiderweb Galaxy. This phase is expected to be rapidly followed by active galactic nucleus (AGN) feedback events, whose onset may have already left an imprint in the radio and X-ray appearance of the Spiderweb protocluster, eventually driving the ICM into a self-regulated, long-term evolution in less than one Gyr.

German CheXpert Chest X-ray Radiology Report Labeler.

The aim of this study was to develop an algorithm to automatically extract annotations from German thoracic radiology reports to train deep learning-based chest X-ray classification models. An automatic label extraction model for German thoracic radiology reports was designed based on the CheXpert architecture. The algorithm can extract labels for twelve common chest pathologies, the presence of support devices, and "no finding". For iterative improvements and to generate a ground truth, a web-based multi-reader annotation interface was created. With the proposed annotation interface, a radiologist annotated 1086 retrospectively collected radiology reports from 2020-2021 (data set 1). The effect of automatically extracted labels on chest radiograph classification performance was evaluated on an additional, in-house pneumothorax data set (data set 2), containing 6434 chest radiographs with corresponding reports, by comparing a DenseNet-121 model trained on extracted labels from the associated reports, image-based pneumothorax labels, and publicly available data, respectively. Comparing automated to manual labeling on data set 1: "mention extraction" class-wise F1 scores ranged from 0.8 to 0.995, the "negation detection" F1 scores from 0.624 to 0.981, and F1 scores for "uncertainty detection" from 0.353 to 0.725. Extracted pneumothorax labels on data set 2 had a sensitivity of 0.997 [95 % CI: 0.994, 0.999] and specificity of 0.991 [95 % CI: 0.988, 0.994]. The model trained on publicly available data achieved an area under the receiver operating curve (AUC) for pneumothorax classification of 0.728 [95 % CI: 0.694, 0.760], while the models trained on automatically extracted labels and on manual annotations achieved values of 0.858 [95 % CI: 0.832, 0.882] and 0.934 [95 % CI: 0.918, 0.949], respectively. Automatic label extraction from German thoracic radiology reports is a promising substitute for manual labeling. By reducing the time required for data annotation, larger training data sets can be created, resulting in improved overall modeling performance. Our results demonstrated that a pneumothorax classifier trained on automatically extracted labels strongly outperformed the model trained on publicly available data, without the need for additional annotation time and performed competitively compared to manually labeled data. · An algorithm for automatic German thoracic radiology report annotation was developed.. · Automatic label extraction is a promising substitute for manual labeling.. · The classifier trained on extracted labels outperformed the model trained on publicly available data.. · Wollek A, Hyska S, Sedlmeyr T et al. German CheXpert Chest X-ray Radiology Report Labeler. Fortschr Röntgenstr 2024; 196: 956 - 965.

Relation-aware graph convolutional network for waste battery inspection based on X-ray images

With the popularization of various electronic devices, the demand for batteries is rapidly increasing. Batteries bring convenience to our daily lives, but how to effectively achieve battery recycling is a challenge. In fact, improper disposal can cause serious environmental pollution and resource waste. Currently, it is inefficient and time-consuming to manually locate batteries from waste electronic devices by employing a large number of professional workers. Therefore, we propose an efficient solution based on deep learning and X-ray images, which can automatically detect batteries without opening waste electronic devices. Specifically, we construct a High-quality X-ray Battery Inspection dataset called HiXray-BI, which not only contains 6000 X-ray images collected from the existing dataset, but also provides 23,994 battery instances. In addition, we propose a novel deep learning model called Relation-Aware Graph Convolutional Network (RA-GCN) for X-ray waste battery inspection, which introduces a novel Multi-scale Feature Relation Module (MFRM) and Semantic Feature Relation Module (SFRM) to improve detection performance. Experimental results on the HiXray-BI dataset show that our method compares favorably with state-of-the-art methods. Our solution provides new insights into recycling waste batteries, which is beneficial for environmental protection and sustainable development.

Auto-detection of the coronavirus disease by using deep convolutional neural networks and X-ray photographs

The most widely used method for detecting Coronavirus Disease 2019 (COVID-19) is real-time polymerase chain reaction. However, this method has several drawbacks, including high cost, lengthy turnaround time for results, and the potential for false-negative results due to limited sensitivity. To address these issues, additional technologies such as computed tomography (CT) or X-rays have been employed for diagnosing the disease. Chest X-rays are more commonly used than CT scans due to the widespread availability of X-ray machines, lower ionizing radiation, and lower cost of equipment. COVID-19 presents certain radiological biomarkers that can be observed through chest X-rays, making it necessary for radiologists to manually search for these biomarkers. However, this process is time-consuming and prone to errors. Therefore, there is a critical need to develop an automated system for evaluating chest X-rays. Deep learning techniques can be employed to expedite this process. In this study, a deep learning-based method called Custom Convolutional Neural Network (Custom-CNN) is proposed for identifying COVID-19 infection in chest X-rays. The Custom-CNN model consists of eight weighted layers and utilizes strategies like dropout and batch normalization to enhance performance and reduce overfitting. The proposed approach achieved a classification accuracy of 98.19% and aims to accurately classify COVID-19, normal, and pneumonia samples.

Deep Learning-based Lung dose Prediction Using Chest X-ray Images in Non-small Cell Lung Cancer Radiotherapy.

This study aimed to develop a deep learning model for the prediction of V20 (the volume of the lung parenchyma that received ≥20 Gy) during intensity-modulated radiation therapy using chest X-ray images. The study utilized 91 chest X-ray images of patients with lung cancer acquired routinely during the admission workup. The prescription dose for the planning target volume was 60 Gy in 30 fractions. A convolutional neural network-based regression model was developed to predict V20. To evaluate model performance, the coefficient of determination (R2), root mean square error (RMSE), and mean absolute error (MAE) were calculated with conducting a four-fold cross-validation method. The patient characteristics of the eligible data were treatment period (2018-2022) and V20 (19.3%; 4.9%-30.7%). The predictive results of the developed model for V20 were 0.16, 5.4%, and 4.5% for the R2, RMSE, and MAE, respectively. The median error was -1.8% (range, -13.0% to 9.2%). The Pearson correlation coefficient between the calculated and predicted V20 values was 0.40. As a binary classifier with V20 <20%, the model showed a sensitivity of 75.0%, specificity of 82.6%, diagnostic accuracy of 80.6%, and area under the receiver operator characteristic curve of 0.79. The proposed deep learning chest X-ray model can predict V20 and play an important role in the early determination of patient treatment strategies.

Deep Hard X-ray Survey of the M81 Field Based on INTEGRAL1 Data

Deep Hard X-ray Survey of the M81 Field Based on INTEGRAL1 Data

First detection of X-ray pulsations and spectrum of the high Galactic latitude pulsar PSR J0837–2454 and direct Urca cooling implications

Abstract PSR J0837 $-$ 2454 is a young 629 ms radio pulsar whose uncertain distance has important implications. A large distance would place the pulsar far out of the Galactic plane and suggest it is the result of a runaway star, while a short distance would mean the pulsar is extraordinarily cold. Here we present further radio observations and the first deep X-ray observation of PSR J0837 $-$ 2454. Data from the Parkes Murriyang telescope show flux variations over short and long timescales and also yield an updated timing model, while the position and proper motion (and, less strongly, parallax) of the pulsar are constrained by a number of low-significance detections with the Very Long Baseline Array. XMM-Newton data enable detection of X-ray pulsations for the first time from this pulsar and yield a spectrum that is thermal and blackbody-like, with a cool blackbody temperature $\approx$ $70\ \mbox{eV}$ or atmosphere temperature $\approx$ $50\ \mbox{eV}$ , as well as a small hotspot. The spectrum also indicates the pulsar is at a small distance of $\lesssim$ $1\ \mbox{kpc}$ , which is compatible with the marginal VLBA parallax constraint that favours a distance of $\gtrsim$ 330 pc. The low implied luminosity ( $\sim7.6\times10^{31}\mbox{erg\, s}^{-1}$ at 0.9 kpc) suggests PSR J0837 $-$ 2454 has a mass high enough that fast neutrino emission from direct Urca reactions operates in this young star and points to a nuclear equation of state that allows for direct Urca reactions at the highest densities present in neutron star cores.

Validation of a Deep Learning Chest X-ray Interpretation Model: Integrating Large-Scale AI and Large Language Models for Comparative Analysis with ChatGPT.

This study evaluates the diagnostic accuracy and clinical utility of two artificial intelligence (AI) techniques: Kakao Brain Artificial Neural Network for Chest X-ray Reading (KARA-CXR), an assistive technology developed using large-scale AI and large language models (LLMs), and ChatGPT, a well-known LLM. The study was conducted to validate the performance of the two technologies in chest X-ray reading and explore their potential applications in the medical imaging diagnosis domain. The study methodology consisted of randomly selecting 2000 chest X-ray images from a single institution's patient database, and two radiologists evaluated the readings provided by KARA-CXR and ChatGPT. The study used five qualitative factors to evaluate the readings generated by each model: accuracy, false findings, location inaccuracies, count inaccuracies, and hallucinations. Statistical analysis showed that KARA-CXR achieved significantly higher diagnostic accuracy compared to ChatGPT. In the 'Acceptable' accuracy category, KARA-CXR was rated at 70.50% and 68.00% by two observers, while ChatGPT achieved 40.50% and 47.00%. Interobserver agreement was moderate for both systems, with KARA at 0.74 and GPT4 at 0.73. For 'False Findings', KARA-CXR scored 68.00% and 68.50%, while ChatGPT scored 37.00% for both observers, with high interobserver agreements of 0.96 for KARA and 0.97 for GPT4. In 'Location Inaccuracy' and 'Hallucinations', KARA-CXR outperformed ChatGPT with significant margins. KARA-CXR demonstrated a non-hallucination rate of 75%, which is significantly higher than ChatGPT's 38%. The interobserver agreement was high for KARA (0.91) and moderate to high for GPT4 (0.85) in the hallucination category. In conclusion, this study demonstrates the potential of AI and large-scale language models in medical imaging and diagnostics. It also shows that in the chest X-ray domain, KARA-CXR has relatively higher accuracy than ChatGPT.

Deep Learning-Based Hip X-ray Image Analysis for Predicting Osteoporosis

Osteoporosis is a common problem in orthopedic medicine, and it has become an important medical issue in orthopedics as Taiwan is gradually becoming an aging society. In the diagnosis of osteoporosis, the bone mineral density (BMD) derived from dual-energy X-ray absorptiometry (DXA) is the main criterion for orthopedic diagnosis of osteoporosis, but due to the high cost of this equipment and the lower penetration rate of the equipment compared to the X-ray images, the problem of osteoporosis has not been effectively solved for many people who suffer from osteoporosis. At present, in clinical diagnosis, doctors are not yet able to accurately interpret X-ray images for osteoporosis manually and must rely on the data obtained from DXA. In recent years, with the continuous development of artificial intelligence, especially in the fields of machine learning and deep learning, significant progress has been made in image recognition. Therefore, it is worthwhile to revisit the question of whether it is possible to use a convolutional neural network model to read a hip X-ray image and then predict the patient’s BMD. In this study, we proposed a hip X-ray image segmentation model and a hip X-ray image recognition classification model. First, we used the U-Net model as a framework to segment the femoral neck, greater trochanter, Ward’s triangle, and the total hip in the hip X-ray images. We then performed image matting and data augmentation. Finally, we constructed a predictive model for osteoporosis using deep learning algorithms. In the segmentation experiments, we used intersection over union (IoU) as the evaluation metric for image segmentation, and both the U-Net model and the U-Net++ model achieved segmentation results greater than or equal to 0.5. In the classification experiments, using the T-score as the classification basis, the total hip using the DenseNet121 model has the highest accuracy of 74%.

Versatility of thermoluminescence materials in thermal history sensing: a review

While several devices can monitor temperature, measuring non-volatile thermal history is far more challenging, especially for sensors placed in hazardous settings like fires and explosions. Herein, the utilisation of thermoluminescent microparticles in thermal history sensors is discussed. These transparent dielectric microparticles can store charge carriers for very extended periods due to their wide distribution of trap states. In its most basic form, an Arrhenius expression is strongly temperature-dependent for determining the population of the trap states. When exposed to high temperatures for a brief period, a particle with filled traps will preferentially lose carriers in shallow traps. To ascertain the temperature and timing of the thermal event, this depopulation creates a mark on the particle luminescence. By being exposed to deep ultraviolet, X-ray, beta, or gamma radiation, which fills the traps with charge carriers, particles are prepared many months before a test, if required. No matter if they are embedded in garbage or other inert materials, one or more particles can nevertheless be made to emit light. Laboratory tests using microheaters and high-energy explosives in the field are used to test and analyse the technology. It is demonstrated that a variety of high-explosive conditions may be used to accurately detect temperatures in the 200 ∘ C to 500 ∘ C range using thermoluminescent materials like LiF:Mg,Ti, MgB4O7:Dy,Li, and CaSO4:Ce,Tb. The article will provide the utilisation of different luminescent materials for sensing thermal history and analysing the particle luminescence mechanism through different thermal models for developing improved thermal history sensing technologies. The thermal history reconstruction techniques are also discussed to predict the thermal history. Therefore, the article will pave the way to understanding the luminescent properties of thermoluminescent microparticles for developing state-of-the-art thermal history sensors for measuring non-volatile thermal history and temperature in extreme environments such as fires and explosions.