Accurate X-ray Research Articles

Contrastive concept-phrase pre-training for generating clinically accurate and interpretable chest X-ray reports

Contrastive concept-phrase pre-training for generating clinically accurate and interpretable chest X-ray reports

An X-ray image-based Pruned Dense Convolution Neural Network for Tuberculosis Detection

According to the Ministry of Health in Kenya, tuberculosis (TB) is the fifth greatest cause of death and the main infectious disease killer in Kenya and across the world. In Kenya and throughout Africa, TB continues to wreak havoc on many vulnerable populations, homes, and communities despite being preventable and treatable. Common TB diagnostics, like blood and skin tests, frequently fail to identify the precise kind of TB. As a result, the World Health Organization (WHO) advises expanding the use of X-rays, for screening. In TB-prevalent regions of Kenya, a shortage of radiologists hampers effective screening and diagnosis, highlighting the need for scalable solutions for accurate X-ray analysis.Recent advancements in deep learning techniques have shown promise in the healthcare sector, particularly in radiology. However, many deep convolutional neural network (CNN) architectures are computationally intensive due to their size and resource requirements. This study designed and developed a Pruned CNN to address this issue by applying pruning techniques to baseline architectures. This approach significantly reduced model sizes while maintaining accuracy levels. Specifically, the pruned version of the DenseNet model achieved an impressive 99% accuracy with a reduction rate of 65.8%. These results highlight the potential of this pruned CNN as an effective and efficient tool for TB detection, particularly in resource-constrained environments. This study addresses the shortage of radiological expertise in many regions by providing a tool that can assist in the interpretation of X-ray images. This capability can help healthcare providers deliver timely and accurate diagnoses, thereby improving patient care.

Learning Scatter Artifact Correction in Cone-Beam X-Ray CT Using Incomplete Projections with Beam Hole Array

X-ray cone-beam computed tomography (CBCT) is a powerful tool for nondestructive testing and evaluation, yet the CT image quality can be compromised by artifact due to X-ray scattering within dense materials such as metals. This problem leads to the need for hardware- and software-based scatter artifact correction to enhance the image quality. Recently, deep learning techniques have merged as a promising approach to obtain scatter-free images efficiently. However, these deep learning techniques rely heavily on training data, often gathered through simulation. Simulated CT images, unfortunately, do not accurately reproduce the real properties of objects, and physically accurate X-ray simulation still requires significant computation time, hindering the collection of a large number of CT images. To address these problems, we propose a deep learning framework for scatter artifact correction using projections obtained solely by real CT scanning. To this end, we utilize a beam-hole array (BHA) to block the X-rays deviating from the primary beam path, thereby capturing scatter-free X-ray intensity at certain detector pixels. As the BHA shadows a large portion of detector pixels, we incorporate several regularization losses to enhance the training process. Furthermore, we introduce radiographic data augmentation to mitigate the need for long scanning time, which is a concern as CT devices equipped with BHA require two series of CT scans. Experimental validation showed that the proposed framework outperforms a baseline method that learns simulated projections where the entire image is visible and does not contain scattering artifacts.

Three-Stage Framework for Accurate Pediatric Chest X-ray Diagnosis Using Self-Supervision and Transfer Learning on Small Datasets.

Pediatric respiratory disease diagnosis and subsequent treatment require accurate and interpretable analysis. A chest X-ray is the most cost-effective and rapid method for identifying and monitoring various thoracic diseases in children. Recent developments in self-supervised and transfer learning have shown their potential in medical imaging, including chest X-ray areas. In this article, we propose a three-stage framework with knowledge transfer from adult chest X-rays to aid the diagnosis and interpretation of pediatric thorax diseases. We conducted comprehensive experiments with different pre-training and fine-tuning strategies to develop transformer or convolutional neural network models and then evaluate them qualitatively and quantitatively. The ViT-Base/16 model, fine-tuned with the CheXpert dataset, a large chest X-ray dataset, emerged as the most effective, achieving a mean AUC of 0.761 (95% CI: 0.759-0.763) across six disease categories and demonstrating a high sensitivity (average 0.639) and specificity (average 0.683), which are indicative of its strong discriminative ability. The baseline models, ViT-Small/16 and ViT-Base/16, when directly trained on the Pediatric CXR dataset, only achieved mean AUC scores of 0.646 (95% CI: 0.641-0.651) and 0.654 (95% CI: 0.648-0.660), respectively. Qualitatively, our model excels in localizing diseased regions, outperforming models pre-trained on ImageNet and other fine-tuning approaches, thus providing superior explanations. The source code is available online and the data can be obtained from PhysioNet.

Low-contrast X-ray image defect segmentation via a novel core-profile decomposition network

Accurate X-ray image defect segmentation is of paramount importance in industrial contexts, as it is the foundation for product quality control and production safety. Deep learning (DL) has demonstrated powerful image scene understanding capabilities and has achieved unprecedented performance in defect segmentation tasks. However, existing DL methods suffer from significant performance degradation when facing low-contrast X-ray images, as the core information of defects is often obscured and the profile details are ambiguous. To address this issue, this paper explicitly decomposes the X-ray defect segmentation task into two subtasks: core feature learning and elasticity profile refinement, allowing task “serial” decomposition and performance “parallel” improvement. On this basis, a novel core-profile decomposition network (CPDNet) is developed to achieve accurate defect segmentation of X-ray images. Specifically, the core feature learning module is designed to construct the effective feature space from two views, discriminative and structural, to extract defect-related core features from X-ray images. Subsequently, the elasticity profile refinement module is developed to further improve the defect segmentation performance, which makes the first attempt to define the profile refinement as an out-of-distribution detection and leverage the elasticity score to refine the profile details at the pixel level. To fully evaluate the presented method, we conduct a series of experiments using two real-world X-ray defect datasets, and the results demonstrate that the CPDNet outperforms state-of-the-art methods.

TO ANALYZE THE LUNGS X-RAY IMAGES USING MACHINE LEARNING ALGORITHM: AN IMPLEMENTATION TO PNEUMONIA DIAGNOSIS

Introduction: Respiratory diseases, particularly pneumonia, pose a significant threat to human life. Pneumonia affects the respiratory function in the human body and is a dangerous lung disease. This study aims to propose a model for detecting pneumonia in chest XR images. By utilizing statistical-based features, relevant and informative features are extracted from lung X-ray images. Objective: The objective is to obtain high accuracy in pneumonia identification; the target of this work is to generate a model that can precisely recognize the presence of pneumonia by evaluating chest X-ray pictures. Method: The Method follows a three-phase approach: preprocessing, categorization, and extraction of features. Preprocessing is the stage when various filters are applied to the chest X-ray images to enhance their eminence and eradicate noise. The feature extraction phase involves extracting statistical-based features from the preprocessed images. These features capture relevant information regarding a pneumonia diagnosis. Finally, in the classification phase, algorithms for machine learning are employed to use the retrieved features to categorize the X-ray pictures as infected or uninfected. Result: The proposed model successfully detects the presence of pneumonia accurately. By leveraging advanced machine learning algorithms, the model achieves accurate X-ray image classification for the chest. Conclusion: This study concludes by presenting a model for detecting pneumonia by examining chest X-ray pictures. To accurately classify infected and non-infected lungs, the proposed model makes use of image dispensation methods and machine learning algorithms. The model's high accuracy in pneumonia detection can significantly contribute to early diagnosis and treatment.

A NEW INSTRUMENT TO SUPPORT EXAMINATION OF DISTAL RADIUS ON WRIST RADIOGRAPH

Objectives: To construct an instrument that supports wrist position control according to a number of shooting angles in anteroposterior and lateral position during X-ray examination. Materials and methods: Term such as criteria for evaluating wrist X-ray films, current technical status, some concepts about posture have been scientifically researched. Based on those things to provide research directions and design tool samples. A male volunteer, 36 years old agree to participate in the study, X-rays were taken in an anteroposterior and lateral position for investigation. Results and Conclusion: Initial evaluation shows that this tool is easy to use and only takes a little time to get used to the adjustment operations. The benefit that the device brings is that it helps stabilize the wrist position during X-rays, helping to adjust the angles within the range of motion of a normal person, produces more accurate wrist X-ray results, without having to take multiple scans, saving effort, machine wear and tear costs, creating practical benefits for patients, medical staff, researchers and hospital.

What Is the Black Hole Spin in Cyg X-1?

We perform a detailed study of the black hole spin of Cyg X-1, using accurate broadband X-ray data obtained in the soft spectral state by simultaneous NICER and NuSTAR observations, supplemented at high energies by INTEGRAL data. We use the relativistic disk model kerrbb together with different models of the Comptonization high-energy tail and the relativistically broadened reflection features. Unlike most previous studies, we tie the spin parameters of the disk and relativistic broadening models, thus combining the continuum and reflection methods of spin determination. We also consider a likely increase of the disk color correction due to a partial support of the disk by large-scale magnetic fields. We find that such models yield a spin parameter of a*=0.87−0.03+0.04 if the disk inclination is allowed to be free, with i=39°−1+1 . Assuming i = 27.°5, as determined by optical studies of the binary, worsens the fit but leads to similar values of the spin, a*=0.90−0.01+0.01 . In addition, we consider the presence of a warm Comptonization layer on top of the disk, motivated by successful modeling of soft X-ray excesses in other sources with such a model. This dramatically lowers the spin, to a * ≲ 0.1, consistent with the spin measurements from black hole mergers. On the other hand, if the natal spin of Cyg X-1 was low but now a * ≈ 0.9, a period of effective supercritical accretion had to take place in the past. Such accretion could be facilitated by photon advection, as proposed for ultraluminous X-ray sources.

Tailoring the Mass Density of 3D Printing Materials for Accurate X-ray Imaging Simulation by Controlled Underfilling for Radiographic Phantoms.

Additive manufacturing and 3D printing allow for the design and rapid production of radiographic phantoms for X-ray imaging, including CT. These are used for numerous purposes, such as patient simulation, optimization of imaging procedures and dose levels, system evaluation and quality assurance. However, standard 3D printing polymers do not mimic X-ray attenuation properties of tissues like soft, adipose, lung or bone tissue, and standard materials like liquid water. The mass density of printing polymers-especially important in CT-is often inappropriate, i.e., mostly too high. Different methods can be applied to reduce mass density. This work examines reducing density by controlled underfilling either realized by using 3D printing materials expanded through foaming during heating in the printing process, or reducing polymer flow to introduce microscopic air-filled voids. The achievable density reduction depends on the base polymer used. When using foaming materials, density is controlled by the extrusion temperature, and ranges from 33 to 47% of the base polymer used, corresponding to a range of -650 to -394 HU in CT with 120 kV. Standard filaments (Nylon, modified PLA and modified ABS) allowed density reductions by 20 to 25%, covering HU values in CT from -260 to 77 (Nylon), -230 to -20 (ABS) and -81 to 143 (PLA). A standard chalk-filled PLA filament allowed reproduction of bone tissue in a wide range of bone mineral content resulting in CT numbers from 57 to 460 HU. Controlled underfilling allowed the production of radiographic phantom materials with continuously adjustable attenuation in a limited but appropriate range, allowing for the reproduction of X-ray attenuation properties of water, adipose, soft, lung, and bone tissue in an accurate, predictable and reproducible manner.

Scanning electron microscopy analysis of non-metallic inclusions for making steels with superior mechanical properties

The objective of this novel research was to further understand the type, size and areal distribution of the non-metallic inclusions in bearing steels by applying state-of-the-art scanning electron microscopy (SEM) analysis in order to develop steels with superior properties. Ten samples were taken from bearing steel grade 52100 and analysed by SEM. It was concluded that SEM allows deeper understanding regarding the type, morphology, size, and distribution of the non-metallic inclusions by covering larger sample areas and accurate automatic energy dispersive X-ray microanalysis of the inclusions. While most of the non-metallic inclusions had an equivalent circle diameter (ECD) smaller than 20 µm, the main type of inclusions with ECDs larger than 20 µm were calcium aluminates. As the higher sulphur will tend to create sulphide type of inclusions and the lower sulphur will shift the inclusions towards calcium aluminate oxide types, it is recommended to keep the sulphur content at amounts above threshold that calcium aluminates appear.

Decoding Pneumonia: Leveraging CNNS for Accurate Chest X-Ray Classification

Pneumonia is a known potentially fatal lung disease that is frequently referred to as a silent killer since it can lead to lung alveoli filling with pus or fluid, mainly from fungal, viral, or bacterial infections. Chest X-rays are the primary diagnostic tool for pneumonia; however, the diagnosis becomes more complex when other pulmonary disorders such volume loss, haemorrhage, lung cancer, fluid overload, and consequences from radiation or surgery are taken into account. As a result, the interpretation of chest X-rays becomes complex, which makes the development of computer-aided diagnosis systems necessary to help physicians make decisions that are more accurate. In order to diagnose pneumonia from chest X-ray pictures, the research reported here uses a convolutional neural network (CNN) enhanced with a self-attention mechanism. 'Normal' and 'pneumonia' classes are included in the dataset used in the study methodology, and data augmentation techniques are applied to improve the model's resilience. By means of extensive evaluation metrics and visualizations, the study highlights the potential of the suggested model as a useful instrument to aid clinicians in diagnosing pneumonia, consequently reducing the difficulties linked to the interpretation of chest X-rays in the context of various pulmonary conditions.

A blurring correction method suitable to analyze quantitative x-ray images derived from energy-resolving photon counting detector

Objective. The purpose of this study is to propose a novel blurring correction method that enables accurate quantitative analysis of the object edge when using energy-resolving photon counting detectors (ERPCDs). Although the ERPCDs have the ability to generate various quantitative analysis techniques, such as the derivations of effective atomic number (Z eff) and bone mineral density values, at the object edge in these quantitative images, accurate quantitative information cannot be obtained. This is because image blurring prevents the gathering of accurate primary x-ray attenuation information. Approach. We developed the following procedure for blurring correction. A 5 × 5 pixels masking region was set as the processing area, and the pixels affected by blurring were extracted from the analysis of pixel value distribution. The blurred pixel values were then corrected to the proper values estimated by analyzing minimum and/or maximum values in the set mask area. The suitability of our correction method was verified by a simulation study and an experiment using a prototype ERPCD. Main results. When Z eff image of aluminum objects (Z eff = 13) were analyzed without applying our correction method, regardless of raw data or correction data applying a conventional edge enhancement method, the proper Z eff values could not be derived for the object edge. In contrast, when applying our correction method, 82% of pixels affected by blurring were corrected and the proper Z eff values were calculated for those pixels. As a result of investigating the applicability limits of our method through simulation, it was proven that it works effectively for objects with 4 × 4 pixels or more. Significance. Our method is effective in correcting image blurring when the quantitative image is calculated based on multiple images. It will become an in-demand technology for putting a quantitative diagnosis into actual medical examinations.

Generative deep learning furthers the understanding of local distributions of fat and muscle on body shape and health using 3D surface scans

BackgroundBody shape, an intuitive health indicator, is deterministically driven by body composition. We developed and validated a deep learning model that generates accurate dual-energy X-ray absorptiometry (DXA) scans from three-dimensional optical body scans (3DO), enabling compositional analysis of the whole body and specified subregions. Previous works on generative medical imaging models lack quantitative validation and only report quality metrics.MethodsOur model was self-supervised pretrained on two large clinical DXA datasets and fine-tuned using the Shape Up! Adults study dataset. Model-predicted scans from a holdout test set were evaluated using clinical commercial DXA software for compositional accuracy.ResultsPredicted DXA scans achieve R2 of 0.73, 0.89, and 0.99 and RMSEs of 5.32, 6.56, and 4.15 kg for total fat mass (FM), fat-free mass (FFM), and total mass, respectively. Custom subregion analysis results in R2s of 0.70–0.89 for left and right thigh composition. We demonstrate the ability of models to produce quantitatively accurate visualizations of soft tissue and bone, confirming a strong relationship between body shape and composition.ConclusionsThis work highlights the potential of generative models in medical imaging and reinforces the importance of quantitative validation for assessing their clinical utility.

Neutrosophic Fuzzy Logic-Based Hybrid CNN- LSTM for Accurate Chest X-ray Classification in COVID-19 PredictionNeutrosophic Fuzzy Logic-Based Hybrid CNN- LSTM for Accurate Chest X-ray Classification in COVID-19 Prediction

Neutrosophic Fuzzy Logic-Based Hybrid CNN- LSTM for Accurate Chest X-ray Classification in COVID-19 PredictionNeutrosophic Fuzzy Logic-Based Hybrid CNN- LSTM for Accurate Chest X-ray Classification in COVID-19 Prediction

A Markedly Expanded Sample of Candidate X-Ray-emitting Isolated Neutron Stars

The current sample of 12 radio-quiet isolated neutron stars that emit strongly in X-rays (XINSs) is both small and heterogeneous, limiting its usefulness for understanding the physics of neutron star atmospheres and cooling rates and for constraining the equation of state of neutron degenerate matter. Utilizing the ROSAT 1RXS and 2RXS data sets, in conjunction with the Sloan Digital Sky Survey Data Release 17 and other companion multiwavelength surveys, we have extended previous searches for blank-field X-ray source candidate XINSs, ultimately recovering two known XINSs while identifying 46 new, unstudied candidate fields devoid of likely multiwavelength counterparts. In this publication, we describe our selection approach and provide detailed information regarding our sample of new candidate XINSs. Future opportunities to verify or to refute these X-ray sources as isolated neutron stars by obtaining more accurate X-ray source positions, quality X-ray spectra, or deeper optical imaging are also discussed.

Unveiling the mechanisms of coating formation during micro-arc oxidation of titanium in Na2HPO4 electrolyte

In the present work, the effect of the electric potential provided by the bipolar pulsed power supply during micro-arc oxidation of titanium in disodium hydrogen phosphate (Na2HPO4) electrolyte on the microstructure and chemical composition of the formed oxide coating was studied. The surface topography and microstructure of obtained coatings were investigated using scanning and transmission electron microscopy. In addition, the chemical analysis was performed with two complementary methods: energy-dispersive X-ray spectroscopy and more accurate but surface-sensitive X-ray photoelectron spectroscopy. These investigations revealed the formation mechanism of the titanium oxide coating, depending on whether an anodic or cathodic cycle is forced by the power supply. During the anodic cycle, the simultaneous interaction of negatively charged O2− ions and HPO42− anions with Ti4+ ions released from the substrate occurs. Introducing the cathodic cycle allows a small amount of Na to be incorporated into the coating material, mainly in the areas close to the porosity, which was not possible with the use of the direct current (DC) or unipolar pulsed power supply. Moreover, an increase in applied voltage increased rutile content and improved the crystallinity of the TiO2-based MAO coating.

Comparing petrophysical properties and pore network characteristics of carbonate reservoir rocks using micro X-ray tomography imaging and microfacies analyses

Carbonate rocks have complex pore structures as a result of sedimentological and diagenetic processes. We investigated the pore network of four carbonate rock samples originating from an oil well in Hungary. Two samples were from the productive part of the Sarmatian limestone, and the other two were from the dry interval. We employed X-ray computed tomography (micro-XCT) in combination with lab measurements and microfacies analysis. To achieve accurate X-ray image segmentation we investigated and compared several segmentation techniques, including entropy and clustering. We then employed the Naïve Bayes classifier and tenfold cross-validation to assess the accuracy of our results. Our study achieved high accuracy with the type-2 fuzzy entropy technique on various metrics, including precision and recall. The comparison between the measured helium porosity and image-derived porosity showed a close match. Micro-XCT measurements revealed connected pore structure in the productive interval, compared to isolated pores in the dry interval. Microfacies analyses indicated that both the rocks of the productive and dry intervals were deposited in a marine environment, but different diagenetic processes altered the sediment into productive and dry rocks, respectively. The diagenesis of the productive rocks occurred in a meteoric phreatic environment, where secondary pores were created (porosity enhancement). In contrast, the dry rocks underwent meteoric vadose and marine diagenesis, where all previously created pores were filled with cement (porosity destruction). Our study highlights the importance of accurately characterizing the pore network of carbonate rocks, which can aid in understanding reservoir properties and predicting fluid flow behavior.

Growth and Structure of Rare-Earth Molybdate Crystals Na0.65La4.35Mo3O15.81±δF0.07±ε

Rare-earth sodium- and fluorine-substituted molybdates with the Na0.65La4.35Mo3O15.81±δF0.07±ε composition were synthesized for the first time as single crystals. An accurate X-ray diffraction analysis of three samples at room temperature showed that the obtained crystals are isostructural to undoped cubic compounds of the Ln5Mo3O16+δ family (space group Pn3¯n). Sodium cations partially occupy one of the two lanthanum positions and center the more distorted LaO8 polyhedra, while fluorine anions occupy over-stoichiometric oxygen positions in the vast interstices (cavities) of the structure. The partial substitution of fluorine atoms for oxygen ones affects the transport characteristics of compounds by activating the oxygen atoms in the anionic subsystem due to the effect of supplementary electrostatic repulsion between anions of different types.

Accurate X-ray diffraction data required for proper evaluation of bond valence sums and global instability indexes: redetermination of the crystal structures of diamond-like Cu2CdSiS4 and Cu2HgSnS4 as a case study.

Our calculations of the global instability index (G) values for some diamond-like materials with the general formula I2-II-IV-VI4 have indicated that the structures may be unstable or incorrectly determined. To compute the G value of a given compound, the bond valence sums (BVSs) must first be calculated using a crystal structure. Two examples of compounds with high G values, based on data from the literature, are the wurtz-stannite-type dicopper cadmium silicon tetrasulfide (Cu2CdSiS4) and the stannite-type dicopper mercury tin tetrasulfide (Cu2HgSnS4), which were first reported in 1967 and 1965, respectively. In the present study, Cu2CdSiS4 and Cu2HgSnS4 were prepared by solid-state synthesis at 1000 and 900 °C, respectively. The phase purity was assessed by powder X-ray diffraction. Optical diffuse reflectance UV/Vis/NIR spectroscopy was used to estimate the optical bandgaps of 2.52 and 0.83 eV for Cu2CdSiS4 and Cu2HgSnS4, respectively. The structures were solved and refined using single-crystal X-ray diffraction data. The structure type of Cu2CdSiS4 was confirmed, where Cd2+, Si4+ and two of the three crystallographically unique S2- ions lie on a mirror plane. The structure type of Cu2HgSnS4 was also verified, where all ions lie on special positions. The S2- ion resides on a mirror plane, the Cu+ ion is situated on a fourfold rotary inversion axis and both the Hg2+ and the Sn4+ ions are located on the intersection of a fourfold rotary inversion axis, a mirror plane and a twofold rotation axis. Using the crystal structures solved and refined here, the G values were reassessed and found to be in the range that indicates reasonable strain for a stable crystal structure. This work, together with some examples gathered from the literature, shows that accurate data collected on modern instrumentation should be used to reliably calculate BVSs and G values.

Attributed Abnormality Graph Embedding for Clinically Accurate X-Ray Report Generation.

Despite the recent success of deep learning models for text generation, generating clinically accurate reports remains challenging. More precisely modeling the relationships of the abnormalities revealed in an X-ray image has been found promising to enhance the clinical accuracy. In this paper, we first introduce a novel knowledge graph structure called an attributed abnormality graph (ATAG). It consists of interconnected abnormality nodes and attribute nodes for better capturing more fine-grained abnormality details. In contrast to the existing methods where the abnormality graph are constructed manually, we propose a methodology to automatically construct the fine-grained graph structure based on annotated X-ray reports and the RadLex radiology lexicon. We then learn the ATAG embeddings as part of a deep model with an encoder-decoder architecture for the report generation. In particular, graph attention networks are explored to encode the relationships among the abnormalities and their attributes. A hierarchical attention attention and a gating mechanism are specifically designed to further enhance the generation quality. We carry out extensive experiments based on the benchmark datasets, and show that the proposed ATAG-based deep model outperforms the SOTA methods by a large margin in ensuring the clinical accuracy of the generated reports.