搜索结果：Journal of X-ray science and technology

BackgroundNon-destructive testing (NDT) is crucial for the preservation and restoration of ancient wooden structures, with Computed Tomography (CT) increasingly utilized in this field. However, practical CT examinations of these structures-often characterized by complex configurations, large dimensions, and on-site constraints-frequently encounter difficulties in acquiring full-angle projection data. Consequently, images reconstructed under limited-angle conditions suffer from poor quality and severe artifacts, hindering accurate assessment of critical internal features such as mortise-tenon joints and incipient damage.ObjectiveThis study aims to develop a novel algorithm capable of achieving high-quality image reconstruction from incomplete, limited-angle projection data.MethodsWe propose CADRE (Contour-guided Alternating Direction Method of Multipliers-optimized Deep Radon Enhancement), an unsupervised deep learning reconstruction framework. CADRE innovatively integrates the ADMM optimization strategy, the learning paradigm of Deep Radon Prior (DRP) networks, and a geometric contour-guidance mechanism. This approach synergistically enhances reconstruction performance by iteratively optimizing network parameters and input images, without requiring large-scale paired training data, rendering it particularly suitable for cultural heritage applications.ResultsSystematic validation using both a digital dougong simulation model of the Yingxian Wooden Pagoda and a physical wooden dougong model from Foguang Temple demonstrates that, under typical 90° and 120° limited-angle conditions, the CADRE algorithm significantly outperforms traditional FBP, iterative reconstruction algorithms SART and ADMM-TV, and other representative unsupervised deep learning methods (Deep Image Prior, DIP; Residual Back-Projection with DIP, RBP-DIP; DRP). This superiority is evident in quantitative metrics such as PSNR and SSIM, as well as in visual quality, including artifact suppression and preservation of structural details. CADRE exhibits exceptional capability in accurately reproducing internal mortise-tenon configurations and fine features within ancient timber.ConclusionThe CADRE algorithm provides a robust and efficient solution for limited-angle CT image reconstruction of ancient wooden structures. It effectively overcomes the limitations of existing methods in handling incomplete data, significantly enhances the quality of reconstructed images and the characterization of internal fine structures, and offers strong technical support for the scientific understanding, condition assessment, and precise conservation of cultural heritage, thereby holding substantial academic value and promising application prospects.

ObjectivesTo evaluate the application of different tube voltages and image-reconstruction algorithms in paranasal-sinus computed tomography (CT) and optimizes the scanning protocols for paranasal-sinus CT while balancing between image quality and radiation dose.MethodsNinety patients were randomly divided into three groups (A, B, and C). Group A used conventional scanning parameters: tube voltage of 120 kVp, tube current uDose level 1, and the Karl iterative reconstruction algorithm. Groups B and C used tube voltages of 100 and 80 kVp, respectively, and tube current uDose level 1. The Karl iterative reconstruction algorithm and artificial intelligence iterative reconstruction (AIIR) algorithm were used. Optimal image reconstruction noise levels were selected for each group, and the image quality and radiation doses of the best images were statistically analyzed.ResultsThe best image reconstruction noise levels for Groups A, B, and C were Karl level 5, AIIR level 5, and AIIR level 4, respectively. The signal-to-noise ratio, contrast-to-noise ratio, figure of merit, and subjective score values of the images in Groups B (AIIR level 5) and C (AIIR level 4) were higher than those in Group A (Karl level 5). The patients from Groups B and C had the CT dose-index volume, dose-length product, and size-specific dose estimate based on the water-equivalent diameter that were 68.86%, 71.76%, 69.84%, 84.39%, 85.95%, and 85.50% lower, respectively, than those of Group A (P&#x2009;<&#x2009;0.001).ConclusionsA low tube voltage combined with the AIIR algorithm effectively improves image quality and decreases the radiation doses for patients undergoing paranasal-sinus CT. The optimal parameters for paranasal-sinus CT are 80 kVp, uDose level 1, and AIIR level 4.

Ghost imaging is an imaging technique that achieves image reconstruction by measuring the intensity correlation function between the reference arm and the object arm. In parallel ghost imaging, each pixel of a position-sensitive detector is further regarded as a bucket detector, enabling the parallel acquisition of hundreds or thousands of ghost imaging subsystems in a single measurement, thus realizing high-resolution imaging with extremely low measurement counts. Relying on synchrotron radiation, we have achieved X-ray parallel ghost imaging with high pixel resolution, low dose, and ultra-large field of view. However, the dependence of X-ray parallel ghost imaging on synchrotron radiation has set extremely high thresholds for the dissemination and application of this technology. In this work, we broke away from synchrotron radiation facility and completed the pipeline-style acquisition of parallel ghost imaging using rough and inexpensive equipment in the most reproducible way for others. Eventually, we achieved ghost imaging with an effective pixel size of 8.03 &#x3bc;m, an image size of 2880 &#xd7; 2280, and a minimum of 10 measurement numbers (a sampling rate of 0.62%) using a laboratory X-ray light source. It can be achieved merely by making minor modifications to any industrial CT device. With a total experimental cost of only $40, this work demonstrates great universality. We have put forward a comprehensive framework for the practical application of parallel ghost imaging, which is an essential prerequisite for the generalization of parallel ghost imaging to enter the commercial and practical arenas.

Cancer remains a leading cause of mortality, where early detection significantly improves survival rates. Advances in technology have enabled automated cancer detection using medical imaging and microarray gene expression data. However, these datasets often contain redundant or noisy features that hinder classification performance. Feature selection is key preprocessing step to enhance accuracy and reduce computational costs. In cancer-related medical research, optimizing deep learning architectures is crucial for better classification outcomes. Metaheuristic algorithms have been popular for tackling both feature selection and deep neural networks (DNN) optimization. This survey reviews 91 peer-reviewed articles (2012-2025) on metaheuristics for feature selection and DNN optimization in cancer classification using medical images and microarray data. Literature was sourced from databases such as Google Scholar, IEEE Xplore, Elsevier, ResearchGate, Springer, MDPI, and ScienceDirect. Our findings indicate that k-Nearest Neighbors (kNN), Support Vector Machines (SVM), and Convolutional Neural Networks (CNN) are the most widely adopted classifiers, used in 23%, 21%, and 18% of cases, respectively. Among metaheuristics, Particle Swarm Optimization (PSO), Genetic Algorithm (GA), and Ant Colony Optimization (ACO) dominate the landscape, appearing in 13%, 11%, and 10% of studies. We also review 39 image-based and 44 microarray cancer datasets. This survey identifies critical gaps in current research and proposes several future directions to enhance model robustness and classification accuracy. Through a detailed comparative analysis, this study provides valuable insights for researchers and decision-makers, highlighting the need for continued innovation in computational methods for cancer detection and diagnosis.

X-ray imaging technology, as the core non-invasive inspection method, plays an irreplaceable role in industrial non-destructive testing and medical diagnosis. However, during signal acquisition, the imaging system faces multiple interferences, such as the quantum effect and electronic noise. This leads to a significant decrease in the image's signal-to-noise ratio, seriously affecting the accuracy of hazardous material identification and lesion detection. Existing X-ray image denoising methods have two major limitations. First, in physical model-driven denoising methods, the existing noise models deviate significantly from realistic ones, resulting in poor denoising results. Second, in mainstream deep learning-based methods, Convolutional Neural Networks (CNNs) have limitations in capturing long-range dependencies, while the Transformer model with a global receptive field has high computational complexity. To address these challenges, a physically grounded noise model is designed for synthesizing realistic X-ray images, trained on the public mainstream X-ray image security inspection datasets and augmented with hybrid real-synthetic data. Based on this, a novel denoising model, XDenoiser, is proposed in this paper. It incorporates a linear attention complexity Receptance Weighted Key-Value (RWKV) into a Transformer-based image restoration structure and combines it with CNNs to support both global and local receptive fields. Experiments on the expanded mainstream X-ray image security inspection datasets demonstrate the reasonableness and effectiveness of the XDenoiser algorithm.

X-ray multimodal imaging, which extracts absorption, refraction, and scattering signals simultaneously, holds significant potential in biomedical and materials science applications. However, laboratory-based X-ray multimodal imaging remains underdeveloped, with existing techniques constrained by system magnification and detector pixel size. This study employs a single-mask edge illumination (SM EI) configuration and establishes the corresponding single-mask illumination curve (SM IC). Using Geant4 simulations, we validate the feasibility of retrieving all three signals under conventional magnification and large-pixel detectors. Results show accurate extraction of both refraction and scattering signals, with model fitting close to unity. We further explore the impact of key system parameters, including focal spot size, tube voltage, mask thickness, duty cycle, pixel count, and detector operation mode on imaging performance. The simulations reveal that small focal spots and low-energy X-rays enhance contrast, thick masks maintain signal quality at high energy, and low duty cycles and high photon counts improve the contrast-to-noise ratio (CNR). Additionally, the charge summing mode increases refraction CNR by approximately three times compared to standard modes. These findings demonstrate the effectiveness of the SM EI method, enhancing spatial resolution and providing optimization insights for designing laboratory-based X-ray multimodal imaging systems.

BackgroundOut-of-plane artifacts in digital breast tomosynthesis (DBT) can affect image quality, even subtly, and are influenced by the size and z-position of features with contrast of clinical images.ObjectiveTo propose a phantom and metric to further characterize out-of-plane artifacts in DBT.MethodsPhantoms with a signal inserted were manufactured, and the reconstructed planes were obtained using the DBT system. Normalized maximum contrast within the plane area was used to quantitatively evaluate out-of-plane artifacts. The spread of out-of-plane artifacts within the reconstructed plane was qualitatively evaluated by observing the profile within the plane area.ResultsThe larger the signal diameter, the stronger the effect of out-of-plane artifacts on the z-position far from the in-focus plane. When the z-position of the signal was on the upper side of the z-position of the center of X-ray tube rotation, out-of-plane artifacts were stronger on the upper side and weaker on the lower side of the signal. The spread of out-of-plane artifacts in the off-focus plane changed from monomodal to bimodal, with movement away from the signal's location in the z-direction.ConclusionsThis work proposes new phantoms and analysis methods to investigate the characteristics of out-of-plane artifacts, supplementing conventional methods.

Accurate X-ray Computed tomography (CT) image segmentation of the abdominal organs is fundamental for diagnosing abdominal diseases, planning cancer treatment, and formulating radiotherapy strategies. However, the existing deep learning based models for three-dimensional (3D) CT image abdominal multi-organ segmentation face challenges, including complex organ distribution, scarcity of labeled data, and diversity of organ structures, leading to difficulties in model training and convergence and low segmentation accuracy. To address these issues, a novel multi-stage training and a deep supervision model based segmentation approach is proposed. It primary integrates multi-stage training, pseudo- labeling technique, and a developed deep supervision model with attention mechanism (DLAU-Net), specifically designed for 3D abdominal multi-organ segmentation. The DLAU-Net enhances segmentation performance and model adaptability through an improved network architecture. The multi-stage training strategy accelerates model convergence and enhances generalizability, effectively addressing the diversity of abdominal organ structures. The introduction of pseudo-labeling training alleviates the bottleneck of labeled data scarcity and further improves the model's generalization performance and training efficiency. Experiments were conducted on a large dataset provided by the FLARE 2023 Challenge. Comprehensive ablation studies and comparative experiments were conducted to validate the effectiveness of the proposed method. Our method achieves an average organ accuracy (AVG) of 90.5% and a Dice Similarity Coefficient (DSC) of 89.05% and exhibits exceptional performance in terms of training speed and handling data diversity, particularly in the segmentation tasks of critical abdominal organs such as the liver, spleen, and kidneys, significantly outperforming existing comparative methods.

In comparison to conventional medical computed tomography (CT), cone-beam CT (CBCT) has become widely used in dental and maxillofacial applications due to its accurate 3D information, high resolution, minimal radiation dose, and affordable machine cost. In this study, we investigated the image quality and radiation doses of dental CBCT and X-ray machines developed in Thailand. Our in-house reconstruction algorithm including artifact reduction was based on GPU calculations of filtered backprojection and was significantly faster than a CPU-based algorithm. The image quality aspects for CBCT were evaluated in terms of high contrast resolution, gray value uniformity, noise, and geometric accuracy, while image quality assessment for 2D images included high contrast resolution, low contrast levels, and distortion rate. Radiation doses were measured and calculated for the dose-area product (DAP). The technical image quality and radiation dose assessment was compared with those of other commercial extraoral imaging machines. The findings demonstrate that, when compared to other units, the proposed 2D and 3D extraoral imaging systems yielded comparable technical image quality and radiation doses. Based on these results, the Thai-made 2D and 3D extraoral imaging machines appear suitable for further clinical evaluation.

Accurate segmentation of stenosis in X-ray angiography (XRA) images is crucial for the objective assessment of stenosis severity and subsequent treatment planning in coronary artery disease. Current clinical practice primarily relies on subjective visual evaluation, which suffers from significant inter-observer variability. In this work, we propose a deep learning model enhanced with a novel Hybrid Context-Aware Attention (HCA) module. HCA employs a parallel dual-pathway design that integrates global inter-channel attention and grouped multi-scale spatial aggregation. This integration enhances feature discriminability and spatial-context modeling, leading to more accurate and anatomically consistent stenosis segmentation in XRA. Evaluated on three independent datasets, our method achieves competitive performance against existing approaches across multiple metrics, demonstrating consistent leading performance. Ablation and attention visualization studies further confirm the contribution of the designed module to reducing segmentation errors and enhancing focus on stenotic regions. These findings demonstrate that the proposed model is an effective and generalizable approach for stenosis segmentation in XRA, with the potential to support standardized assessment in clinical practice.

Recently, semi-supervised learning has demonstrated significant potential in the field of medical image segmentation. However, the majority of the methods fail to establish connections among diverse sample data. Moreover, segmentation networks that utilize fixed parameters can impede model training and even amplify the risk of overfitting. To address these challenges, this paper proposes an adversarial consistency-based semi-supervised segmentation method, leveraging a dual multiscale mean teacher model. First, by designing a discriminator network with adaptive feature selection and training it alternately with the segmentation network, the method enhances the segmentation network's ability to transfer knowledge from the limited labeled data to the unlabeled data. The discriminator evaluates the quality of the segmentation network's results for both labeled and unlabeled data, while simultaneously guiding the network to learn consistency in segmentation performance throughout the training process. Second, we design a Triple-attention dynamic convolutional (TADC) module, which allows the convolution kernel parameters to be adjusted flexibly according to different input data. This improves the feature representation capability of the network model and helps reduce the risk of overfitting. Finally, we propose a novel feature selection and fusion module (FSFM) within the segmentation network, which dynamically selects and integrates important features to enhance the saliency of key information, improving the overall performance of the model. The proposed adversarial consistency-based semi-supervised segmentation method is applied to the MosMedData dataset. The results demonstrate that the segmentation network outperforms the baseline model, achieving improvements of 3.83%, 3.97%, 3.14% in terms of Dice, Jaccard, and NSD scores, respectively, for the segmentation of pneumonia lesions. The proposed segmentation method outperforms state-of-the-art segmentation networks and demonstrates superior potential for segmenting pneumonia lesions, as evidenced by extensive experiments conducted on the MosMedData and COVID-19-P20 datasets.

ObjectiveThis study aimed to develop a robust framework for breast cancer diagnosis by integrating advanced segmentation and classification approaches. Transformer-based and U-Net segmentation models were combined with radiomic feature extraction and machine learning classifiers to improve segmentation precision and classification accuracy in mammographic images.Materials and MethodsA multi-center dataset of 8000 mammograms (4200 normal, 3800 abnormal) was used. Segmentation was performed using Transformer-based and U-Net models, evaluated through Dice Coefficient (DSC), Intersection over Union (IoU), Hausdorff Distance (HD95), and Pixel-Wise Accuracy. Radiomic features were extracted from segmented masks, with Recursive Feature Elimination (RFE) and Analysis of Variance (ANOVA) employed to select significant features. Classifiers including Logistic Regression, XGBoost, CatBoost, and a Stacking Ensemble model were applied to classify tumors into benign or malignant. Classification performance was assessed using accuracy, sensitivity, F1 score, and AUC-ROC. SHAP analysis validated feature importance, and Q-value heatmaps evaluated statistical significance.ResultsThe Transformer-based model achieved superior segmentation results with DSC (0.94&#x2009;&#xb1;&#x2009;0.01 training, 0.92&#x2009;&#xb1;&#x2009;0.02 test), IoU (0.91&#x2009;&#xb1;&#x2009;0.01 training, 0.89&#x2009;&#xb1;&#x2009;0.02 test), HD95 (3.0&#x2009;&#xb1;&#x2009;0.3 mm training, 3.3&#x2009;&#xb1;&#x2009;0.4 mm test), and Pixel-Wise Accuracy (0.96&#x2009;&#xb1;&#x2009;0.01 training, 0.94&#x2009;&#xb1;&#x2009;0.02 test), consistently outperforming U-Net across all metrics. For classification, Transformer-segmented features with the Stacking Ensemble achieved the highest test results: 93% accuracy, 92% sensitivity, 93% F1 score, and 95% AUC. U-Net-segmented features achieved lower metrics, with the best test accuracy at 84%. SHAP analysis confirmed the importance of features like Gray-Level Non-Uniformity and Zone Entropy.ConclusionThis study demonstrates the superiority of Transformer-based segmentation integrated with radiomic feature selection and robust classification models. The framework provides a precise and interpretable solution for breast cancer diagnosis, with potential for scalability to 3D imaging and multimodal datasets.

BackgroundGlaucoma is a leading cause of irreversible vision loss and is characterized by subtle structural changes in the optic disc and optic cup. However, existing automated detection systems often suffer from weak boundary delineation, dataset variability, and unstable feature learning, which limit their generalizability and clinical reliability.ObjectiveThis study aims to develop a unified and anatomically guided framework for accurate and reliable automated glaucoma detection from fundus images.MethodsThe proposed pipeline begins with contrast-enhanced preprocessing to improve image quality, followed by an Attention-guided Multi-scale Edge-aware Segmentation Network (AME-SegNet) for precise segmentation of the optic disc and optic cup. Both deep convolutional features and clinically relevant geometric features are extracted and optimized using Bitterling Colony Optimization (BCO) to select the most discriminative attributes. A Convolutional Transformer (CT) is then employed to integrate local convolutional representations with global attention mechanisms for robust classification. Additionally, the Honey Badger Algorithm (HBA) is used for automatic parameter tuning to ensure stable convergence.ResultsExperimental evaluation demonstrates high segmentation performance with Dice scores of 97.36% for the optic disc and 96.72% for the optic cup on the Drishti-GS1 dataset. The classification model achieves accuracies of 98.63% on RIM-ONE and 98.96% on ORIGA-Light datasets, indicating strong generalization capability.ConclusionsThe proposed framework exhibits robust performance, high accuracy, and strong generalization across multiple datasets. These results highlight its effectiveness and clinical potential for reliable automated glaucoma screening and early diagnosis.

ObjectiveDetecting and accurately diagnosing rib fractures in chest radiographs is a challenging and time-consuming task for radiologists. This study presents a novel deep learning system designed to automate the detection and segmentation of rib fractures in chest radiographs.MethodsThe proposed method combines CenterNet with HRNet v2 for precise fracture region identification and HRNet-W48 with contextual representation to enhance rib segmentation. A dataset consisting of 1006 chest radiographs from a tertiary hospital in Korea was used, with a split of 7:2:1 for training, validation, and testing.ResultsThe rib fracture detection component achieved a sensitivity of 0.7171, indicating its effectiveness in identifying fractures. Additionally, the rib segmentation performance was measured by a dice score of 0.86, demonstrating its accuracy in delineating rib structures. Visual assessment results further highlight the model's capability to pinpoint fractures and segment ribs accurately.ConclusionThis innovative approach holds promise for improving rib fracture detection and rib segmentation, offering potential benefits in clinical practice for more efficient and accurate diagnosis in the field of medical image analysis.

BackgroundIt is common for X-ray computed tomography (CT) images to be reconstructed differently for various clinical examination purposes. This is primarily because we aim to meet two clinical requirements: improving spatial resolution and reducing noise by changing the reconstruction parameters.ObjectiveTwo deep learning-based methods, super resolution (SR) and denoising, respectively, have been proposed to address these requests. We present a single neural network that can perform SR and denoising simultaneously.MethodsWe propose using existing ultra-high-resolution CT (UHR-CT) data to achieve high spatial resolution and reduces noise. We generated specific input data, which is normal-resolution and high-noise data, simulated from UHR-CT data. Afterwards, we apply the network to NR-CT data, the resulting method, called SR-Denoise deep-learning reconstruction (DLR). In experiments, we measured modulation transfer function as the quantitative study. We also evaluated the performance using both simulated and real clinical data in NR-CT data, with UHR-CT data serving as the ground truth.ResultsSR-Denoise DLR achieved performance on both SR and denoising tasks that was equivalent to training them individually and outperformed the methods currently used in clinical settings.ConclusionsSR-Denoise DLR utilizes the spatial resolution of UHR-CT and takes advantage of NR-CT to significantly reduce noise.

BackgroundInterior tomography is a crucial technique in computed tomography (CT) that aims to minimize radiation exposure by limiting X-ray imaging to the region of interest (ROI) while maintaining diagnostic accuracy. However, traditional reconstruction algorithms often suffer from severe cupping artifacts caused by data truncation, which significantly degrades image quality.ObjectiveThis study aims to develop a parallel network that effectively integrates information between the projection and image domains to improve interior tomography reconstruction.MethodsIn this paper, we propose an end-to-end deep learning framework, the Two-Module Parallel Dual-Domain Network (TPDDN), which consists of two key modules. The Initial Restoration Module generates high-quality prior sinograms and images, providing a robust foundation for subsequent processing and effectively mitigating the impact of data truncation. The Interactive Fusion Module, the core of the network, employs two parallel and interactive branches that operate simultaneously on the projection and image domains. These branches enable bidirectional feature interaction and information fusion, significantly enhancing the accuracy and quality of the reconstructed images.ResultsExtensive experiments were conducted under both normal-dose and high-dose noise conditions to evaluate the performance of TPDDN. The results demonstrate that TPDDN achieves superior qualitative and quantitative performance compared to existing representative methods.ConclusionsThe proposed TPDDN offers a robust and effective approach for interior tomography reconstruction by synergistically integrating information from both the projection and image domains. It effectively suppresses cupping artifacts and enhances reconstructed image quality under both normal-dose and high-noise conditions, demonstrating promising potential for safer and more accurate diagnostic imaging.

Introduction and HypothesisAccurate segmentation of cesarean scar disorder (CSDi) in ultrasound images is crucial for clinical diagnosis, disease monitoring, and personalized treatment. However, the ambiguous boundaries and complex anatomical structures of CSDi pose significant challenges. To address this, we propose UMamba-Dual, a dual-branch model derived from UMamba, designed to enhance segmentation performance in CSDi regions and provide reliable imaging support for clinical decision-making.MethodsUMamba-Dual integrates the strengths of two enhanced branches: Dual-Bot, incorporating squeeze-and-excitation (SE) attention, and Dual-Enc, employing a feature pyramid network (FPN) for improved feature representation and multi-scale perception. The training dataset included 1200 augmented 2D ultrasound images from 300 originals via flipping and rotation. An independent test set of 32 images was randomly selected and excluded from training and validation. Model performance was evaluated using Dice similarity coefficient, Intersection over Union (IoU), and Normalized Surface Dice (NSD), and compared with classical segmentation models such as nnUNet and VM-UNet.ResultsUMamba-Dual achieved superior performance with Dice = 0.832, IoU = 0.782, and NSD = 0.788, consistently outperforming both classical models (UNet, nnUNet) and the recent VM-UNet, as well as the internal baselines (UMamba_Bot, UMamba_Enc).ConclusionsUMamba-Dual enables more accurate and robust segmentation of CSDi regions in ultrasound images, particularly in cases characterized by ambiguous boundaries or irregular anatomical structures. These results highlight its potential for reliable clinical application.

BackgroundMeasuring an X-ray source's focal spot size is vital for Micro-CT resolution. Standard methods are often too complex or inaccurate. The popular JIMA resolution test card is simple to use but lacks a clear, quantitative formula to determine the actual focal spot size.ObjectiveThis study aims to create a reliable quantitative link between JIMA resolution and focal spot size using simulations and experiments.MethodsWe used Monte Carlo simulations and practical experiments to establish the relationship between JIMA resolution and focal spot size.ResultsWe found that the focal spot size is twice the line pair width on the JIMA card when the image contrast (MTF) is at 10%. This method is highly accurate, with a maximum measurement error of less than 8.7% compared to a high-precision technique.ConclusionsOur findings provide a simple, fast, and validated method for measuring focal spot size using the JIMA test card. This makes it a practical and reliable alternative to more complex procedures.

BackgroundPneumoconiosis is one of the most severe occupational diseases, and accurate staging is essential for treatment planning and disease management. However, the visual features on chest X-rays are often subtle and exhibit gradual transitions between stages, posing challenges for traditional classification models.ObjectiveThe study aims to overcome the limitations of current staging methods, and to develop a model that simultaneously captures the ordinal progression of pneumoconiosis and enhances feature discrimination for reliable staging.MethodsWe propose a Prototype-enhanced Contrastive Ordinal Regression Network (PCOR-Net) for pneumoconiosis staging. PCOR-Net adopts a dual-branch architecture, where a momentum-updated teacher encoder builds dynamic class prototypes, and a student encoder learns more discriminative features under prototype-guided supervision. To capture the ordinal structure of disease progression, we introduce an ordinal-aware prototype contrastive mechanism and a learnable-threshold ordinal regression module that adapts to the non-uniform nature of stage transitions. Three loss functions-prototype contrastive loss, feature distillation loss, and ordinal regression loss-are jointly optimized in a unified framework.ResultsWe conducted experiments on the pneumoconiosis dataset, where PCOR-Net achieved an accuracy of 91.18% and a Quadratic Weighted Kappa (QWK) of 92.72%, outperforming existing state-of-the-art methods. To assess generalizability, PCOR-Net was also evaluated on a COVID-19 severity dataset, demonstrating good transferability.ConclusionsPCOR-Net demonstrates strong effectiveness and robustness in pneumoconiosis staging and generalizes well to the COVID-19 grading dataset, providing reliable support for clinical diagnosis with improved accuracy and ordinal consistency.