Search Results (4,574)

Shipping containers are vital to the transportation industry due to their cost-effectiveness and compatibility with intermodal systems. With the significant increase in container usage since the mid-20th century, manual tracking at port terminals has become inefficient and prone to errors. Recent advancements in Deep Learning for object detection have introduced Computer Vision as a solution for automating this process. However, challenges such as low-quality images, varying font sizes & illumination, and environmental conditions hinder recognition accuracy. This study explores various architectures and proposes a Container Code Localization Network (CCLN), utilizing ResNet and UNet for code identification, and a Container Code Recognition Network (CCRN), which combines Convolutional Neural Networks with Long Short-Term Memory to convert the image text into a machine-readable format. By enhancing existing shipping container localization and recognition datasets with additional images, our models exhibited improved generalization capabilities on other datasets, such as Syntext, for text recognition. Experimental results demonstrate that our system achieves $97.93\%$ accuracy at $64.11$ frames per second under challenging conditions such as varying font sizes, illumination, tilt, and depth, effectively simulating real port terminal environments. The proposed solution promises to enhance workflow efficiency and productivity in container handling processes, making it highly applicable in modern port operations. Full article

Forage crops play a vital role in ensuring livestock productivity and food security in Northern Kazakhstan, a region characterized by highly variable weather conditions. However, traditional methods for assessing crop maturity remain time-consuming and labor-intensive, underscoring the need for automated monitoring solutions. Recent advances in remote sensing and artificial intelligence (AI) offer new opportunities to address this challenge. In this study, unmanned aerial vehicle (UAV)-based multispectral imaging was used to monitor the development of forage crops—pea, sudangrass, common vetch, oat—and their mixtures under field conditions in Northern Kazakhstan. A multispectral dataset consisting of five spectral bands was collected and processed to generate vegetation indices. Using a ResNet-based neural network model, the study achieved a high predictive accuracy ($R^2$ = 0.985) for estimating the continuous maturity index. The trained model was further integrated into a web-based platform to enable real-time visualization and analysis, providing a practical tool for automated crop maturity assessment and long-term agricultural monitoring. Full article

As a fundamental and indispensable component of modern electronic devices, the printed circuit board (PCB) has a complex structure and highly integrated functions, with its manufacturing quality directly affecting the stability and reliability of electronic products. However, during large-scale automated PCB production, its surfaces are prone to various defects and imperfections due to uncontrollable factors, such as diverse manufacturing processes, stringent machining precision requirements, and complex production environments, which not only compromise product functionality but also pose potential safety hazards. At present, PCB defect detection in industry still predominantly relies on manual visual inspection, the efficiency and accuracy of which fall short of the automation and intelligence demands in modern electronics manufacturing. To address this issue, in this paper, we have made improvements based on the classical Faster-RCNN object detection framework. Firstly, ResNet-101 is employed to replace the conventional VGG-16 backbone, thereby enhancing the ability to perceive small objects and complex texture features. Then, we extract features from images by using deformable convolution in the backbone network to improve the model’s adaptive modeling capability for deformed objects and irregular defect regions. Finally, the Convolutional Block Attention Module is incorporated into the backbone, leveraging joint spatial and channel attention mechanisms to improve the effectiveness and discriminative power of feature representations. The experimental results demonstrate that the improved model achieves a 4.5% increase in mean average precision compared with the original Faster-RCNN. Moreover, the proposed method exhibits superior detection accuracy, robustness, and adaptability compared with mainstream object detection models, indicating strong potential for engineering applications and industrial deployment. Full article

This paper presents a novel DEVA-ConvNeXt model to address challenges in beef freshness grading, including data collection difficulties, complex backgrounds, and model accuracy issues. The Alpha-Background Generation Shift (ABG-Shift) technology enables rapid generation of beef image datasets with complex backgrounds. By incorporating the Dynamic Non-Local Coordinate Attention (DNLC) and Enhanced Depthwise Convolution (EDW) modules, the model enhances feature extraction in complex environments. Additionally, Varifocal Loss (VFL) accelerates key feature learning, reducing training time and improving convergence speed. Experimental results show that DEVA-ConvNeXt outperforms models like ResNet101 and ShuffleNet V2 in terms of overall performance. Compared to the baseline model ConvNeXt, it achieves significant improvements in recognition Accuracy (94.8%, a 6.2% increase), Precision (94.8%, a 5.4% increase), Recall (94.6%, a 5.9% increase), and F1 score (94.7%, a 6.0% increase). Furthermore, real-world deployment and testing on embedded devices confirm the feasibility of this method in terms of accuracy and speed, providing valuable technical support for beef freshness grading and equipment design. Full article

Intelligent Fault Diagnosis (IFD) systems are integral to predictive maintenance and real-time monitoring but often encounter challenges such as data scarcity, non-linearity, and changing operational conditions. To address these challenges, we propose an enhanced transfer learning framework that integrates the Universal Adaptation Network (UAN) with Spectral-normalized Neural Gaussian Process (SNGP), WideResNet, and attention mechanisms, including self-attention and an outlier attention layer. UAN’s flexibility bridges diverse fault conditions, while SNGP’s robustness enables uncertainty quantification for more reliable diagnostics. WideResNet’s architectural depth captures complex fault patterns, and the attention mechanisms focus the diagnostic process. Additionally, we employ Optuna for hyperparameter optimization, using a structured study to fine-tune model parameters and ensure optimal performance. The proposed approach is evaluated on benchmark datasets, demonstrating superior fault identification accuracy, adaptability to varying operational conditions, and resilience against data anomalies compared to existing models. Our findings highlight the potential of advanced machine learning techniques in IFD, setting a new standard for applying these methods in complex diagnostic environments. Full article

Three-dimensional cell model systems such as tumour organoids allow for in vitro modelling of self-organized tissue with functional and histologic similarity to in vivo tissue. However, there is a need for standard protocols and techniques to confirm the presence of cancer within organoids derived from tumour tissue. The aim of this study was to assess the utility of a Nanostring gene expression-based machine learning classifier to determine the presence of cancer or normal organoids in cultures developed from both benign and cancerous stomach biopsies. A prospective cohort of normal and cancer stomach biopsies were collected from 2019 to 2022. Tissue specimens were processed for formalin-fixed paraffin-embedding (FFPE) and a subset of specimens were established in organoid cultures. Specimens were labelled as normal or cancer according to analysis of the FFPE tissue by two pathologists. The gene expression in FFPE and organoid tissue was measured using a 107 gene Nanostring codeset and normalized using the Removal of Unwanted Variation III algorithm. Our machine learning model was developed using five-fold nested cross-validation to classify normal or cancer gastric tissue from publicly available Asian Cancer Research Group (ACRG) gene expression data. The models were externally validated using the Cancer Genome Atlas (TCGA), as well as our own FFPE and organoid gene expression data. A total of 60 samples were collected, including 38 cancer FFPE specimens, 5 normal FFPE specimens, 12 cancer organoids, and 5 normal organoids. The optimal model design used a Least Absolute Shrinkage and Selection Operator model for feature selection and an ElasticNet model for classification, yielding area under the curve (AUC) values of 0.99 [95% CI: 0.99–1], 0.90 [95% CI: 0.87–0.93], and 0.79 [95% CI: 0.74–0.84] for ACRG (internal test), FFPE, and organoid (external test) data, respectively. The performance of our final model on external data achieved AUC values of 0.99 [95% CI: 0.98–1], 0.94 [95% CI: 0.86–1], and 0.85 [95% CI: 0.63–1] for TCGA, FFPE, and organoid specimens, respectively. Using a public database to create a machine learning model in combination with a Nanostring gene expression assay allows us to allocate organoids and their paired whole tissue samples. This platform yielded reasonable accuracy for FFPE and organoid specimens, with the former being more accurate. This study re-affirms that although organoids are a high-fidelity model, there are still limitations in validating the recapitulation of cancer in vitro. Full article

The relationship between retinal images and intraocular cytokine profiles remains largely unexplored, and no prior work has systematically compared fundus- and OCT-based deep learning models for cytokine prediction. We aimed to predict intraocular cytokine concentrations using color fundus photographs (CFP) and retinal optical coherence tomography (OCT) with deep learning. Our pipeline consisted of image preprocessing, convolutional neural network–based feature extraction, and regression modeling for each cytokine. Deep learning was implemented using AutoGluon, which automatically explored multiple architectures and converged on ResNet18, reflecting the small dataset size. Four approaches were tested: (1) CFP alone, (2) CFP plus demographic/clinical features, (3) OCT alone, and (4) OCT plus these features. Prediction performance was defined as the mean coefficient of determination (R²) across 34 cytokines, and differences were evaluated using paired two-tailed t-tests. We used data from 139 patients (152 eyes) and 176 aqueous humor samples. The cohort consisted of 85 males (61%) with a mean age of 73 (SD 9.8). Diseases included 64 exudative age-related macular degeneration, 29 brolucizumab-associated endophthalmitis, 19 cataract surgeries, 15 retinal vein occlusion, and 8 diabetic macular edema. Prediction performance was generally poor, with mean R² values below zero across all approaches. The CFP-only model (–0.19) outperformed CFP plus demographics (–24.1; p = 0.0373), and the OCT-only model (–0.18) outperformed OCT plus demographics (–14.7; p = 0.0080). No significant difference was observed between CFP and OCT (p = 0.9281). Notably, VEGF showed low predictability (31st with CFP, 12th with OCT). Full article

Background/Objectives: Major Depressive Disorder (MDD) is a severe psychiatric disorder, and effective, efficient automated diagnostic approaches are urgently needed. Traditional methods for assessing MDD face three key challenges: reliance on predefined features, inadequate handling of multi-site data heterogeneity, and suboptimal feature fusion. To address these issues, this study proposes the Multimodal Multitask Dynamic Disentanglement (MMDD) Framework. Methods: The MMDD Framework has three core innovations. First, it adopts a dual-pathway feature extraction architecture combining a 3D ResNet for modeling gray matter volume (GMV) data and an LSTM–Transformer for processing time series data. Second, it includes a Bidirectional Cross-Attention Fusion (BCAF) mechanism for dynamic feature alignment and complementary integration. Third, it uses a Gradient Reversal Layer-based Multitask Learning (GRL-MTL) strategy for enhancing the model’s domain generalization capability. Results: MMDD achieved 77.76% classification accuracy on the REST-meta-MDD dataset. Ablation studies confirmed that both the BCAF mechanism and GRL-MTL strategy played critical roles: the former optimized multimodal fusion, while the latter effectively mitigated site-related heterogeneity. Through interpretability analysis, we identified distinct neurobiological patterns: time series were primarily localized to subcortical hubs and the cerebellum, whereas GMV mainly involved higher-order cognitive and emotion-regulation cortices. Notably, the middle cingulate gyrus showed consistent abnormalities across both imaging modalities. Conclusions: This study makes two major contributions. First, we develop a robust and generalizable computational framework for objective MDD diagnosis by effectively leveraging multimodal data. Second, we provide data-driven insights into MDD’s distinct neuropathological processes, thereby advancing our understanding of the disorder. Full article

As technology advances, developers continually create innovative solutions to enhance smartphone security. However, the rapid spread of Android malware poses significant threats to devices and sensitive data. The Android Operating System (OS)’s open-source nature and Software Development Kit (SDK) availability mainly contribute to this alarming growth. Conventional malware detection methods, such as signature-based, static, and dynamic analysis, face challenges in detecting obfuscated techniques, including encryption, packing, and compression, in malware. Although developers have created several visualization techniques for malware detection using deep learning (DL), they often fail to accurately identify the critical malicious features of malware. This research introduces MalVis, a unified visualization framework that integrates entropy and N-gram analysis to emphasize meaningful structural and anomalous operational patterns within the malware bytecode. By addressing significant limitations of existing visualization methods, such as insufficient feature representation, limited interpretability, small dataset sizes, and restricted data access, MalVis delivers enhanced detection capabilities, particularly for obfuscated and previously unseen (zero-day) malware. The framework leverages the MalVis dataset introduced in this work, a publicly available large-scale dataset comprising more than 1.3 million visual representations in nine malware classes and one benign class. A comprehensive comparative evaluation was performed against existing state-of-the-art visualization techniques using leading convolutional neural network (CNN) architectures, MobileNet-V2, DenseNet201, ResNet50, VGG16, and Inception-V3. To further boost classification performance and mitigate overfitting, the outputs of these models were combined using eight distinct ensemble strategies. To address the issue of imbalanced class distribution in the multiclass dataset, we employed an undersampling technique to ensure balanced learning across all types of malware. MalVis achieved superior results, with 95% accuracy, 90% F1-score, 92% precision, 89% recall, 87% Matthews Correlation Coefficient (MCC), and 98% Receiver Operating Characteristic Area Under Curve (ROC-AUC). These findings highlight the effectiveness of MalVis in providing interpretable and accurate representation features for malware detection and classification, making it valuable for research and real-world security applications. Full article

Breast cancer is the second leading cause of cancer-related mortality among women. Early detection enables timely treatment, improving survival outcomes. This paper presents a comparative evaluation of federated learning (FL) frameworks for multiclass breast cancer classification using ultrasound images drawn from three datasets: BUSI, BUS-UCLM, and BCMID, which include 600, 38, and 323 patients, respectively. Five state-of-the-art networks were tested, with MobileNet, ResNet and InceptionNet identified as the most effective for FL deployment. Two aggregation strategies, FedAvg and FedProx, were assessed under varying levels of data heterogeneity in two and three client settings. Results from experiments indicate that the FL models outperformed local and centralized training, bypassing the adverse impacts of data isolation and domain shift. In the two-client federations, FL achieving up to 8% higher accuracy and almost 6% higher macro-F1 scores on average that local and centralized training. FedProx on MobileNet maintained a stable performance in the three-client federation with best average accuracy of 73.31%, and macro-F1 of 67.3% despite stronger heterogeneity. Consequently, these results suggest that the proposed multiclass model has the potential to support clinical workflows by assisting in automated risk stratification. If deployed, such a system could allow radiologists to prioritize high-risk patients more effectively. The findings emphasize the potential of federated learning as a scalable, privacy-preserving infrastructure for collaborative medical imaging and breast cancer diagnosis. Full article

Background/Objectives: Breast cancer diagnosis based on histopathological examination remains the most reliable and widely accepted approach in clinical practice, despite being time-consuming and prone to inter-observer variability. While deep learning methods have achieved high accuracy in medical image classification, their cross-site generalization remains limited due to differences in staining protocols and image acquisition. This study aims to evaluate and compare three clinically relevant adaptation strategies to improve model robustness under domain shift. Methods: The ResNet50V2 model, pretrained on ImageNet and further fine-tuned on the Kaggle Breast Histopathology Images dataset, was subsequently adapted to the BreaKHis dataset under three clinically relevant transfer strategies: (i) threshold calibration without retraining (site calibration), (ii) head-only fine-tuning (light FT), and (iii) full fine-tuning (full FT). Experiments were performed on an internal balanced dataset and on the public BreaKHis dataset using strict patient-level splitting to avoid data leakage. Evaluation metrics included accuracy, precision, recall, F1-score, ROC-AUC, and PR-AUC, computed per magnification level (40×, 100×, 200×, 400×). Results: Full fine-tuning consistently yielded the highest performance across all magnifications, reaching up to 0.983 ROC-AUC and 0.980 sensitivity at 400×. At 40× and 100×, the model correctly identified over 90% of malignant cases, with ROC-AUC values of 0.9500 and 0.9332, respectively. Head-only fine-tuning led to moderate gains (e.g., sensitivity up to 0.859 at 200×), while threshold calibration showed limited improvements (ROC-AUC ranging between 0.60–0.73). Grad-CAM analysis revealed more stable and focused attention maps after full fine-tuning, though they did not always align with diagnostically relevant regions. Conclusions: Our findings confirm that full fine-tuning is essential for robust cross-site deployment of histopathology AI systems, particularly at high magnifications. Lighter strategies such as threshold calibration or head-only fine-tuning may serve as practical alternatives in resource-constrained environments where retraining is not feasible. Full article

Early and accurate diagnosis of Alzheimer’s disease is critical for effective clinical intervention, particularly in distinguishing it from mild cognitive impairment, a prodromal stage marked by subtle structural changes. In this study, we propose a hybrid deep learning ensemble framework for Alzheimer’s disease classification using structural magnetic resonance imaging. Gray and white matter slices are used as inputs to three pretrained convolutional neural networks: ResNet50, NASNet, and MobileNet, each fine-tuned through an end-to-end process. To further enhance performance, we incorporate a stacked ensemble learning strategy with a meta-learner and weighted averaging to optimally combine the base models. Evaluated on the Alzheimer’s Disease Neuroimaging Initiative dataset, the proposed method achieves state-of-the-art accuracy of 99.21% for Alzheimer’s disease vs. mild cognitive impairment and 91.02% for mild cognitive impairment vs. normal controls, outperforming conventional transfer learning and baseline ensemble methods. To improve interpretability in image-based diagnostics, we integrate Explainable AI techniques by Gradient-weighted Class Activation Mapping, which generates heatmaps and attribution maps that highlight critical regions in gray and white matter slices, revealing structural biomarkers that influence model decisions. These results highlight the framework’s potential for robust and scalable clinical decision support in neurodegenerative disease diagnostics. Full article

This study presents an optimized Real-Time Detection Transformer (RT-DETRv2) deep learning model for the automated assessment of Tartary buckwheat germination and evaluates the influence of soaking and ultrasonic pretreatments on the germination ratio. Model optimization revealed that image chip size critically affected performance. The 512 × 512-pixel chip size was optimal, providing sufficient image context for detection and achieving a robust F1-score (0.9754 at 24 h, tested with a ResNet-101 backbone). In contrast, smaller chips (e.g., 128 × 128 pixels) caused severe performance degradation (24 h F1 = 0.3626 and 48 h F1 = 0.1211), which occurred because the 128 × 128 chip was too small to capture the entire object, particularly as the elongated and highly variable 48 h sprouts exceeded the chip dimensions. The optimized model, incorporating a ResNet-34 backbone, achieved a peak F1-score of 0.9958 for 24 h germination detection, demonstrating its robustness. The model was applied to assess germination dynamics, indicating that 24 h of treatment with 0.1% CaCl₂ and ultrasound enhanced total polyphenol accumulation (6.42 mg GAE/g). These results demonstrate that RT-DETRv2 enables accurate and efficient automated germination monitoring, providing a promising AI-assisted tool for seed quality evaluation and the optimization of agricultural pretreatments. Full article

Automated ultrasonic testing (AUT) serves as a vital method for evaluating critical infrastructure in industries such as oil and gas. However, a significant challenge in deploying artificial intelligence (AI)-based interpretation methods for AUT data lies in improving their reliability and effectiveness, particularly due to the inherent scarcity of real-world defective data. This study directly addresses data scarcity in a weld defect classification task, specifically concerning the detection of lack of fusion (LOF) defects in weld inspections using a proprietary industrial ultrasonic B-scan image dataset. This paper leverages state-of-the-art generative models, including Generative Adversarial Networks (GANs) and Denoising Diffusion Probabilistic Models (DDPM) (StyleGAN3, VQGAN with an unconditional transformer, and Stable Diffusion), to produce realistic B-scan images depicting LOF defects. The fine-tuned Transformer-based models, including ViT-Base, Swin-Tiny, and MobileViT-Small classifiers, on the regular B-scan image dataset are then applied to retain only high-confidence positive synthetic samples from each method. The impact of these synthetic images on the classification performance of a ResNet-50 model is evaluated, where it is fine-tuned with cumulative additions of synthetic images, ranging from 10 to 200 images. Its accuracy on the test set increases by 38.9% relative to the baseline with the addition of either 80 synthetic images using VQGAN with an unconditional transformer or 200 synthetic images by StyleGAN3 to the training set, and by 36.8% with the addition of 150 synthetic images by Stable Diffusion. This also outperforms Transformer-based vision models that are trained on regular training data. Concurrently, knowledge distillation experiments involve training ResNet-50 as a student model, leveraging the expertise of ViT-Base and Swin-Tiny as teacher models to demonstrate the effectiveness of adding the synthetic data to the training set, where the greatest enhancement is observed to be 34.7% relative to the baseline. This work contributes to advancing robust, AI-assisted tools for critical infrastructure inspection and offers practical pathways for enhancing available models in resource-constrained industrial environments. Full article

The internal quality assessment of potato tubers is a crucial task in agro-laboratory processing. Traditional methods struggle to detect internal defects such as hollow heart, internal bruises, and insect galleries using only surface features. We present a novel, fully modular hybrid AI architecture designed for defect detection using RGB images of potato slices, suitable for integration in laboratory. Our pipeline combines high-recall multi-threshold YOLO detection, contextual patch validation using ResNet, precise segmentation via the Segment Anything Model (SAM), and skin-contact analysis using VGG16 with a Random Forest classifier. Experimental results on a labeled dataset of over 6000 annotated instances show a recall above 95% and precision near 97.2% for most defect classes. The approach offers both robustness and interpretability, outperforming previous methods that rely on costly hyperspectral or MRI techniques. This system is scalable, explainable, and compatible with existing 2D imaging hardware. Full article