Search Results (134)

Background: Pneumonia in children poses a serious threat to life and health, making early detection critically important. In this regard, artificial intelligence methods can provide valuable support. Methods: Capsule networks and Bayesian optimization are modern techniques that were employed to build effective models for predicting pneumonia from chest X-ray images. The medical images underwent essential preprocessing, were divided into training, validation, and testing sets, and were subsequently used to develop the models. Results: The designed capsule neural network model with Bayesian optimization achieved the following final results: an accuracy of 95.1%, sensitivity of 98.9%, specificity of 85.4%, precision (PPV) of 94.8%, negative predictive value (NPV) of 96.2%, F1-score of 96.8%, and a Matthews correlation coefficient (MCC) of 0.877. In addition, the model was complemented with an explainability analysis using Grad-CAM, which demonstrated that its predictions rely predominantly on clinically relevant pulmonary regions. Conclusions: The proposed model demonstrates high accuracy and shows promise for potential use in clinical practice. It may also be applied to other tasks in medical image analysis. Full article

Automated quality control is critical in modern manufacturing, especially for metallic cast components, where fast and accurate surface defect detection is required. This study evaluates classical Machine Learning (ML) algorithms using extracted statistical parameters and deep learning (DL) architectures including ResNet50, Capsule Networks, and a 3D Convolutional Neural Network (CNN3D) using 3D image inputs. Using the Dataset Original, ML models with the selected parameters achieved high performance: RF reached 99.4 ± 0.2% precision and 99.4 ± 0.2% sensitivity, GB 96.0 ± 0.2% precision and 96.0 ± 0.2% sensitivity. ResNet50 trained with extracted parameters reached 98.0 ± 1.5% accuracy and 98.2 ± 1.7% F1-score. Capsule-based architectures achieved the best results, with ConvCapsuleLayer reaching 98.7 ± 0.2% accuracy and 100.0 ± 0.0% precision for the normal class, and 98.9 ± 0.2% F1-score for the affected class. CNN3D applied on 3D image inputs reached 88.61 ± 1.01% accuracy and 90.14 ± 0.95% F1-score. Using the Dataset Expanded with ML and PCA-selected features, Random Forest achieved 99.4 ± 0.2% precision and 99.4 ± 0.2% sensitivity, K-Nearest Neighbors 99.2 ± 0.0% precision and 99.2 ± 0.0% sensitivity, and SVM 99.2 ± 0.0% precision and 99.2 ± 0.0% sensitivity, demonstrating consistent high performance. All models were evaluated using repeated train-test splits to calculate averages of standard metrics (accuracy, precision, recall, F1-score), and processing times were measured, showing very low per-image execution times (as low as $3.69\times {10}^{-4}$ s/image), supporting potential real-time industrial application. These results indicate that combining statistical descriptors with ML and DL architectures provides a robust and scalable solution for automated, non-destructive surface defect detection, with high accuracy and reliability across both the original and expanded datasets. Full article

Pose estimation and localization within the gastrointestinal tract, particularly the small bowel, are crucial for invasive medical procedures. However, the task is challenging due to the complex anatomy, homogeneous textures, and limited distinguishable features. This study proposes a hybrid deep learning (DL) method combining Convolutional Neural Network (CNN)-based pose estimation and optical flow to address these challenges in a simulated small bowel environment. Initial pose estimation was used to assess the performance of simultaneous localization and mapping (SLAM) in such complex settings, using a custom endoscope prototype with a laser, micromotor, and miniaturized camera. The results showed limited feature detection and unreliable matches due to repetitive textures. To improve this issue, a hybrid CNN-based approach enhanced with Farneback optical flow was applied. Using consecutive images, three models were compared: Hybrid ResNet-50 with Farneback optical flow, ResNet-50, and NASNetLarge pretrained on ImageNet. The analysis showed that the hybrid model outperformed both ResNet-50 (0.39 cm) and NASNetLarge (1.46 cm), achieving the lowest RMSE of 0.03 cm, with feature-based SLAM failing to provide reliable results. The hybrid model also gained a competitive inference speed of 241.84 ms per frame, outperforming ResNet-50 (316.57 ms) and NASNetLarge (529.66 ms). To assess the impact of the optical flow choice, Lucas–Kanade was also implemented within the same framework and compared with the Farneback-based results. These results demonstrate that combining optical flow with ResNet-50 enhances pose estimation accuracy and stability, especially in textureless environments where traditional methods struggle. The proposed method offers a robust, real-time alternative to SLAM, with potential applications in clinical capsule endoscopy. The results are positioned as a proof-of-concept that highlights the feasibility and clinical potential of the proposed framework. Future work will extend the framework to real patient data and optimize for real-time hardware. Full article

Single-phase ground (SPG) faults pose significant challenges in three-phase four-wire distribution systems due to their complex transient characteristics and the presence of multiple influencing factors. To solve the aforementioned issues, a comprehensive fault identification framework is proposed, which uses the TabNet deep learning architecture with hyperparameters optimized by Optuna. Firstly, a 10 kV simulation model is developed in Simulink to generate a diverse fault dataset. For each simulated fault, voltage and current signals from eight channels (L1–L4 voltage and current) are collected. Secondly, multi-domain features are extracted from each channel across time, frequency, waveform, and wavelet perspectives. Then, an attention-based fusion mechanism is employed to capture cross-channel dependencies, followed by L2-norm-based feature selection to enhance generalization. Finally, the optimized TabNet model effectively classifies 24 fault categories, achieving an accuracy of 97.33%, and outperforms baseline methods including Temporal Convolutional Network (TCN), Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM), Capsule Network with Sparse Filtering (CNSF), and Dual-Branch CNN in terms of accuracy, macro-F1 score, and kappa coefficient. It also exhibits strong stability and fast convergence during training. These results demonstrate the robustness and interpretability of the proposed method for SPG fault detection. Full article

Transcriptomic profiling is a powerful tool for dissecting the cellular and molecular complexity of ocular tissues, providing insights into retinal development, corneal disease, macular degeneration, and glaucoma. With the expansion of microarray, bulk RNA sequencing (RNA-seq), and single-cell RNA-seq technologies, artificial intelligence (AI) has emerged as a key strategy for analyzing high-dimensional gene expression data. This review synthesizes AI-enabled transcriptomic studies in ophthalmology from 2019 to 2025, highlighting how supervised and unsupervised machine learning (ML) methods have advanced biomarker discovery, cell type classification, and eye development and ocular disease modeling. Here, we discuss unsupervised techniques, such as principal component analysis (PCA), t-distributed stochastic neighbor embedding (t-SNE), uniform manifold approximation and projection (UMAP), and weighted gene co-expression network analysis (WGCNA), now the standard in single-cell workflows. Supervised approaches are also discussed, including the least absolute shrinkage and selection operator (LASSO), support vector machines (SVMs), and random forests (RFs), and their utility in identifying diagnostic and prognostic markers in age-related macular degeneration (AMD), diabetic retinopathy (DR), glaucoma, keratoconus, thyroid eye disease, and posterior capsule opacification (PCO), as well as deep learning frameworks, such as variational autoencoders and neural networks that support multi-omics integration. Despite challenges in interpretability and standardization, explainable AI and multimodal approaches offer promising avenues for advancing precision ophthalmology. Full article

Background: Gastrointestinal (GI) disorders present significant healthcare challenges, requiring rapid, accurate, and effective diagnostic methods to improve treatment outcomes and prevent complications. Wireless capsule endoscopy (WCE) is an effective tool for diagnosing GI abnormalities; however, precisely identifying diverse lesions with similar visual patterns remains difficult. Methods: Many existing computer-aided diagnostic (CAD) systems rely on manually crafted features or single deep learning (DL) models, which often fail to capture the complex and varied characteristics of GI diseases. In this study, we proposed “EndoNet,” a multi-stage hybrid DL framework for eight-class GI disease classification using WCE images. Features were extracted from two different layers of three pre-trained convolutional neural networks (CNNs) (Inception, Xception, ResNet101), with both inter-layer and inter-model feature fusion performed. Dimensionality reduction was achieved using Non-Negative Matrix Factorization (NNMF), followed by selection of the most informative features via the Minimum Redundancy Maximum Relevance (mRMR) method. Results: Two datasets were used to evaluate the performance of EndoNer, including Kvasir v2 and HyperKvasir. Classification using seven different Machine Learning algorithms achieved a maximum accuracy of 97.8% and 98.4% for Kvasir v2 and HyperKvasir datasets, respectively. Conclusions: By integrating transfer learning with feature engineering, dimensionality reduction, and feature selection, EndoNet provides high accuracy, flexibility, and interpretability. This framework offers a powerful and generalizable artificial intelligence solution suitable for clinical decision support systems. Full article

Artificial intelligence (AI) has already played a significant role in the healthcare sector, particularly in image-based medical diagnosis. Deep learning models have produced satisfactory and useful results for accurate decision-making. Among the various types of medical images, magnetic resonance imaging (MRI) is frequently utilized in deep learning applications to analyze detailed structures and organs in the body, using advanced intelligent software. However, challenges related to performance and data privacy often arise when using medical data from patients and healthcare institutions. To address these issues, new approaches have emerged, such as federated learning. This technique ensures the secure exchange of sensitive patient and institutional data. It enables machine learning or deep learning algorithms to establish a client–server relationship, whereby specific parameters are securely shared between models while maintaining the integrity of the learning tasks being executed. Federated learning has been successfully applied in medical settings, including diagnostic applications involving medical images such as MRI data. This research introduces an analytical intelligence system based on an Internet of Medical Things (IoMT) framework that employs federated learning to provide a safe and effective diagnostic solution for brain tumor identification. By utilizing specific brain MRI datasets, the model enables multiple local capsule networks (CapsNet) to achieve improved classification results. The average accuracy rate of the CapsNet model exceeds 97%. The precision rate indicates that the CapsNet model performs well in accurately predicting true classes. Additionally, the recall findings suggest that this model is effective in detecting the target classes of meningiomas, pituitary tumors, and gliomas. The integration of these components into an analytical intelligence system that supports the work of healthcare personnel is the main contribution of this work. Evaluations have shown that this approach is effective for diagnosing brain tumors while ensuring data privacy and security. Moreover, it represents a valuable tool for enhancing the efficiency of the medical diagnostic process. Full article

Artificial intelligence (AI) is rapidly transforming imaging modalities in inflammatory bowel disease (IBD), particularly in endoscopy and ultrasound. Despite their critical roles, both modalities are challenged by interobserver variability, subjectivity, and accessibility issues. AI offers significant potential to address these limitations by enhancing lesion detection, standardizing disease activity scoring, and supporting clinical decision-making. In endoscopy, deep convolutional neural networks have achieved high accuracy in detecting mucosal abnormalities and grading disease severity, reducing observer dependency and improving diagnostic consistency. AI-assisted colonoscopy systems have also demonstrated improvements in procedural quality metrics, including adenoma detection rates and withdrawal times. Similarly, AI applications in intestinal ultrasound show promise in automating measurements of bowel wall thickness, assessing vascularity, and distinguishing between inflammatory and fibrotic strictures, which are critical for tailored therapy decisions. Video capsule endoscopy has likewise benefited from AI, reducing interpretation times and enhancing the detection of subtle lesions. Despite these advancements, implementation challenges, including dataset quality, standardization, AI interpretability, clinician acceptance, and regulatory and ethical considerations, must be carefully addressed. The current review focuses on the most recent developments in the integration of AI into experimental designs, medical devices, and clinical workflows for optimizing diagnostic accuracy, treatment strategies, and patient outcomes in IBD management. Full article

In view of the problems of data pollution, incomplete feature extraction, and poor multi-network parameter sharing and transmission under the federated learning framework of deep learning, this article proposes an improved asynchronous federated learning algorithm of multi-model fusion based on data injection pollution. Through data augmentation, the existing dataset is preprocessed to enhance the algorithm’s ability to identify the noise data. In our approach, the residual network is used to extract the static information of the image, the capsule network is used to extract the spatial dependence among the internal structures of the image, several layers of convolution are used to reduce the dimensions of both features, and the two extracted features are fused. In order to reduce the transmission overhead of parameters shared between the residual network and capsule network, we adopt an asynchronous parameter transmission between the global trainer and the local trainer. When the global trainer broadcasts the parameters to each local trainer, several trainers are randomly selected to avoid communication link blockage. Finally, through conducting various experiments, the results show that our alogrithm can effectively extract the pathological features in the image and achieve higher accuracy, outperforming the current mainstream algorithms. Full article

Extracting robust features against geometric transformations and adversarial perturbations remains a critical challenge in deep learning. Although capsule networks exhibit resilience through vector-encapsulated features and dynamic routing, they suffer from computational inefficiency due to iterative routing, dense matrix operations, and extra activation scalars. To address these limitations, we propose a method that integrates (1) compact vector-grouped neurons to eliminate activation scalars, (2) a non-iterative voting algorithm that preserves spatial relationships with reduced computation, and (3) efficient weight-sharing strategies that balance computational efficiency with generalizability. Our approach outperforms existing methods in image classification on CIFAR-10 and SVHN, achieving up to a 0.31% increase in accuracy with fewer parameters and lower FLOPs. Evaluations demonstrate superior performance over competing methods, with 0.31% accuracy gains on CIFAR-10/SVHN (with reduced parameters and FLOPs) and 1.93%/1.09% improvements in novel-view recognition on smallNORB. Under FGSM and BIM attacks, our method reduces attack success rates by 47.7% on CIFAR-10 and 32.4% on SVHN, confirming its enhanced robustness and efficiency. Future work will extend vexel representations to MLPs and RNNs and explore applications in computer graphics, natural language processing, and reinforcement learning. Full article

Variations in terrain elevation cause images acquired under different imaging modalities to deviate from a linear mapping relationship. This effect is particularly pronounced between optical and SAR images, where the range-based imaging mechanism of SAR sensors leads to significant local geometric distortions, such as perspective shrinkage and occlusion. As a result, it becomes difficult to represent the spatial correspondence between optical and SAR images using a single geometric model. To address this challenge, we propose a global optical-SAR image registration method that leverages local distortion characteristics. Specifically, we introduce a Superpixel-based Local Distortion Division (SLDD) method, which defines superpixel region features and segments the image into local distortion and normal regions by computing the Mahalanobis distance between superpixel features. We further design a Multi-Feature Fusion Capsule Network (MFFCN) that integrates shallow salient features with deep structural details, reconstructing the dimensions of digital capsules to generate feature descriptors encompassing texture, phase, structure, and amplitude information. This design effectively mitigates the information loss and feature degradation problems caused by pooling operations in conventional convolutional neural networks (CNNs). Additionally, a hard negative mining loss is incorporated to further enhance feature discriminability. Feature descriptors are extracted separately from regions with different distortion levels, and corresponding transformation models are built for local registration. Finally, the local registration results are fused to generate a globally aligned image. Experimental results on public datasets demonstrate that the proposed method achieves superior performance over state-of-the-art (SOTA) approaches in terms of Root Mean Squared Error (RMSE), Correct Match Number (CMN), Distribution of Matched Points (Scat), Edge Fidelity (EF), and overall visual quality. Full article

Transfer learning is the act of using the data or knowledge in a problem to help solve different but related problems. In a brain computer interface (BCI), it is important to deal with individual differences between topics and/or tasks. A kind of capsule decision neural network (CDNN) based on transfer learning is proposed. In order to solve the problem of feature distortion caused by EEG feature extraction algorithm, a deep capsule decision network was constructed. The architecture includes multiple primary capsules to form a hidden layer, and the connection between the advanced capsule and the primary capsule is determined by the neural decision routing algorithm. Unlike the dynamic routing algorithm that iteratively calculates the similarity between primary capsules and advanced capsules, the neural decision network computes the relationship between each capsule in the deep and shallow hidden layers in a probabilistic manner. At the same time, the distribution of the EEG covariance matrix is aligned in Riemann space, and the regional adaptive method is further introduced to improve the independent decoding ability of the capsule decision neural network for the subject’s EEG signals. Experiments on two motor imagery EEG datasets show that CDNN outperforms several of the most advanced transfer learning methods. Full article

Diabetes, a chronic medical condition, affects millions of people worldwide and requires consistent monitoring of blood glucose levels (BGLs). Traditional invasive methods for BGL monitoring can be challenging and painful for patients. This study introduces a non-invasive, deep learning (DL)-based approach to estimate BGL using photoplethysmography (PPG) signals. Specifically, a Deep Sparse Capsule Network (DSCNet) model is proposed to provide accurate and robust BGL monitoring. The proposed model’s workflow includes data collection, preprocessing, feature extraction, and predictions. A hardware module was designed using a PPG sensor and Raspberry Pi to collect patient data. In preprocessing, a Savitzky–Golay filter and moving average filter were applied to remove noise and preserve pulse form and high-frequency components. The DSCNet model was then applied to predict the sugar level. Two models were developed for prediction: a baseline model, DSCNet, and an enhanced model, DSCNet with self-attention. DSCNet’s performance was evaluated using Mean Absolute Percentage Error (MAPE), Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Relative Difference (MARD), and coefficient of determination (R²), yielding values of 3.022, 0.05, 0.058, 0.062, 10.81, and 0.98, respectively. Full article

Celiac disease and gluten sensitivity affect a significant portion of the population and require adherence to a gluten-free diet. Dining in social settings, such as family events, workplace gatherings, or restaurants, makes it difficult to ensure that certain foods are gluten-free. Despite the availability of portable gluten testing devices, these instruments have high costs, disposable capsules, depend on user preparation and technique, and cannot analyze an entire meal or detect gluten levels below the legal thresholds, potentially leading to inaccurate results. In this study, we propose RGB (Recognition of Gluten in Bread), a novel deep learning-based method for automatically detecting gluten in bread images. RGB is a decision-support tool to help individuals with celiac disease make informed dietary choices. To develop this method, we curated and annotated three unique datasets of bread images collected from Pinterest, Instagram, and a custom dataset containing information about flour types. Fine-tuning pre-trained convolutional neural networks (CNNs) on the Pinterest dataset, our best-performing model, ResNet50V2, achieved 77% accuracy and recall. Transfer learning was subsequently applied to adapt the model to the Instagram dataset, resulting in 78% accuracy and 77% recall. Finally, further fine-tuning the model on a significantly different dataset, the custom bread dataset, significantly improved the performance, achieving an accuracy of 86%, precision of 87%, recall of 86%, and F1-score of 86%. Our analysis further revealed that the model performed better on gluten-free flours, achieving higher accuracy scores for these types. This study demonstrates the feasibility of image-based gluten detection in bread and highlights its potential to provide a cost-effective non-invasive alternative to traditional testing methods by allowing individuals with celiac disease to receive immediate feedback on potential gluten content in their meals through simple food photography. Full article

The development of emotionally intelligent computers depends on emotion recognition based on richer multimodal inputs, such as text, speech, and visual cues, as multiple modalities complement one another. The effectiveness of complex relationships between modalities for emotion recognition has been demonstrated, but these relationships are still largely unexplored. Various fusion mechanisms using simply concatenated information have been the mainstay of previous research in learning multimodal representations for emotion classification, rather than fully utilizing the benefits of deep learning. In this paper, a unique deep multimodal emotion model is proposed, which uses the meaningful neural network to learn meaningful multimodal representations while classifying data. Specifically, the proposed model concatenates multimodality inputs using a graph convolutional network to extract acoustic modality, a capsule network to generate the textual modality, and vision transformer to acquire the visual modality. Despite the effectiveness of MNN, we have used it as a methodological innovation that will be fed with the previously generated vector parameters to produce better predictive results. Our suggested approach for more accurate multimodal emotion recognition has been shown through extensive examinations, producing state-of-the-art results with accuracies of 69% and 56% on two public datasets, MELD and MOSEI, respectively. Full article