Search Results (2,003)

This paper presents a comprehensive study on automated defect detection in complex structures using phased array ultrasonic testing data, focusing on both traditional signal processing and advanced deep learning methods. As a non-AI baseline, the well-known signal-to-noise ratio algorithm was improved by introducing automatic depth gate calculation using derivative analysis and eliminated the need for manual parameter tuning. Even though this method demonstrates robust flaw indication, it faces difficulties for automatic defect detection in highly noisy data or in cases with large pore zones. Considering this, multiple DL architectures—including fully connected networks, convolutional neural networks, and a novel Convolutional Attention Temporal Transformer for Sequences—are developed and trained on diverse datasets comprising simulated CIVA data and real-world data files from welded and composite specimens. Experimental results show that while the FCN architecture is limited in its ability to model dependencies, the CNN achieves a strong performance with a test accuracy of 94.9%, effectively capturing local features from PAUT signals. The CATT-S model, which integrates a convolutional feature extractor with a self-attention mechanism, consistently outperforms the other baselines by effectively modeling both fine-grained signal morphology and long-range inter-beam dependencies. Achieving a remarkable accuracy of 99.4% and a strong F1-score of 0.905 on experimental data, this integrated approach demonstrates significant practical potential for improving the reliability and efficiency of NDT in complex, heterogeneous materials. Full article

Background. Automatic detection of Sit phases in people with Parkinson’s disease (PD) using a single body-worn sensor is crucial for enhancing long-term, home-based monitoring of mobility. Aim. The aim of this study was to enhance the accuracy of detecting and segmenting Sit phases in people with PD using a single SmartWatch worn at the ankle. Method. Twenty-two patients living with PD performed activities of daily living that incorporate repeated transitions to a seated position in a simulated free-living environment during 3 min, 4 min, and 5 min trials. Tri-axial accelerations and angular velocities of the right or left ankle were recorded at 50 Hz using a SmartWatch. Random forest algorithms were trained using raw and filtered data to automatically detect and segment Sit phases. Sensibility, specificity, and F-scores were calculated based on manual segmentation using the OptiTrack motion capture system. Results. Sensibility, specificity, and F-score achieved 78.3%, 93.8%, and 84.7% for Sit phase detection of the 3 min trial; 78.8%, 85.5%, and 80.6% for Sit phase detection of the 4 min trial; and 71.6%, 84.8%, and 75.6% for Sit phase detection of the 5 min trial. The median time difference between the manual and automatic segmentation was 0.95s, 0.89s, and 0.84s, respectively, for the 3 min, 4 min, and 5 min trial. Conclusion. This study demonstrates that a random forest algorithm can accurately detect and segment Sit phases in people with PD using data from a single ankle-worn SmartWatch. The algorithm’s performance was comparable to manual segmentation, while substantially reducing the time and effort required. These findings represent a meaningful step forward in enabling efficient, long-term, and home-based monitoring of mobility and symptom progression in people with PD. Full article

Robot-assisted radical prostatectomy (RARP) has become the most prevalent treatment for patients with organ-confined prostate cancer. Despite superior outcomes, suboptimal vesicourethral anastomosis (VUA) may lead to serious complications, including urinary leakage, prolonged catheterization, and extended hospitalization. A precise localization of both the surgical needle and the surrounding vesical and urethral tissues to coadapt is needed for fine-grained assessment of this task. Nonetheless, the identification of anatomical structures from endoscopic videos is difficult due to tissue distortions, changes in brightness, and instrument interferences. In this paper, we propose and compare two Deep Learning (DL) pipelines for the automatic segmentation of the mucosal layers and the suturing needle in real RARP videos by exploiting different architectures and training strategies. To train the models, we introduce a novel, annotated dataset collected from four VUA procedures. Experimental results show that the nnU-Net 2D model achieved the highest class-specific metrics, with a Dice Score of 0.663 for the mucosa class and 0.866 for the needle class, outperforming both transformer-based and baseline convolutional approaches on external validation video sequences. This work paves the way for computer-assisted tools that can objectively evaluate surgical performance during the critical phase of suturing tasks. Full article

We propose an edge-enhanced federated learning framework for real-time itinerary optimization in elderly oriented adventure tourism, addressing the critical need for adaptive scheduling that balances activity intensity with health constraints. The system integrates lightweight convolutional neural networks with a priority-based scheduling algorithm, processing participant profiles and real-time biometric data through a decentralized computation model to enable dynamic adjustments. A modified Hungarian algorithm incorporates physical exertion scores, temporal proximity weights, and health risk factors, then optimizes activity assignments while respecting physiological recovery requirements. The federated learning architecture operates across distributed edge nodes, preserving data privacy through localized model training and periodic global aggregation. Furthermore, the framework interfaces with transportation systems and medical monitoring infrastructure, automatically triggering itinerary modifications when vital sign anomalies exceed adaptive thresholds. Implemented on NVIDIA Jetson AGX Orin modules, the system achieves 300 ms end-to-end latency for real-time schedule updates, meeting stringent safety requirements for elderly participants. The proposed method demonstrates significant improvements over conventional itinerary planners through its edge computing efficiency and personalized adaptation capabilities, particularly in handling the latency-sensitive demands of intensive tourism scenarios. Experimental results show robust performance across diverse participant profiles and activity types, confirming the system’s practical viability for real-world deployment in elderly adventure tourism operations. Full article

Various types of interfering signals are an integral part of ECGs recorded using wearable electronics, specifically during field monitoring, outside the controlled environment of a medical doctor’s office, or laboratory. The frequency spectrum of several types of interfering signals overlaps significantly with the ECG signal, making effective filtration impossible without losing clinically relevant information. In this article, we proceed from the practical assumption that it is unnecessary to analyze the entire ECG recording in real long-term recordings. Conversely, in the preprocessing phase, it is necessary to detect unreadable segments of the ECG signal. This paper proposes a novel method for automatically detecting unreadable segments distorted by superimposed interference in ECG recordings. The method is based on a convolutional neural network (CNN) and is comparable in quality to annotation performed by a medical expert, but incomparably faster. In a series of controlled experiments, the ECG signal was recorded during physical activities of varying intensities, and individual segments of the recordings were manually annotated based on visual assessment by a medical expert, i.e., divided into four different classes based on the intensity of distortion to the useful ECG signal. A deep convolutional model was designed and evaluated, exhibiting a 87.62% accuracy score and the same F1-score in automatic recognition of segments distorted by superimposed interference. Furthermore, the model exhibits an accuracy and F1-score of 98.70% in correctly identifying segments with visually detectable and non-detectable heart rate. The proposed interference detection procedure appears to be sufficiently effective despite its simplicity. It facilitates subsequent automatic analysis of undisturbed ECG waveform segments, which is crucial in ECG monitoring using wearable electronics. Full article

Remote sensing provides a viable alternative for understanding landscape modifications attributed to beaver activity. The central objective of this study is to integrate multi-source remote sensing observations in tandem with a deep learning (DL) (convolutional neural net or transformer) model to automatically map beaver-influenced floodplain inundations (BIFI) over large geographical extents. We trained, validated, and tested eleven different model configurations in three architectures using five ResNet and five B-Finetuned encoders. The training dataset consisted of >25,000 manually annotated aerial image tiles of BIFIs in Connecticut. The YOLOv8 architecture outperformed competing configurations and achieved an F1 score of 80.59% and pixel-based map accuracy of 98.95%. SegFormer and U-Net++’s highest-performing models had F1 scores of 68.98% and 78.86%, respectively. The YOLOv8l-seg model was deployed at a statewide scale based on 1 m resolution multi-temporal aerial imagery acquired from 1990 to 2019 under leaf-on and leaf-off conditions. Our results suggest a variety of inferences when comparing leaf-on and leaf-off conditions of the same year. The model exhibits limitations in identifying BIFIs in panchromatic imagery in occluded environments. Study findings demonstrate the potential of harnessing historical and modern aerial image datasets with state-of-the-art DL models to increase our understanding of beaver activity across space and time. Full article

Accurate quantification of minor mineral phases is important in Powder X-Ray Diffraction (PXRD) and Rietveld phase quantification. The precise limit of quantification for the various phases is rarely considered but rather approximated to 0.2–2 wt% by applying a global minimum weight percentage threshold. This approximation often leads to false positive or false negative phase quantity, jeopardizing the trustworthiness of the analytic method in general. In this work (1) we propose a dynamic and adaptable false positive filtering method for Rietveld Quantitative X-ray diffraction (QXRD) based on a phase-specific signal-to-noise ratio referred to as “Phase-SNR”; (2) we introduce the method baptized “Phase Guard” which is implemented in the software HighScore Plus. Phase Guard is based on peaks counting statistics and it automatically adapts to different mineral scattering powers, different mineral crystallinity, instrumental configuration and measurement time. Its applicability and benefits are demonstrated with several examples in cement and mining applications. The adoption of Phase Guard is especially beneficial for industrial black-box solutions, where all “probable” phases are included in the model, even when they are absent from the sample. Phase Guard eliminates false positives, it reduces the likelihood of false negatives, and it is an essential tool to answer the question “what is the limit of quantification for Rietveld analysis?” Full article

During the automatic leveling process of small low-cost unmanned aerial vehicles (UAVs) after airdrop, their state parameters and control surface efficiency undergo drastic changes. It is difficult to achieve good control effects using controllers with fixed parameters. To solve these problems, this study proposes a parameter adaptive PID controller based on indicated airspeed. When tuning the controller parameters, in order to ensure the successful pulling of the UAV and the safety of structure and flight, it is necessary to optimize the success rate of pulling up, normal overload, angle of attack (AOA), airspeed, and descent altitude simultaneously. These five indicators are of different importance to the UAV. To facilitate parameter tuning based on these differences, an improved second-generation non-dominated sorting genetic algorithm (NSGA-II) is proposed, which combines a comprehensive fitness mechanism based on target priority and segmented scoring and an adaptive genetic strategy. In this study, different priorities were set for all indicators, and segmented scores were given based on individual indicators to calculate the comprehensive fitness, which guided the evolutionary direction of the population. Then, while the genetic parameters were modified, elite individuals were retained to balance search ability and convergence. Finally, the effectiveness of this mechanism was confirmed through comparative simulation. The flight test results show significant differences from the simulation results of the controller designed in this study, but the basic trend remains consistent. The controller can effectively suppress the oscillations caused by the initial state. Full article

This study aimed to evaluate short-term clinical outcomes in COVID-19 pneumonia patients using parameters derived from a commercial deep learning-based automatic detection algorithm (DLAD) applied to serial chest radiographs (CXRs). We analyzed 391 patients with COVID-19 who underwent serial CXRs during isolation at a residential treatment center (median interval: 3.57 days; range: 1.73–5.56 days). Patients were categorized into two groups: the improved group (n = 309), who completed the standard 7-day quarantine, and the deteriorated group (n = 82), who showed worsening symptoms, vital signs, or CXR findings. Using DLAD’s consolidation probability scores and gradient-weighted class activation mapping (Grad-CAM)-based localization maps, we quantified the consolidation area through heatmap segmentation. The weighted area was calculated as the sum of the consolidation regions’ areas, with each area weighted by its corresponding probability score. Change rates (Δ) were defined as per-day differences between consecutive measurements. Prediction models were developed using Cox proportional hazards regression and evaluated daily from day 1 to day 7 after the subsequent CXR acquisition. Among the imaging factors, baseline probability and ΔProbability, ΔArea, and ΔWeighted area were identified as prognostic indicators. The multivariate Cox model incorporating baseline probability and ΔWeighted area demonstrated optimal performance (C-index: 0.75, 95% Confidence Interval: 0.68–0.81; integrated calibration index: 0.03), with time-dependent AUROC (Area Under Receiver Operating Curve) values ranging from 0.74 to 0.78 across daily predictions. These findings suggest that the Δparameters of DLAD can aid in predicting short-term clinical outcomes in patients with COVID-19. Full article

This study aims to improve the automatic detection of unwanted emails using advanced machine learning and deep learning methods. By reviewing current research over the past five years, a comprehensive combined dataset structure was created containing a total of 81,586 email samples from seven different spam datasets. Class imbalance was addressed through the application of random oversampling and class-weighted loss, and the decision threshold was subsequently tuned for deployment. Among classical machine learning solutions, Random Forest (RF) emerged as the most successful method, while deep learning approaches, such as Transformer-based models like Distilled Bidirectional Encoder Representations from Transformers (DistilBERT) and Robustly Optimized BERT Pretraining Approach (RoBERTa), demonstrated superior performance. The highest test score (99.62%) on a combined static dataset was achieved with a multimodal architecture that combines deep meaningful text representations from DistilBERT with structural text features. Beyond this static performance benchmark, the study investigates the critical challenge of concept drift by performing a temporal analysis on datasets from different eras. The results reveal a significant performance degradation in all models when tested on modern spam, highlighting a critical vulnerability of statically trained systems. Notably, the Transformer-based model demonstrated greater robustness against this temporal decay compared to traditional methods. This study offers not only an effective classification solution but also provides crucial empirical evidence on the necessity of adaptive, continually learning systems for robust spam detection. Full article

Face morphing attacks have become a serious threat to Face Recognition Systems (FRSs). A de-morphing-based morphing attack detection method has been proposed and studied, which uses suspect and live capture, but the unknown morphing parameters in the used morphing algorithm make applying de-morphing methods challenging. This paper proposes a robust face morphing attack detection (FMAD) method (pipeline) leveraging deep learning de-morphing networks. Inspired by differences in similarity score (i.e., cosine similarity between feature vectors) variations between morphed and non-morphed images, the detection pipeline was proposed to learn the variation patterns of similarity scores between live capture and de-morphed face/bona fide images with different de-morphing factors. An effective deep de-morphing network based on StyleGAN and the pSp (pixel2style2pixel) encoder was developed. The network generates de-morphed images from suspect and live images with multiple de-morphing factors and calculates similarity scores between feature vectors from the ArcFace network, which are then classified by the detection network. Experiments on morphing datasets from the Color FERET, FRGCv2, and SYS-MAD databases, including landmark-based and deep learning attacks, demonstrate that the proposed method performs high accuracy in detecting unseen morphing attacks across different databases. It attains an Equal Error Rate (EER) of less than 1–4% and a Bona Fide Presentation Classification Error Rate (BPCER) of approximately 11% at an Attack Presentation Classification Error Rate (APCER) of 0.1%, outperforming previous methods. Full article

The study of dramatic plays has long relied on qualitative methods to analyze character interactions, making little assumption about the structural patterns of communication involved. Our approach bridges NLP and literary studies, enabling scalable, data-driven analysis of interaction patterns and power structures in drama. We propose a novel method to supplement addressee identification in tragedies using Large Language Models (LLMs). Unlike conventional Social Network Analysis (SNA) approaches, which often diminish dialogue dynamics by relying on co-occurrence or adjacency heuristics, our LLM-based method accurately records directed speech acts, joint addresses, and listener interactions. In a preliminary evaluation of an annotated multilingual dataset of 14 scenes from nine plays in four languages, our top-performing LLM (i.e., Llama3.3-70B) achieved an F1-score of $88.75\%$ (P = $94.81\%$, R = $84.72\%$), an exact match of 77.31%, and an 86.97% partial match with human annotations, where partial match indicates any overlap between predicted and annotated receiver lists. Through automatic extraction of speaker–addressee relations, our method provides preliminary evidence for the potential scalability of SNA for literary analyses, as well as insights into power relations, influence, and isolation of characters in tragedies, which we further visualize by rendering social network graphs. Full article

Assessing teaching behavior is essential for improving instructional quality, particularly in Physical Education, where classroom interactions are fast-paced and complex. Traditional evaluation methods such as questionnaires, expert observations, and manual discourse analysis are often limited by subjectivity, high labor costs, and poor scalability. These challenges underscore the need for automated, objective tools to support pedagogical assessment. This study explores and compares the use of Transformer-based language models for the automatic classification of teaching behaviors from real classroom transcriptions. A dataset of over 1300 utterances was compiled and annotated according to the teaching styles proposed in the circumplex approach (Autonomy Support, Structure, Control, and Chaos), along with an additional category for messages in which no style could be identified (Unidentified Style). To address class imbalance and enhance linguistic variability, data augmentation techniques were applied. Eight pretrained BERT-based Transformer architectures were evaluated, including several pretraining strategies and architectural structures. BETO achieved the highest performance, with an accuracy of 0.78, a macro-averaged F1-score of 0.72, and a weighted F1-score of 0.77. It showed strength in identifying challenging utterances labeled as Chaos and Autonomy Support. Furthermore, other BERT-based models purely trained with a Spanish text corpus like DistilBERT also present competitive performance, achieving accuracy metrics over 0.73 and and F1-score of 0.68. These results demonstrate the potential of leveraging Transformer-based models for objective and scalable teacher behavior classification. The findings support the feasibility of leveraging pretrained language models to develop scalable, AI-driven systems for classroom behavior classification and pedagogical feedback. Full article

Background/Objectives: Frank’s sign, a diagonal earlobe crease, may reflect systemic microvascular dysfunction. We investigated whether Frank’s sign serves as a clinical marker of white matter hyperintensity (WMH) burden in Cerebral Autosomal Dominant Arteriopathy with Subcortical Infarcts and Leukoencephalopathy (CADASIL), a monogenic model of pure cerebral small vessel disease. Methods: We analyzed 81 genetically confirmed CADASIL patients (61.8 ± 12.6 years, 40.7% female) and 54 age/sex-matched controls (70.3 ± 6.6 years, 48.1% female). Frank’s sign was detected using deep learning from brain MRI-reconstructed 3D facial surfaces. WMH volumes were automatically quantified and adjusted for confounders using Random Forest regression residuals. We compared Frank’s sign prevalence between groups, assessed within-CADASIL associations, and evaluated dose–response relationships across WMH tertiles. Results: Frank’s sign prevalence was significantly higher in CADASIL versus controls (66.7% vs. 42.6%, p = 0.020), with strengthened association after multivariate adjustment (OR = 4.214, 95% CI: 1.128–15.733, p = 0.032). Within CADASIL, Frank’s sign-positive patients showed 72% greater WMH burden (51.5 ± 27.1 vs. 30.0 ± 26.1 mL, p < 0.001) and lower Consortium to Establish a Registry for Alzheimer’s Disease (CERAD) total scores (57.7 ± 19.6 vs. 71.2 ± 22.8, p = 0.006), but similar lacunes, microbleeds, and hippocampal volumes. A robust dose–response relationship emerged across WMH tertiles, with Frank’s sign prevalence increasing from 37.0% (lowest) to 74.1% (highest tertile; adjusted OR = 3.571, 95% CI: 1.134–11.253, p = 0.030). Conclusions: Frank’s sign represents an accessible biomarker of WMH burden in CADASIL, demonstrating disease-specificity and dose–response characteristics independent of vascular risk factors. The automated MRI-based detection method of Frank’s sign enables retrospective analysis of existing neuroimaging databases, transforming a bedside observation into a quantifiable neuroimaging biomarker for genetic small vessel disease stratification. Full article

Agricultural drainage canals provide critical habitats for fish species that are highly sensitive to agricultural practices. However, conventional monitoring methods such as capture surveys are invasive and labor-intensive, which means they can disturb fish populations and hinder long-term ecological assessment. Therefore, there is a strong need for effective and non-invasive monitoring techniques. In this study, we developed a practical method using the YOLOv8n deep learning model to automatically detect and quantify fish occurrence in underwater images from a canal in Ibaraki Prefecture, Japan. The model showed high performance in validation (F1-score = 91.6%, Precision = 95.1%, Recall = 88.4%) but exhibited reduced performance under real field conditions (F1-score = 61.6%) due to turbidity, variable lighting, and sediment resuspension. By correcting for detection errors, we estimated that approximately 7300 individuals of Pseudorasbora parva and 80 individuals of Cyprinus carpio passed through the observation site during a seven-hour monitoring period. These findings demonstrate the feasibility of deep learning-based monitoring to capture temporal patterns of fish occurrence in agricultural drainage canals. This approach provides a promising tool for sustainable aquatic ecosystem management in agricultural landscapes and emphasizes the need for further improvements in recall under turbid and low-visibility conditions. Full article