Search Results (1,211)

Manual industrial patrol inspection relies heavily on on-site visual observation, while the visual attention of inspectors toward specific target equipment is difficult to quantify. This study proposes an industrial patrol inspection visual attention analysis method integrating eye tracking and deep learning-based object detection. A wearable eye-tracking device is used to acquire first-person inspection videos, eye-tracking data, and camera parameters. After temporal alignment of the multimodal data, the 3D gaze points are projected onto the corresponding 2D inspection video frames. YOLO12m is employed to detect target equipment, followed by frame-by-frame matching between the 2D gaze points and the equipment bounding boxes. Based on the matching results, seven visual attention metrics are calculated to quantitatively characterize visual attention during patrol inspection. On-site experiments with 10 participants yielded 21 inspection records, with overall visual attention scores ranging from 43.16 to 89.71 and averaging 75.99. Three experts independently rated the 21 records, and the system-generated ratings agreed with the majority expert ratings for 19 records. These results demonstrate that, under the present experimental conditions, the proposed method can quantify the visual attention of inspectors toward specific target equipment during industrial patrol inspection using multiple metrics and provide interpretable results. Full article

On-skin non-invasive Wearable Health-Monitoring Devices (WHMDs) have rapidly evolved from laboratory prototypes into commercially viable systems capable of continuously tracking physiological and biochemical signals. By integrating epidermal temperature sensors, electrophysiological electrodes, biochemical sensing platforms, low-power electronics, and wireless communication technologies, these systems are emerging as key enablers of personalized and decentralized healthcare through the continuous acquisition of clinically relevant information directly from the skin surface. In this Perspective, we present our view on the state-of-the-art across the key technological pillars that define modern on-skin WHMDs, including non-invasive sensing strategies, advanced materials, processing and wireless communication units, energy-storage solutions, energy-harvesting techniques, and power-management architectures, with a particular focus on technologies that have already reached high Technology Readiness Levels (TRLs). We highlight how the next-generation of on-skin WHMDs must balance performance with sustainability and long-term reliability. This includes the adoption of biodegradable and recyclable materials, low-power and reconfigurable electronics, solid-state batteries, and hybrid energy-harvesting systems. By aligning technological innovation with human-centric and eco-friendly design principles, on-skin WHMDs can evolve into scalable, equitable, and environmentally responsible tools for future digital healthcare. Full article

Colles fracture is a common type of wrist fracture, typically resulting from falling onto an outstretched hand. Postoperative patients often require long-term self-rehabilitation to restore wrist function. However, traditional rehabilitation relies on medical personnel and lacks real-time feedback and progress tracking, which may lead to poor rehabilitation or even deterioration of the condition. This article proposes an intelligent wearable system based on edge computing to enhance the efficiency of self-rehabilitation, improve system portability, and ensure comprehensive recording of rehabilitation data. The system performs real-time data processing and feedback to assist patients and healthcare providers in monitoring rehabilitation progress and optimizing recovery outcomes. The proposed system integrates an STM32 microcontroller, NanoEdge AI, and a 9-axis inertial sensor to detect and evaluate the accuracy of hand rehabilitation movements, ensuring the precision of self-rehabilitation. All rehabilitation movements are schemed under the guidance of professional physicians to ensure correctness and minimize the risk of secondary injuries caused by improper rehabilitation. Data such as motion records, training duration, and count are transmitted via Bluetooth to the Local database for further analysis. A custom web interface allows healthcare providers to monitor and analyze collected data. This approach improves diagnostic accuracy, supports personalized rehabilitation recommendations, and improves treatment outcomes and patient recovery success rates. Full article

This paper addresses the task of binary violence classification in short surveillance video clips, i.e., deciding whether a given clip contains violent interactions (Fight) or not (Non-Fight). Although this task is often discussed in the broader context of video-based abnormal behavior detection driven by smart cities, Closed-Circuit Television (CCTV) networks, and intelligent surveillance systems, existing methods for violence classification largely rely on appearance features from single frames or global motion cues across whole scenes and therefore fail to adequately capture the interactions among multiple objects and the structural changes in their relationships that arise in real surveillance environments. In particular, violent behavior is rarely defined by a specific pose or a single moment; rather, it emerges as a cumulative process in which relational changes—such as inter-person approach, distance variation, collision, and repeated contact—unfold over time. Detecting such behavior accurately therefore requires an approach that can analyze inter-object relationships in a spatiotemporal manner. To this end, this paper proposes a violence detection method that combines multi-object tracking with spatiotemporal graph-based relational pattern mining. The proposed method first detects and tracks person objects using YOLO and DeepSORT, and extracts time-series features—including position, velocity, pose, and inter-object distance variation—to construct a spatiotemporal graph. Relational event sequences are then generated from the edge features of the graph, and class-representative relational patterns are automatically extracted based on discriminative power through PrefixSpan-based frequent sequential pattern mining. In parallel, the spatiotemporal graph is fed into a Spatial Temporal Graph Convolutional Network (ST-GCN) to learn the structural relationships among objects and their temporal evolution. Finally, the pattern-matching score and the ST-GCN classification score are combined to classify each input video as either violent or non-violent. By jointly exploiting interpretable relational pattern information and graph-based structural learning, the proposed approach compensates for the limitations of appearance-centric anomaly detection and demonstrates its applicability to complex real-world surveillance environments. Performance is evaluated in terms of Accuracy, Precision, Recall, and F1-score, with Recall considered a primary metric to reflect the importance of not missing violent events. Full article

Allogeneic hematopoietic stem cell transplantation (HSCT) serves as a critical model for investigating clonal dynamics, as hematopoietic reconstitution occurs amidst significant proliferative stress and immune-mediated pressures. This study utilized longitudinal next-generation sequencing of a 30-gene panel to track clonal evolution in 20 patients and their matched-related or haploidentical donors. The findings revealed that post-transplant hematopoiesis is frequently driven by donor-derived clones, with DNMT3A being the most prevalent mutation in the study population. A key finding was the significant association between the presence of the donor TET2 Leu1721Trp variant and reduced overall survival in recipients (median 14.3 months compared with 50.8 months for wild type; p = 0.0100). Additionally, the study showed that the re-emergence of recipient-derived clones, particularly oncogenic DNMT3A mutations, frequently preceded graft failure or disease relapse. These findings suggest that, although donor clonal hematopoiesis is generally considered clinically neutral, specific genetic variants can profoundly influence clinical outcomes, suggesting that comprehensive molecular screening of donors and longitudinal molecular monitoring of recipients may be essential to optimize transplant outcomes and personalize post-HSCT management. Full article

Adaptive and explainable systems are two critical concepts in personalized learning that enhance the customized experience based on the adjusted content according to the individual learning of a learner. In this paper, we present a modular and interpretable architecture for the development of a System for Adaptive and Explainable Feedback, which is a potential solution for the above-mentioned issues in personalized learning environments. The adaptive feedback provided by the System for Adaptive and Explainable Feedback is in real time. Still, even more importantly, it is justified clearly and transparently so that both the learner and instructor understand the why behind the recommendations for action given by the system. It consists of modular technologies for data collection (word and phrase occurrence, time tracking), analysis (latency, process mining), feedback generation (adaptive after-action review), and result presentation, making a modular, flexible, and scalable system to suit many educational scenarios. To evaluate the system’s functional performance, the SAEF pipeline was applied to a dataset derived from the ASSISTments platform, a well-established educational dataset widely used in learning analytics research. On a cohort of 500 student profiles, the SAEF achieved an overall recommendation accuracy of 83.6%, a weighted F1-score of 82.9%, and a mean system response time of 42.1 ms, demonstrating both the internal computational consistency and efficiency of the adaptive pipeline. These results indicate strong agreement with score-derived difficulty categories and support the internal computational consistency of the recommendation pipeline as a proof-of-concept system, though they do not constitute independent evidence of instructional appropriateness. The SAEF is designed to support learner engagement, personalize learning pathways, and foster transparency in AI-driven educational environments; a full empirical evaluation involving real-world deployment is identified as the primary direction for future work. The perceived understandability, trustworthiness, and pedagogical usefulness of the SAEF’s explanations by learners and instructors represent a complementary dimension yet to be empirically explored. The SAEF’s architecture is designed with the explicit objective of supporting learner engagement and fostering transparency; however, these pedagogical benefits are architectural design goals rather than empirically demonstrated outcomes in the present study, which focuses exclusively on computational validation. Full article

Lower-limb rehabilitation exoskeletons have emerged as a promising complementary technology to conventional therapy, enabling repetitive, intensive, and personalized gait training. However, achieving accurate trajectory tracking while maintaining robustness against parametric uncertainties, external disturbances, and unpredictable human–robot interaction remains a significant control challenge due to the highly nonlinear dynamics of coupled human–exoskeleton systems. This paper presents an experimental performance comparison of five control strategies for gait rehabilitation exoskeletons, including a classical proportional–integral–derivative (PID) controller, a model-based proportional–derivative controller with gravity compensation (PD+G), a computed torque sliding mode controller (CT-SMC), a computed torque–super-twisting sliding mode controller (CT–ST-SMC) and a hybrid backstepping–super-twisting sliding mode controller (BS–ST-SMC). All the controllers were implemented on the same lower-limb rehabilitation exoskeleton under identical operating conditions. The experimental results demonstrate that the proposed BS–ST-SMC architecture outperforms classical and traditional robust approaches, particularly in mitigating chattering and managing human–robot interaction uncertainties. Specifically, the BS–ST-SMC achieved the highest tracking precision with a mean squared position error (MSEp) of $1.32\times {10}^{-3}{\mathrm{rad}}^2$ and effectively synchronized with the user by reducing the phase lag to just 4.22° at the knee joint. Their overall performance was evaluated using the following metrics: mean squared position error (MSEP), mean squared velocity error (MSEv), peak error, phase lag, jerk index, peak torque, and peak power. Full article

Graphs are essential tools in introductory STEM courses, yet students often read them as pictures rather than quantitative representations of changing quantities. This study examined how 18 first-semester general chemistry students interpreted two structurally parallel, three-question graphs in chemistry and physics: a concentration-time graph of a chemical reaction and a position–time graph of a walking person. Using eye-tracking, written work, and accuracy data, we found high but asymmetric performance (82% in chemistry; 89% in physics), with rate calculations as the most challenging items in both domains. Students commonly conflated value with rate and sometimes averaged from the origin across discontinuities, with these difficulties more prevalent in chemistry. Trials resulting in correct answers typically followed a question → axes → curve pattern with brief returns to the prompt, whereas trials resulting in incorrect answers often showed early and repeated attention to salient non-target graph features without careful attention to axes. These findings support teaching graph comprehension as a discipline; it is a general but context-sensitive skill and has potential value in an axes-first, rise-over-run-based protocol for rate-of-change reasoning. Full article

Large Language Models (LLMs) are increasingly incorporated into software systems. Their non-deterministic behavior, external hosting, response latency, and operational cost create challenges for established design approaches such as Clean Architecture. This paper examines the integration of an LLM component into a layered software system and compares three possible placements within Clean Architecture: Domain, Application, and Infrastructure. The evaluation considers dependency management, testability, separation of concerns, and implementation complexity. The study proposes an approach in which the LLM is implemented in the Infrastructure layer and accessed through an interface defined in the Application layer. This approach is combined with the Command and Query Responsibility Segregation pattern to isolate LLM interaction within dedicated query handlers. The proposed pattern is demonstrated through the implementation of Budget, a personal finance tracking system that uses GPT-4.1 to convert free-form natural language input into structured transaction records. The results show that placement in the infrastructure layer provides the clearest separation between business logic and external AI services and avoids the introduction of non-deterministic behavior into the core application logic. Full article

Fall detection in multi-person environments, such as nursing homes and rehabilitation centers, is essential for ensuring the safety of vulnerable populations. Despite advances in deep learning, current vision- and skeleton-based fall detection systems often exhibit false negatives and reduced reliability in real-world scenarios due to scene complexity. This study presents a system-level analysis of fall detection errors by comparing four approaches—two skeleton-based methods using ST-GCN and ProtoGCN, a rule-based method, and a VIRA-GCN-based 3D joint method—on 95 RGB video sequences captured under minimally constrained multi-person conditions. We define six error types: skeleton structural interference, localized joint recognition failure, temporal skeleton identity inconsistency, object-to-skeleton association failure, viewpoint-induced observation limitation, and action-level ambiguity with similar activities. Although most methods achieved high event-level recall, their false-positive and false-negative patterns differed. The rule-based approach showed the most balanced performance, under the present experimental conditions, whereas the ST-GCN-based skeleton approach was more sensitive to joint-level and tracking instability. ProtoGCN reduced false positives but increased false negatives, showing a more conservative decision pattern. The VIRA-GCN-based 3D joint approach provided spatial cues but did not eliminate upstream pose and tracking errors. These results highlight the need for skeleton–depth fusion, robust identity tracking, occlusion handling, and enhanced joint recognition in real-world multi-person fall detection. Full article

Background/Objectives: Healthcare burnout is a pervasive problem affecting both professionals and patients. Resilience training may be especially important for healthcare teams practicing in underserved areas. This article describes the development and pilot testing of a curriculum innovation developed as part of California Area Health Education Centers’ (CA AHEC) resilience training program. Methods: A team of experts from CA AHEC and Hennepin Healthcare Systems led the development of a whole-person approach to resilience for health professionals, trainees, and other healthcare workers in medically underserved areas. A curriculum was piloted with 7 two-hour Train-the-Trainer video conferences and 1 six-hour in-person workshop. Participants (n = 44) included representatives from 12 CA AHEC centers. Attendance was tracked, and pre/post Knowledge, Skills, and Attitudes (KSA) and Mini Z Burnout Reduction surveys were administered. An engagement survey was fielded after training. Results: Attendance averaged 26 participants per session. Engagement surveys reflected high engagement, utility, and program appreciation, with KSA surveys reflecting large gains in understanding burnout mechanisms/prevention strategies (13.5% high pre- vs. 94.7% post-, p < 0.001; Absolute difference of 81.22, 95% CI 66.87–95.58), large Cohen’s h Effect Size (ES) of 1.9) and the perceived ability to use the training to cope with stress (8.11% high pre- vs. 94.7% post, p < 0.001; Absolute Difference of 86.63%, 95% CI 73.55–99.71, large Cohen’s h ES of 2.1). Mini-Z burnout scores remained relatively unchanged. Mindfulness, mind/body approaches to resilience, community support, and system change were rated as the most impactful curricular aspects. Conclusions: This whole-person resilience curriculum represents an emerging approach to healthcare burnout. While the curriculum was greatly appreciated by attendees, working conditions and burnout remained unchanged, suggesting a need for programs that address underlying systemic stressors. Full article

Background: Whether routinely available laboratory measures are associated with longitudinal bone mineral density (BMD) change in clinical practice remains uncertain, particularly when treatment exposure and other major skeletal determinants are incompletely recorded. This study examined BMD trajectories and incident osteoporosis in a Saudi dual-energy X-ray absorptiometry (DXA) cohort. Anti-osteoporosis therapy, glucocorticoid exposure and menopausal status were unavailable, limiting causal interpretation. Methods: This retrospective longitudinal cohort included DXA examinations performed at King Saud University Medical City (KSUMC), Riyadh, from 2016 to 2021. The trajectory analysis comprised 2147 patients with at least two scans and a minimum one-year interval. Diagnostic category was based on the lowest T-score across the lumbar spine, bilateral femoral necks and distal radius. Twenty linear mixed-effects (LME) models evaluated time-by-biomarker interactions with Benjamini–Hochberg correction. Cox proportional hazards (PH) regression was the primary analysis of incident osteoporosis among 789 patients without osteoporosis at baseline. Results: During 2045 person-years of observation, 107 patients developed osteoporosis (5.2 events per 100 person-years). No event occurred among 124 patients with normal baseline BMD, compared with 107 events among 665 patients with baseline osteopenia (log-rank p < 0.001). In the complete-case Cox model (n = 384; 53 events; C-index = 0.679), baseline osteopenia had a hazard ratio (HR) of 3.12 (95% Confidence interval (CI) 0.89–10.97; p = 0.076); no modelled covariate reached statistical significance. None of the 20 time-by-biomarker interactions remained significant after multiplicity correction (smallest q = 0.214). Four nominal terms with raw p < 0.10 were below the measurement-based clinical threshold and were compatible with chance variation. Conclusions: No routinely measured biomarker demonstrated clinical utility for identifying BMD trajectory in this cohort. Baseline DXA category separated the observed progression pattern, although this partly reflects the distance from the diagnostic threshold and should not be interpreted as a validated prediction model. The principal contribution is a carefully characterised null result. Prospective studies with explicit treatment tracking and repeated bone-turnover measurements are required for confirmation. Full article

The integration of digital therapeutics (DTx), wearable electronic devices, and artificial intelligence (AI) represents a significant advancement in personalized healthcare. The primary purpose of this structured narrative review is to evaluate the convergence of these technologies, providing a consolidated framework that bridges the gap between raw biometric data acquisition and actionable, AI-driven clinical insights. This paper synthesizes the latest literature on the intersection of mobile health (mHealth), machine learning (ML), and physiological tracking, with a primary focus on heart rate variability (HRV) and associated biochemical markers, such as cortisol, salivary alpha-amylase, and interleukins. Instead of viewing wearable outputs simply as raw data, we critically evaluate the technical verification and clinical validation required to define them as true “digital biomarkers.” By evaluating multimodal sensor technologies and advanced predictive algorithms, this paper outlines the clinical utility of digital biomarkers in diagnosing and proactively managing cardiovascular, neurological, metabolic, and psychiatric conditions, noting classification accuracies frequently exceeding 85% in controlled settings. However, we strongly caution that internally validated performance in controlled settings does not inherently demonstrate external clinical utility. The clinical relevance of this study lies in its holistic approach to identifying how continuous monitoring can broaden healthcare accessibility while improving precision medicine. Furthermore, it deeply addresses the technical challenges of highly variable ambulatory data quality, the necessity for robust artifact reduction (e.g., via LSTM and GAN architectures), and the limitations of small, homogeneous training datasets. We highlight the essential need for demographic-aware algorithmic models, external validation, and decentralized privacy-preserving models (e.g., federated learning) in diverse populations to ensure the safe, equitable clinical translation of DTx, mHealth, ML, and AI technologies. Full article

Smartwatches generate continuous personal health data, yet their value for self-care depends on users’ ability to engage with these data meaningfully. Using a cross-sectional survey of 838 adult smartwatch users in Taiwan, this study examined digital data engagement (DDE) as a data-specific construct linking smartwatch use to everyday self-care. DDE captured monitoring, interpretation, verification, and goal-oriented use of smartwatch-generated data, while health behavior diversity (HBD) represented the breadth of routine self-care practices across preventive, mind–body, physical activity/exercise, and restorative or expressive domains. The DDE-centered model showed that DDE was positively associated with health motivation, body awareness, and HBD. Health motivation showed the strongest association with HBD, whereas body awareness was not directly associated with HBD, suggesting that bodily cue awareness alone may not broaden self-care repertoires. HBD analysis further showed that smartwatch-enabled self-care was not merely exercise-oriented; preventive self-care had the highest item-adjusted coverage. Gender differences appeared mainly in domain composition, whereas health concern profiles were associated with both repertoire breadth and domain composition. These findings contribute a data-engagement perspective to smartwatch research and introduce HBD as a multidomain lens for evaluating everyday digital self-care beyond device access, tracking frequency, or single behavior change. Full article

This review examines the clinical and practical implications of embedding molecular profiles directly into the 2023 FIGO staging system for endometrial carcinoma, in the context of the 2025 ESGO/ESTRO/ESP guidelines. The primary purpose is to navigate a central conflict in modern oncology: how to deliver increasingly personalized care while maintaining a globally accessible, equitable, and standardized cancer classification system. The 2023 FIGO update represents a paradigm shift from the traditional dualistic model (Type I versus Type II) by allowing molecular findings to redefine stage itself. While this integration offers clear benefits, it introduces significant challenges. First, the system depends on advanced molecular testing, creating a “rich-poor” divide where patients in resource-limited settings are systematically overtreated because testing is unavailable. Second, stage becomes unstable, changing with sequential histologic and molecular re-review, which causes confusion for patients and clinicians. Third, the system lumps prognostically distinct histotypes (Serous, Clear Cell, Carcinosarcoma, and Grade 3 Endometrioid) into a single aggressive stage, obscuring meaningful differences in survival. Fourth, it relies on subjective parameters such as “substantial” lymphovascular space invasion, for which no standardized definition exists, leading to high inter-observer variability. After analyzing these controversies, the review proposes a pragmatic solution: decouple anatomical staging from molecular risk stratification. Staging should remain a purely anatomical, universally applicable descriptor of tumor extent, while molecular and histologic data are used separately within a dynamic risk assessment model, as suggested by the European guidelines. This dual-track approach preserves global comparability, reduces inequity, and maintains diagnostic stability, while still enabling personalized treatment where advanced diagnostics are available. Full article