Search Results (2,547)

Medical image segmentation is a critical task in computer-assisted diagnosis, but accurate delineation remains challenging in sensor-captured dermoscopic and endoscopic images because of low contrast, blurred boundaries, acquisition artifacts, and large appearance variations. Although CNN-based methods are effective in capturing local texture and boundary cues, they often struggle to explicitly model long-range dependencies and global structural relationships. Transformer-based architectures can capture global context, but their self-attention mechanism may become computationally costly when processing high-resolution feature maps. To address these challenges, we propose MGA-UNet, a frequency-aware multi-scale encoder–decoder segmentation framework that integrates wavelet-based frequency decomposition with Mamba-based long-range dependency modelling. Specifically, the Wavelet-Mamba feature extraction backbone (WMB) decomposes features into low- and high-frequency components to enhance boundary-aware representation, the Gated Multi-scale Aggregation Module (GMAM) aggregates parallel multi-scale encoder features and applies a content-dependent gate to the fused response, and the Adaptive Sparse Attention Module (ASAM) refines bottleneck representations with sparse attention for global semantic modelling. Across three independent runs with random seeds 42, 123, and 2026, MGA-UNet achieves mean Dice Similarity Coefficients of $88.92\pm 0.04\%$, $88.01\pm 0.07\%$, and $85.91\pm 0.04\%$ on ISIC2018, ISIC2017, and Kvasir-SEG, respectively. These results demonstrate competitive segmentation performance among the compared representative CNN-based, Transformer-based, and Mamba-based methods, including the recent H-VMUNet baseline. These results indicate that frequency-domain decomposition and state-space modelling can complement each other for accurate medical image segmentation, particularly in images with ambiguous boundaries and complex background interference. Full article

Cyber-Physical Systems (CPS) are using more AI for smart decisions and automation, but also faces new security issues. This study surveys existing CPS security assessment methodologies across healthcare, automotive, energy, and critical infrastructure, identifying their limitations in addressing emerging digital-physical threats. We find that traditional risk assessment and testing approaches, often network-centric and compliance-driven, are insufficient for AI-powered CPS. New vulnerabilities arise from the tight coupling of cyber and physical components, such as adversarial manipulation of sensors that can cause dangerous misbehavior, supply chain attacks on AI models, and the inability to patch critical devices on the fly. We use AI techniques and problem-solving methods to improve the security of CPS. The framework helps detect threats, monitor system activities, and reduce security risks in real time. It also follows important security and privacy standards such as NIST, IEC 62443, ISO 21434, and GDPR. The system continuously checks CPS operations, uses AI tools to find weaknesses, and supports security compliance. We also study real-world CPS attacks, including industrial malware, car hacking, and medical device attacks, to show the importance of the framework. In this research, we present prototype implementation and experimental evaluation along with a case study of protecting a smart manufacturing plant during a ransomware attack using the proposed approach. Full article

Artificial intelligence (AI)-based digital twins (DTs) are emerging as a groundbreaking paradigm in rehabilitation and physiotherapy, enabling the creation of dynamic virtual representations of patients for continuous monitoring, prognostic assessment and personalised therapeutic interventions. This article presents a global, interdisciplinary review of AI-based DT technologies in rehabilitation settings utilising the Internet of Medical Things (IoMT), with particular emphasis on the integration of wearable and implantable sensor systems in next-generation wireless healthcare applications. The article analyses how multimodal wearable sensors, implantable devices and smart wireless communication networks can support the acquisition of real-time biomechanical and physiological data for adaptive rehabilitation. By combining perspectives from biomedical engineering, physiotherapy, computational intelligence and wireless healthcare systems, this article highlights the emerging opportunities and challenges associated with the creation of scalable digital twin ecosystems for precision rehabilitation. The proposed vision contributes to the development of smart, connected and personalized rehabilitation infrastructures, in line with future paradigms of healthcare and wireless communication. Full article

The growing problem of air pollution and its direct impact on human health have created an urgent need for reliable, intelligent, and machine learning (ML)-enabled gas-sensing technologies. Among various sensing platforms, metal oxide gas sensors (MO-GSs) have emerged as promising candidates owing to their low cost, high sensitivity, and scalability. However, their practical application is limited by poor selectivity, cross-sensitivity, sensor drift, and high operating temperatures. Recent advances in ML have provided effective strategies to overcome these limitations through data-driven optimization of sensing performance. This review summarizes recent progress in ML-assisted MO-GSs, covering sensor array design, feature engineering, and classification algorithms, including support vector machines (SVMs), random forests (RFs), and deep neural networks (DNNs). In addition, key data-processing techniques such as preprocessing, dimensionality reduction, and hybrid learning approaches are critically discussed. The application of ML-enabled MO-GSs in medical diagnostics, environmental monitoring, industrial safety, and food quality assessment is also reviewed. Despite significant progress, challenges including limited dataset availability, sensor drift, and poor model generalization remain. Future research should focus on developing adaptive, energy-efficient, and IoT-enabled smart sensing systems. The integration of machine learning with metal oxide gas sensors represents a significant step toward intelligent, next-generation, high-performance gas-sensing technologies. Full article

There is a growing trend toward employing two-coil and three-coil systems in magnetically coupled resonance wireless power-transfer (MCR-WPT) technology. However, these configurations have limitations at longer transmission distances and are sensitive to load variations. This paper provides a comprehensive review of four-coil WPT systems, focusing on their design, optimization, and applications. The reviewed literature indicates that four-coil configurations generally maintain higher power-transfer efficiency (PTE) over longer transmission distances, exhibit greater tolerance to misalignment, and show reduced sensitivity to load variations. The literature review indicates that while symmetric designs are easier to analyze, asymmetric configurations may provide higher efficiency and extended transmission range. The review also discusses how four-coil technology can be used in medical implants, sensors, and consumer electronics. It also addresses the current challenges of power and compliance in the electric vehicle (EV) industry. Finally, this paper identifies important areas of research that need to be addressed. Future research should aim to optimize power and efficiency together, explore multi-receiver systems for public charging, and investigate the potential of four-coil technology for Electric Micromobility Vehicles (EMVs). Full article

IoT-connected healthcare sensor networks require authenticated and privacy-preserving data exchange among wearable sensors, mobile medical terminals, edge gateways, cloud servers, and medical institutions. Existing authentication schemes for healthcare IoT often bind signatures directly to user identities, exposing sensitive personal or institutional information and imposing heavy verification costs on resource-constrained sensing devices. To address this problem, we propose a fine-grained and flexible dual authentication scheme for healthcare sensor networks. In the proposed scheme, health data and diagnoses are signed with a fresh signing key and a fine-grained access control policy each time, so that the signer identity remains hidden while authorized entities can still modify permitted parts of signed data. No entity other than an authorized entity can trace a malicious signer or modify signed data without changing the data source. To support lightweight verification in sensor-edge-cloud deployments, we further present a verifiable outsourced authentication scheme that outsources time-consuming pairing operations to cloud servers; the online verification process then requires only six multiplication operations. As a fundamental technical component, we present a practical attribute-based sanitizable signature with shorter signature and key lengths and more efficient signing and signature-changing operations than the state-of-the-art policy-based sanitizable signature (P3S). Formal security analysis and experiments demonstrate the security and practicality of the proposed scheme for privacy-preserving healthcare sensing and medical data exchange. Full article

Taiwan has one of the highest dialysis prevalences worldwide, making safe and reliable hemodialysis monitoring a critical sensor-based healthcare challenge. Modern hemodialysis machines integrate heterogeneous multimodal sensors (pressure, flow, conductivity, temperature, and cardiovascular signals), but differences in machine brands, data formats, and privacy constraints hinder centralized learning and robust complication prediction. This work proposes a Medical IoT-oriented federated learning framework, PSOFed-HD, that performs dual-layer Particle Swarm Optimization (PSO) to enhance heterogeneous sensor fusion for predicting dialysis-related hypotension and discomfort events. The events are defined as abnormal blood-pressure states, defined as systolic blood pressure <90 mmHg. Each hemodialysis machine is paired with an edge gateway acting as an FL client, where local PSO optimizes CNN feature weights over non-IID sensor subsets, while the central server applies PSO-driven aggregation to adaptively weight client models according to validation performance. Experiments on real-world hemodialysis datasets with 17 most commonly seen HD physiological features demonstrate that standard FedAvg yields an accuracy of 65.24% and F1-score of 0.5318, server-side PSO improves accuracy to 75.11%, and client-side PSO further raises accuracy to 81.97%. The proposed dual-layer PSO framework achieves the best performance, with 90.56% accuracy and an F1-score of 0.8533, along with superior ROC characteristics (AUC = 0.908) and stable cross-validation across 11 folds. State-of-the-art federated learning techniques for non-IID data such as SCAFFOLD and FedProx are also examined using the same heterogeneous HD dataset. The performance is close to our client-only PSO techniques, proving the merits of our dual-layer PSO architecture. These results confirm that jointly optimizing local feature representations and global aggregation weights enables effective fusion of heterogeneous hemodialysis sensor data under privacy-preserving Medical IoT constraints, providing a practical decision-support approach for real-time complication prediction in dialysis units. Future work will incorporate temporal models such as LSTM or Transformer architectures to achieve early event prediction. Full article

The Timed Up and Go (TUG) test is widely used to assess functional mobility in Parkinson’s disease (PD), although total test duration may overlook phase-specific biomechanical alterations. This cross-sectional study investigated whether phase-specific analysis of the instrumented TUG (iTUG) could identify candidate biomechanical markers characterizing differences among early-stage PD subgroups and healthy controls. Seventy-nine participants (38 PD, 41 controls) performed four iTUG trials in the OFF-medication state. Movement data were collected using synchronized inertial measurement units and optoelectronic motion capture systems. The iTUG was segmented into six phases, and temporal, spatiotemporal, variability, and multisegmental kinematic parameters were analyzed. Participants with PD at modified Hoehn and Yahr (mH&Y) stage 2 performed the iTUG more slowly than controls (p < 0.001), mainly due to impairments during walking (p = 0.012) and turning. Turning was the only phase that distinguished controls from individuals with PD at mH&Y stages 1–1.5 (p = 0.015), who also showed increased elbow asymmetry (p = 0.009). Participants with PD at mH&Y stage 2 exhibited broader differences, including increased double-support time, shorter stride length, lower walking speed, reduced frontal-plane control, and reduced trunk, elbow, hip, and ankle motion, whereas gait variability did not differ significantly between groups. Sagittal trunk ROM on the less affected side was the only variable that significantly differed between the two PD subgroups. These findings suggest that phase-specific iTUG analysis may reveal candidate biomechanical markers associated with mH&Y stage and functional mobility impairment and support sensor-based assessment for objective characterization of functional mobility in PD. Full article

The Internet of Medical Things (IoMT) is a promising technology for enabling smart and efficient healthcare systems through continuous physiological monitoring, early anomaly detection, and proactive patient management. However, IoMT sensors are vulnerable to False Data Injection Attacks (FDIAs), in which adversaries can manipulate sensor measurements to compromise diagnostic accuracy, mislead clinical decision-making, and threaten patient safety. Existing detection approaches often rely on population-level statistical models that may not fully capture individual physiological variations or residual-based thresholds designed for relatively simple attack scenarios, limiting their ability to exploit the spatio-temporal dependencies of multi-sensor physiological streams and detect stealthy or adversarial FDIAs. This paper proposes a patient-specific FDIA detection framework based on a Graph Convolutional Network–Gated Recurrent Unit (GCN–GRU) digital twin that learns an individual patient’s normal physiological behaviour from clean baseline telemetry. The trained digital twin is integrated into a Kalman filter as the state prediction model, and the resulting standardised innovation residuals are used as detection features. To characterise stealthy attack behaviours, four complementary window-based feature groups are extracted from the innovation sequence: innovation statistics, sensor correlation drift, temporal smoothness, and uncertainty mismatch. A CNN-1D classifier is then trained to learn discriminative temporal attack patterns from these features for accurate detection. A structured attack taxonomy comprising five stealthy and adversarial FDIA scenarios is developed, where attacks are injected as smooth gradual or abrupt coordinated modifications to sensor measurements while remaining within plausible physiological ranges. Experiments conducted on the WUSTL-EHMS-2020 benchmark dataset demonstrate that the proposed framework achieves an F1-score of 94.3%, outperforming Isolation Forest and PCA Reconstruction by 34 percentage points. Furthermore, the proposed framework reduces the false alarm rate to 3.6%, compared with 35.1% and 9.2% achieved by Isolation Forest and PCA Reconstruction, respectively. These results demonstrate the effectiveness of the proposed framework for reliable detection of stealthy FDIAs in IoMT-based healthcare systems. Full article

Flexible piezoresistive sensors are highly desirable for wearable health monitoring, yet balancing ultrahigh sensitivity and wide pressure detection range is a major bottleneck restricting their applications in electronic skin and soft robots. This work constructs a hierarchical piezoresistive sensor through a simple three-step fabrication: electrospinning PVDF/PAN fiber networks, polydopamine (PDA) surface modification, and in situ polypyrrole (PPy) polymerization for conductive sensing layers. As a dual-function interlayer, PDA forms hydrogen and covalent bonds with PPy to yield uniform, firm conductive coatings. The link between PPy morphology and sensing performance is clarified by regulating polymerization parameters. At a pyrrole concentration of 3 g/L, the optimized sensor achieves a high sensitivity of 220.88 kPa⁻¹ (0–10 kPa) and stable linear signals up to 1 MPa, with superior cycling durability over 5000 cycles and good biocompatibility. This scalable fabrication resolves the sensitivity–range tradeoff, promising wearable medical monitoring and human–machine interaction devices. Full article

Durable medical devices (DMD) are essential healthcare assets, yet their management in public hospitals is constrained by fragmentation, limited traceability, reactive maintenance, and weak lifecycle integration. This study proposes a framework for digitally enabled, sustainable, and circular DMD management. A mixed-methods design integrated a literature review, expert consultation using the Best–Worst Method, weighted technology nominations, and case-based process mapping in Brazilian hospitals. Eleven experts assessed the criteria guiding Industry 4.0 technology selection for DMD management and the technologies best responding to these priorities; nine consistent judgments were aggregated. Patient-Centered Care, Operational Efficiency, and Resource Efficiency and Cost Reduction emerged as the leading influences on technology selection. Big Data and Analytics, Artificial Intelligence, the Internet of Things, Cloud Computing, Cyber-Physical Systems, Smart Sensors, and Machine Learning formed the priority portfolio, accounting for 84% of the weighted score. The cases contextualized these priorities by revealing discontinuous information flows, limited asset visibility, corrective maintenance, fragmented governance, and weak end-of-life practices. By connecting decision priorities and technological capabilities with observed gaps, the TO-BE framework organizes sustainable procurement, traceable use, predictive maintenance, redeployment, refurbishment, and responsible disposal through material and information flows, providing a pathway for digital and circular transformation in resource-constrained healthcare systems. Full article

The growing demand for wearable and Internet of Medical Things (IoMT) devices is driving the development of compact, low-power platforms for continuous physiological monitoring. Bioimpedance analysis represents a versatile non-invasive technique for the assessment of tissue properties, body composition, and respiratory dynamics. This work presents a comprehensive experimental validation of a compact bioimpedance measurement platform based on the SENSIPLUS chip, a CMOS sensor interface integrating a frequency-programmable lock-in amplifier for Electrochemical Impedance Spectroscopy in the 10 kHz–1 MHz range. The platform was validated at three complementary levels: (i) electrical characterization on Debye tissue-equivalent circuits using a three-point bilinear calibration, with analysis of the electrode–skin contribution and repeatability assessment; (ii) in vivo multi-frequency bioimpedance spectroscopy (BIS) with Cole–Cole model fitting and hook-effect correction; and (iii) single-frequency thoracic impedance plethysmography for respiratory monitoring. Results were compared against an Agilent E4980A precision Inductance (L), Capacitance (C), and Resistance (R) meter and a calibrated spirometer. The presented device achieved a maximum resistance error below 5.7% and reactance deviation under 6 $\mathsf{\Omega }$ across the investigated frequency range, Cole–Cole parameters consistent with reference values, and strong linear correlation ($R^2=0.97$) between thoracic impedance variations and tidal volume, with respiratory rate estimation errors below 2% across the ten sessions, specifically 1.43% during normal breathing and 1.96% during deep breathing. These results demonstrate that the SENSIPLUS-based platform achieves metrological performance compatible with the requirements of wearable IoMT applications, here demonstrated in a single-subject proof-of-concept study, while relying for all critical analog functions on a compact ($1.5\times 1.5$) mm² system-on-chip with low power consumption (1.5 mW). Full article

Raman spectroscopy has emerged as a powerful tool for gas detection due to its label-free operation, molecular specificity, and multi-component analysis capabilities. However, its widespread application is hindered by limited sensitivity, particularly for trace gas analysis. To overcome this challenge, this study introduced an effective Raman spectroscopy detection system that synergistically combines a 532 nm high-power laser with a near-concentric multipass cell (MPC), enabling dual enhancement of both Raman excitation and signal collection. We simultaneously determined three critical performance metrics, including gas Raman signal intensity, signal-to-noise ratio (SNR), and limit of detection (LOD). Under the optimized experimental conditions, the CO₂ Raman signal reached an SNR of approximately 58 for laboratory air containing 916 ppm CO₂, corresponding to a concentration-equivalent detection limit of 48 ppm according to the 3σ criterion. Notably, the intensity of the generated Raman scattering signal is proportional to the average power. The intensity of Raman signals at different wave numbers increases at different rates with the increase in the average excitation power. The system achieves a remarkable LOD of 48 ppm for CO₂, representing an advancement over conventional Raman gas sensors. This work validates high-power near-concentric cavity-enhanced Raman spectroscopy as a reliable method for trace gas detection, with potential implications for multi-component gas analyzers in environmental monitoring, industrial safety, and medical diagnostics. Full article

High-performance ammonia (NH₃) sensors play a critical role in environmental protection and noninvasive medical diagnosis. This work reports a new NH₃ sensor based on Au-microsphere-decorated polyaniline (PANI) manufactured by a precise in situ inkjet printing method on a MEMS micro-hotplate. The in situ oxidative polymerization of aniline was performed directly on the MEMS platform using AuCl₃ as a bifunctional oxidant and precursor, with a hierarchical morphology of microspheres (~750 nm) and nanorods (~250 nm). Reduced from Au³⁺ in the polymerization reaction, Au microparticles achieve substantial catalytic promotion by virtue of chemical sensitization and spillover effect. The optimized Au–PANI MEMS sensor exhibits a superlative response of 201% toward 1 ppm NH₃ at room temperature, with an ultra-low theoretical limit of detection (LOD) of 0.42 ppb. Furthermore, the device demonstrates rapid response/recovery kinetics (76 s/72 s), exceptional selectivity against common interfering gases (SO₂, CO, H₂, etc.), and robust long-term stability with high response retention over two months. This research provides a scalable, cost-effective strategy for the mass production of miniaturized, high-sensitivity gas sensors for industrial and healthcare applications. Full article

Microelectromechanical systems (MEMS) gyroscopes are now standard rate sensors in inertial navigation, automotive electronics, industrial automation, and medical instrumentation because they are inexpensive, compact, and readily integrated. Their accuracy, however, degrades with temperature: damping and quadrature coupling change, and readout-electronics behavior shifts, producing temperature-dependent zero-rate-output drift, elevated random noise, and poorer long-term stability. Hardware- and structure-based temperature compensation address part of the problem but carry cost and generality penalties, which has moved recent work toward data-driven software-based temperature-drift compensation. This review focuses on the fusion algorithms that have come to dominate that literature, organized as a four-stage pipeline: signal decomposition, learning-based drift modeling, adaptive filtering, and signal reconstruction. We examine how optimizer-tuned variational mode decomposition and improved empirical-mode-decomposition variants separate temperature-related components from noise; how deep temporal networks and optimizer-coupled learners model the nonlinear, time-lagged drift; and how adaptive Kalman variants and time-frequency filtering reconstruct a stable output. We close by identifying four open problems that recur across the recent gyroscope-specific work—cross-device generalization, temperature hysteresis, embedded real-time deployment, and physics-informed lightweight modeling. Full article