Enhancing sEMG recognition with dynamic asymmetric training in context-informed virtual–real fusion system

The evolution of information technology has positioned multimedia content as a pillar of digital communication, but at the same time, it has opened a gap for serious threats in the form of deepfakes. This highly realistic media manipulation challenges information authenticity, privacy, and cybersecurity, which, for Information Technology professionals, presents both technical and ethical challenges. This Systematic Literature Review (SLR) aims to map the development of Artificial Intelligence based algorithms in deepfake detection. Using the PRISMA methodology on 20 selected primary articles (2021-2025), this study aims to identify trends in the use of AI algorithms for deepfake detection, determine the most effective approaches, and analyze the factors contributing to their effectiveness. The analysis results show a paradigm shift from single models (such as CNN) to hybrid architectures (CNN-LSTM-Transformer) and complex multimodal fusion systems. It was found that hybrid algorithms are the closest approach to best practice due to their ability to handle spatial and temporal dimensions simultaneously. Key contributing factors include hierarchical feature extraction, generative data augmentation, and the integration of Explainable AI (XAI).

Delamination is a major failure Mode in laminated composites, typically triggered by premature interlaminar matrix cracking and leading to severe structural degradation. To address this, various through-thickness reinforcement strategies have been explored, including three-dimensional woven architecture. Although these designs significantly improve delamination resistance, their industrial adoption stays limited due to reproducibility challenges and the high cost and operational complexity of advanced manufacturing systems needed for controlled through-thickness reinforcement. This study investigates an alternative interlaminar reinforcement method, through-thickness stitching, aimed at enhancing Mode-I delamination resistance of a commercial fiberglass laminate without changing its native architecture. Composites were manufactured using a low-viscosity epoxy infusion system (MAX 1618 A/B) and a [0/90] biaxial fiberglass fabric. An eight-filament polyethylene thread (Ø = 0.12 mm) was introduced in predefined stitch architectures consisting of three longitudinal patterns having two, three, and five continuous stitch lines, referred to as AV, BV and CV samples, respectively. Results show that stitching highly increases Mode-I interlaminar fracture toughness GIC by 0.3808, 0.4152 and 0.5192 kJ/m2 for AV, BV and CV respectively, compared to 0.0265 kJ/m2 for the unstitched composite O, highlighting the strong influence of stitch orientation and spacing on interlaminar performance. But scanning electron microscopy revealed added failure mechanisms in stitched specimens, including localized fiber misalignment of up to 33° and resin-rich regions approximately 0.6 mm in length, suggesting that while stitching enhances delamination resistance, it may also influence other mechanical properties.

In urban environments, autonomous vehicles face critical challenges in localization and perception under extreme lighting conditions, including rapid illumination changes, high contrast, and nighttime low-light scenarios. To address the performance degradation of traditional LiDAR-inertial odometry systems under such conditions, this study proposes a high-precision FAST-LIO2-EC algorithm that fuses event cameras into the FAST-LIO2 framework. Event cameras, with their microsecond temporal resolution and 140 dB dynamic range, provide asynchronous edge information that complements LiDAR point clouds and IMU measurements. We validate the proposed system through real-world road tests conducted on public roads and closed test tracks, covering three typical extreme lighting scenarios: tunnel entrance/exit transitions, high-contrast shadow boundaries, and nighttime sparse-lighting conditions. The experimental platform is equipped with a 32-beam LiDAR, a 6-axis IMU, a DVS event camera, and an RTK-GNSS system for ground truth trajectory acquisition. Real-world results demonstrate that the FAST-LIO2-EC system achieves significant improvements in localization accuracy and robustness. In illumination change scenarios, the Absolute Trajectory Error (ATE) is reduced by 32.5% compared to the baseline FAST-LIO2 system, with zero tracking loss events. The point cloud quality is substantially enhanced, with more uniform distribution and clearer obstacle boundaries. In high-contrast scenarios, both systems maintain comparable performance with ATE below 0.15 m. However, in nighttime scenarios, the fusion system shows moderate improvement (15.3% ATE reduction) but reveals sensitivity to event camera noise, indicating the need for adaptive thresholding strategies. Supplementary simulation experiments validate the system’s robustness under varying speeds and sensor noise levels. This work provides a practical solution for autonomous vehicle deployment in complex urban lighting environments, with a comprehensive analysis of real-world performance boundaries and deployment considerations.

Blood pressure is closely monitored during anaesthesia, yet the optimal intraoperative target remains uncertain. This narrative review synthesizes contemporary observational and randomized evidence and explores emerging strategies for individualized haemodynamic management. We reviewed major observational cohort studies, randomized controlled trials (RCTs), consensus statements, and recent technological developments addressing intraoperative hypotension, MAP thresholds, and strategies to prevent perioperative organ injury in adult noncardiac surgery. Large observational datasets consistently demonstrate graded, duration-dependent associations between intraoperative MAP 60-70 mmHg and postoperative myocardial injury, acute kidney injury, and mortality . These findings have informed international recommendations to avoid MAP below 60-65 mmHg. However, contemporary multicentre RCTs enrolling more than 13,000 patients show that targeting higher or individualized MAP thresholds does not improve patient-centred outcomes compared with routine care (typically MAP ≥ 65 mmHg) . Only one small trial reported benefit with individualized systolic targets. Emerging evidence suggests that hypotension reflects heterogeneous haemodynamic endotypes (vasodilation, hypovolaemia, myocardial depression, bradycardia), potentially explaining why uniform pressure targets fail to improve outcomes. Continuous blood pressure monitoring, proactive norepinephrine infusion, predictive analytics, and closed-loop vasopressor systems reliably reduce hypotension exposure, although definitive outcome benefits remain unproven. Observational and randomized data are concordant: MAP ≥ 60-65 mmHg appears sufficient for most noncardiac surgical patients. Future progress will likely depend on mechanistic endotyping, integration of advanced monitoring, and precision-guided haemodynamic strategies rather than escalation of universal MAP targets alone.

This work introduces a new benchmark problem for calculating shutdown dose rates (SDDRs) aimed at fusion reactor applications. The model is designed to represent a simplified version of a typical ITER port plug. The responses of interest include neutron flux, gamma flux, and gamma SDDR at 12 different locations scattered throughout the port. This article outlines the geometry specifications of the problem, provides material definitions for the components, specifies the required responses to be calculated, and presents the source definition information. The need for this benchmark arises from the limited availability of publicly accessible references, with only one benchmark representing the typical dimensions and materials found in fusion systems. This existing benchmark has been cited extensively, reflecting the demand within the scientific community to test both established and novel workflows for SDDR calculations. However, since its presentation at a conference in 2011, the results have become increasingly well known. Moreover, the absence of formal publication and peer review has led to the details of this benchmark being extracted from secondary sources, such as subsequent studies that reference it. As a result, analysts are left with significant flexibility in interpreting the key parameters, which can be adjusted to account for unknown systematic errors, ultimately reproducing the already well-known responses. This new benchmark serves as an updated version of that earlier work, with the aim of providing a more reliable description of the materials and their impurities, which is crucial for assessing activation and subsequent gamma emission. Additionally, it seeks to provide a geometry that more closely represents an ITER port plug. The improvements in the problem definition will lead to a more reproducible benchmark problem, while also presenting the radiation transport community with a completely new challenge. The results will be published in a future article to allow analysts adequate time to analyze this problem independently.

Diabetes summer camp is a demanding, although joyful, setting with many unpredictable activities affecting glycemic control. Recent technological advances, such as automated insulin delivery (AID) systems, offer promising real-world benefits. Our aim was to evaluate the efficacy and safety of continuous subcutaneous insulin infusion (CSII) systems during a diabetes summer camp and compare the performance of the MiniMed 780G (AID) system with the sensor-augmented MiniMed 640G system. This is a retrospective, observational study collecting data from six summer camp weeks organized from 2019 to 2025 in Northern Greece. Children, adolescents, and adult staff with type 1 diabetes (T1D) using MiniMed insulin pumps and continuous glucose monitoring were included. Glycemic metrics (time in range, glucose coefficient of variation, time in hypoglycemia/hyperglycemia, glycemia risk index) were collected from CareLink platform across three weekly periods: during camp, pre-camp, and post-camp. Data from 93 participants/year (67 females,72.04%) were included. Mean time in range (TIR) during camp was 72.72%, with best outcomes in years 2023 to 2025 (TIR > 77%). Across all periods, MiniMed 780G users demonstrated markedly superior outcomes compared to 640G users: during the camp week, TIR was 78.68% vs 62.83% (P < .001), and post-camp TIR remained higher (70.02% vs 55.43%, P < .001), with lower time in hyperglycemia >180 mg/dL (22.67% vs 30.71%, P < .001). Camp weeks were associated with improved TIR and reduced hyperglycemia overall without increased hypoglycemia rates. Diabetes camps promote satisfactory glycemic control in youth with T1D, particularly when using AID systems. MiniMed 780G users maintained better outcomes even the week after camp compared to MiniMed 640G users.

The study introduces CrossMF as a unified transformer-based model which performs emotion recognition from speech and text data. CrossMF combines dynamic attention between modalities with a memory enhancement system for effective fusion between textual and acoustic information. The training of CrossMF involves a two-step tactic where the acoustic encoder receives clean audio inputs from Toronto emotional speech set (TESS) for optimization yet the text encoder and fusion module acquire training from audio-text pairs from the Multimodal EmotionLines Dataset (MELD) dataset. The system allows emotion predictions from three combination types including audio-only, text-only and audio with text inputs due to adjustable modality access. The evaluation process takes place across the three different input scenarios to show that the system performs well with high generalization capability in both laboratory-recorded and natural conversational conditions. Fusion between multiple sources only occurs when both inputs are available so that integrity can be maintained until integration occurs. The model reaches a maximum validation accuracy of 97.68% while demonstrating sustained high-test performance which proves its effectiveness when operated in various conditions without depending on manually created features. The architecture also supports future extensions, allowing developers to easily incorporate ablation studies and adaptive training strategies for real-world emotion-aware systems.

Dempster-Shafer Theory (DST) and Dezert-Smarandache Theory (DSmT) are prominent frameworks within evidence theory for managing uncertainty and fusing information. While both utilize basic belief assignments (BBAs) on sets of hypotheses, they diverge critically in their foundational mathematical assumptions and conflict handling. This article provides a comprehensive comparison, detailing the philosophical and practical shift from DST to its generalized superset, DSmT. DST is built upon the power set 2Θ of a mutually exclusive frame of discernment, forcing conflict to be resolved via global normalization (Dempster's Rule), which can lead to counter-intuitive results in high-conflict scenarios (e.g., Zadeh's paradox). DSmT, conversely, operates on the hyper-power set DΘ of an exhaustive but non-exclusive frame, allowing for the representation of vague, paradoxical, or overlapping concepts (e.g., A∩B). Crucially, DSmT employs the Proportional Conflict Redistribution (PCR) family of rules, which resolves disagreement by locally redistributing conflicting mass back only to the propositions that generated it. This mechanism ensures stability and interpretability, even under extreme conflict. We analyze the trade-offs in expressiveness and computational cost, illustrating that the choice between the two theories hinges on two core factors: the expected level of conflict and the degree of conceptual overlap in the problem domain. Ultimately, DSmT is demonstrated to fully encompass DST, functioning as a robust, flexible alternative for modern multi-sensor fusion, intelligence analysis, and decision-making systems characterized by ambiguity and high disagreement.

Abstract With respect to considerations of nuclear fusion systems, since the electromagnetic field in the wide frequency range from the static magnetic field to extremely high frequency was utilized for the plasma confinement, heating at the magnetic confinement fusion facilities, not only ionizing radiation but also non-ionizing radiation should be investigated and discussed in terms of occupational protection as one of the issues in fusion facilities. To clarify the behavior of non-ionizing radiation, leakage from electric and magnetic fields has been observed at the Large Helical Device (LHD), because a strong magnetic field generated by large superconducting magnet coils for plasma confinement, in addition to high-frequency and high-power oscillators for ion cyclotron range of frequencies (ICRF) and electron cyclotron resonance heating (ECRH) for plasma heating, are used. In the LHD experiments, magnetic field leakage has been observed at extremely low frequencies (ELFs) around a power supply system for superconducting magnet coils. The magnetic field outside the experimental hall and the intensity of electric field leakage from the oscillators for ICRF and ECRH are significantly below the International Commission on Non-Ionizing Radiation Protection (ICNIRP) guidelines and radio frequency exposure protection standards in Japan (RCR-STD38). The ICNIRP and RCR-STD38 provide guidelines for limiting exposure to electric and magnetic fields to protect against established adverse health effects. The guidelines for safety management to protect the working people in fusion facilities should be made clearer with respect to occupational protection from non-ionizing radiation.

Abstract Time synchronization is a key step in Simultaneous Localization and Mapping (SLAM) and multi-source sensor data fusion. Current time synchronization methods rely on professional-grade devices with high-precision time synchronization interfaces, which are often unavailable in consumer-grade sensors. To address this limitation, this paper proposes a high-precision time synchronization method for camera and LiDAR based on feature matching of MEMS gyroscope time series (SFM-Sync). By constructing integrated sensor units that combine MEMS gyroscopes with consumer-grade cameras or LiDARs, the method collects data in parallel, extracts peak features from gyroscope signals, and performs template matching to achieve accurate synchronization without dedicated hardware interfaces. Experimental results demonstrate that SGM-Sync effectively enables high-precision multi-sensor time synchronization, reducing both technical barriers and hardware dependency while offering a practical solution for multi-sensor fusion systems.

This study evaluated the optimal depth of anesthesia guided by closed-loop target-controlled infusion (TCI) in preschool-aged children undergoing dental surgery. Dental procedures in this population are frequently associated with heightened anxiety and procedural distress, often necessitating general anesthesia; however, the appropriate depth of anesthesia remains insufficiently defined. A randomized, double-blind, controlled trial was conducted involving 60 children aged 3 to 6 years scheduled for dental surgery. Participants were allocated to 3 groups (A, B, and C), with bispectral index (BIS) values maintained at 50, 55, and 60, respectively, using a closed-loop TCI system. The primary outcomes were heart rate (HR) and mean arterial pressure (MAP), recorded at 7 intraoperative time points. Secondary outcomes included surgical duration, time to first eye opening (defined as eyelid movement in response to verbal command without implying full orientation), extubation time, discharge time, and total propofol dose. No statistically significant differences were observed among groups in HR or MAP at any intraoperative time point (between-subject effect: HR, F=2.73, P=0.074; MAP, F=1.54, P=0.222). Surgical duration (F=0.521, P=0.596) and total propofol dosage (P=0.165) also did not differ significantly. Recovery parameters demonstrated significant group differences. Time to first eye opening was 37.75±8.81 minutes in group A, 28.35±2.45 minutes in group B, and 20.70±5.21 minutes in group C (P<0.001). Extubation time was 39.40±8.08 minutes, 29.60±2.58 minutes, and 22.70±5.63 minutes in groups A, B, and C, respectively (P<0.001). Discharge time was shortest in group C (135.6±30.9min), followed by group B (160.5±25.3min), and longest in group A (202.5±26.5min) (F=29.77, P<0.001). The incidence of adverse events was comparable across groups, and no serious adverse events occurred. Maintaining a BIS value of 60 (±5) with a closed-loop infusion system provided adequate anesthetic depth for dental procedures in preschool-aged children, while significantly reducing recovery and discharge times. This anesthetic strategy may support enhanced recovery and improve perioperative efficiency in pediatric populations.

Thermal Analysis and Simulation of Key Parameters in Target Design for Laboratory Fusion Systems

A Knowledge-Driven dual trigger mechanism for enhancing model security in multi-Source fusion systems

Role of Implantable Intrathecal Pumps in Refractory Cancer Pain: Systematic Review and Meta-Analysis.

Implementation of fusion energy requires processing the deuterium-tritium (D-T) mixture used to fuel the reaction, and separation of hydrogen isotopes from other gases is imperative. Specifically, the separation of hydrogen isotopes from helium is a matter of importance to the fusion fuel cycle community. Initial testing with a palladium-silver (Pd-Ag) membrane indicates that even moderate vacuum (~100 torr permeate pressure) can provide a high degree of separation (>90%) at a high ratio of H2 to He. Given the presence of He in many fusion systems, a high technology readiness level (TRL) for Q2/He (where Q represents any isotope of hydrogen) separations is needed. This study demonstrates the efficacy of H2 removal from He via permeation and potential applications for direct internal recycle. Modeling will accompany the experimental campaign to generate a predictive capability and quantify the separation performance. Modeling from previous hydrogen permeation studies has demonstrated that the typical Sieverts’ law fails to predict the measured permeation rates at high hydrogen fluxes. Existing models are being refined to integrate the effects of surface phenomena into permeation predictions, which have been expanded to account for mixtures with large ranges of Q2 concentrations. These data will improve the TRL of permeators as a separation technology for the fusion fuel cycle.

Multi-stage Kalman filtering system for sensor fusion integrated with MoS2 memtransistor featuring 1024 conductance levels

To address the semantic gap in physical sensor data for fault diagnosis of heavy-duty railway maintenance machinery and the underuse of semantic information in maintenance logs, this study proposes a model that treats fault-related text as a virtual semantic sensor. The goal is to explore a semantic-aware approach to fault diagnosis and its role in multisensor fusion. A classification model combining a BERT pretrained model with a convolutional neural network (BERT-CNN) was built. To improve the focus on key semantic units and strengthen links between textual features and sensor modalities, a dual self-attention (DSA) mechanism was added, forming the BERT-DSA-CNN model. It extracts structured semantic feature vectors from unstructured logs, which serve as outputs of the virtual semantic sensor. Experiments show that (1) incorporating DSA significantly increases performance, with BERT-DSA-CNN and Word2vec-DSA-CNN outperforming baselines (BERT-CNN and Word2vec-CNN) in terms of accuracy, precision, recall, and F1-score; (2) BERT's contextual embeddings clearly surpass Word2vec, as BERT-DSA-CNN consistently outperforms Word2vec-DSA-CNN; (3) CNN effectively captures local features of short fault texts, as BERT-CNN outperforms BERT-BiLSTM on most metrics; and (4) deep semantic feature learning substantially outperforms traditional machine learning, confirming the superiority of deep semantic feature learning. This study validates that the proposed semantic-aware model can efficiently transform fault texts into semantic features for identification. More importantly, the structured semantic features extracted by this model have the potential to be fused with physical sensor data in future work, which could provide a foundation for more accurate, robust, and interpretable intelligent fault diagnosis systems for heavy-duty railway maintenance machinery.

Artificial intelligence (AI) is redefining the precision of surgical wound closure through innovations in deep learning, computer vision, and robotic control. These technologies enable automated wound assessment, trajectory planning, and suturing assistance, with applications emerging across the United States, the United Kingdom, Japan, and India. Yet, the art of closure balancing tension, perfusion, and tissue integrity remains a fundamentally human skill. This work additionally highlights the current gap between theoretical AI capabilities and clinical validation, clarifies tissue-specific and interpretability challenges, and outlines concrete research pathways, including prospective trials, interpretable AI, and multimodal fusion systems. This letter discusses the clinical and technical aspects of AI-guided wound closure, emphasizing the importance of maintaining surgeon oversight. It argues that AI can enhance, but not supplant, human judgment, aligning with international guidelines on transparency and safety in AI deployment. Widespread clinical adoption will require rigorous validation, equitable access, and adherence to frameworks such as the TITAN Guidelines to ensure ethical and effective integration into surgical practice.

The CRISPR/Cas9 system is a cornerstone technology in genome editing. Delivery of pre-assembled Cas9 ribonucleoprotein (RNP) complexes exhibits distinct advantages, including reduced off-target effects and lower immunogenicity. Conventional methods for purifying Cas9 protein typically involve multi-step chromatography and the cleavage of fusion tag, which are time-consuming and result in diminished yields. In this study, we present a simplified, one-step purification strategy for functional Streptococcus pyogenes Cas9 (SpCas9) using the ubiquitin (Ub) fusion system in Escherichia coli. The N-terminal Ub fusion not only improves protein solubility but also facilitates high-yield production of the His-Ub-Cas9 fusion protein. Importantly, the Ub tag does not require proteolytic removal during purification, allowing direct one-step purification of the fusion protein via nickel-affinity chromatography. The purified His-Ub-Cas9 retains robust DNA cleavage activity in vivo, as validated in zebrafish embryos. This protocol greatly simplifies the production of functional Cas9 protein, facilitating its broad application in genome editing.Key features• The Ub fusion system enables single-step purification of Cas9 in E. coli using Ni-NTA chromatography, eliminating the protease cleavage step.• This method yields over 8 mg/L of high purity (>95%), functional Cas9 protein, suitable for direct use in RNP complex assembly.• The purified His-Ub-Cas9 maintains high genome editing activity in vivo, as demonstrated in zebrafish embryos.