Search Results (1,390)

In-band full-duplex (IBFD) communication systems offer a promising means of improving spectral efficiency by enabling simultaneous transmission and reception on the same frequency channel. Despite this advantage, self-interference (SI) remains a major challenge to their practical deployment. Among the different SI cancellation (SIC) techniques, this paper focuses on digital SIC methodologies tailored for multiple-input multiple-output (MIMO) wireless transceivers operating under digital beamforming architectures. Two distinct digital SIC approaches are evaluated, employing a generalized memory polynomial (GMP) model augmented with Itô–Hermite polynomial basis functions and a phase-normalized neural network (PNN) to effectively model the nonlinearities and memory effects introduced by transmitter and receiver hardware impairments. The robustness of the SIC is further evaluated under both single off-line training and closed-loop real-time adaptation, employing estimation techniques such as least squares (LS), least mean squares (LMS), and fast Kalman (FK) for model coefficient estimation. The performance of the proposed digital SIC techniques is evaluated through detailed simulations that incorporate realistic power amplifier (PA) characteristics, channel conditions, and high-order modulation schemes. Metrics such as error vector magnitude (EVM) and total bit error rate (BER) are used to assess the quality of the received signal after SIC under different signal-to-interference ratio (SIR) and signal-to-noise ratio (SNR) conditions. The results show that, for time-variant scenarios, a low-complexity adaptive SIC can be realized using a GMP model with FK parameter estimation. However, in time-invariant scenarios, an open-loop SIC approach based on PNN offers superior performance and maintains robustness across various modulation schemes. Full article

This research contrasts model-centric (MCAI) and data-centric (DCAI) strategies in artificial intelligence, focusing specifically on optical quality control. It addresses the necessity for a thorough empirical study to evaluate both approaches under identical conditions. By examining casting and leather datasets, the study highlights that the quality and diversity of data play a more vital role in the success of models than merely fine-tuning hyperparameters. While MCAI delivers dependable results with superior datasets, DCAI methods—such as label correction, data augmentation, and generating synthetic data through diffusion models—significantly enhance recognition performance. For the casting dataset, accuracy increased from 83% to 93%, and for the leather dataset, from 53% to 62%. These results indicate that robust AI systems are built on high-quality, balanced data. Full article

Nowadays, information obtained through Global Navigation Satellite Systems (GNSSs) is widely employed in modern geodesy. The Precise Point Positioning (PPP) approach, which leverages signals from multiple GNSS constellations (e.g., GPS, GLONASS, Galileo, and BeiDou), enables high-precision positioning—crucial for seismic monitoring and early tsunami warning systems (EEWs). Recent advances, such as increased satellite availability and additional frequency bands, have significantly improved PPP performance, particularly in terms of positioning accuracy and convergence time. This study focuses on the Rete Integrata Nazionale GNSS (RING) network, managed by the Istituto Nazionale di Geofisica e Vulcanologia (INGV), which comprises dual-frequency GNSS receivers distributed across the Italian peninsula and parts of the Mediterranean Basin. We evaluate the performance of the RING data (GPS and GNSS) acquired in a period of three weeks between 19 January 2024 and 9 February 2024 and analyzed in real time by using different PPP strategies: standard PPP and PPP with Regional Augmentation (PPP-RA). The preliminary results show that the PPP-RA approach enhances positioning accuracy and reduces convergence time, especially when comparing GPS-only datasets with those incorporating full multi-GNSS configurations. For the daily solution, in the optimal setup (i.e., full GNSS with RA), real-time solutions exhibit average accuracies of 2.05, 1.73, and 4.35 cm for the North, East, and vertical components, respectively. Sub-daily accuracies’ analysis, using 300 s sliding windows, showed even better uncertainties, exhibiting median values of 0.41, 0.32, and 0.9 cm for the North, East and vertical components, respectively. Based on the outcomes for network-wide sub-daily accuracies, 84% of the stations demonstrate average errors within 2 cm for North and East components and 3 cm for the vertical one. The analysis on the convergence time after data gaps occurred during the investigation period shows that 87% of the RING stations experienced convergence times lower than five minutes in the GNSS PPP-RA solution. These findings underscore the potential of RT-GNSS RING data for enhancing seismic monitoring and early warning systems, particularly in tectonically active regions. Full article

Vibration-based predictive maintenance is an essential element of reliability engineering for modern automotive powertrains including internal combustion engines, hybrids, and battery-electric platforms. This review synthesizes advances in sensing, signal processing, and artificial intelligence that convert raw vibration into diagnostics and prognostics. It characterizes vibration signatures unique to engines, transmissions, e-axles, and power electronics, emphasizing order analysis, demodulation, and time–frequency methods that extract weak, non-stationary fault content under real driving conditions. It surveys data acquisition, piezoelectric and MEMS accelerometry, edge-resident preprocessing, and fleet telemetry, and details feature engineering pipelines with classical machine learning and deep architectures for fault detection and remaining useful life prediction. In contrast to earlier reviews focused mainly on stationary industrial systems, this review unifies vibration analysis across combustion, hybrid, and electric vehicles and connects physics-based preprocessing to scalable edge and cloud implementations. Case studies show that this integrated perspective enables practical deployment, where physics-guided preprocessing with lightweight models supports robust on-vehicle inference, while cloud-based learning provides cross-fleet generalization and model governance. Open challenges include disentangling overlapping sources in compact e-axles, coping with domain and concept drift from duty cycles, software updates, and aging, addressing data scarcity through augmentation, transfer, and few-shot learning, integrating digital twins and multimodal fusion of vibration, current, thermal, and acoustic data, and deploying scalable cloud and edge AI with transparent governance. By emphasizing inverter-aware analysis, drift management, and benchmark standardization, this review uniquely positions vibration-based predictive maintenance as a foundation for next-generation vehicle reliability. Full article

Background/objectives: High-quality pediatric critical care includes supporting children nearing the end-of-life (EOL) and their families. Cataloging existing interventions for children dying in the neonatal or pediatric intensive care unit (NICU, PICU) establishes critical areas for future research. In this scoping review, we evaluated characteristics of PICU EOL interventions. Methods: A librarian guided a search of OVID Medline, CINAHL, OVID PsycINFO, OVID Embase, Cochrane Central, and Web of Science, plus backwards and forwards reference searching. We included interprofessional interventions, defined as any systematic change (e.g., educational programs, symptom management, electronic medical record, etc.), for children dying from any cause. Studies were independently screened by two reviewers. Data were extracted by one team member and reviewed by a second. We extracted intervention elements, contextual factors, implementation barriers/facilitators, and generated frequencies from qualitative coding. Results: Of 11,643 screened articles, 44 met the inclusion criteria. Most were in neonatal ICUs (n = 28/44, 64%) and general PICUs (n = 10/44, 23%). Most interventions aimed to improve clinician knowledge (25/44, 57%), augment clinical structures and processes (n = 11/44, 25%), or enhance communication (n = 8/44, 18%). Common delivery methods included clinical practice changes (n = 25/44, 57%; e.g., protocols, order sets [n = 12]), and educational sessions (n = 20/44, 45%). Outcomes included clinician knowledge (n = 17/44, 39%), qualitative feedback (n = 18/44, 41%), feasibility/acceptability (n = 12/44, 27%), or treatment utilization (n = 11/44, 25%). Few examined families’ mental health (n = 3, 7%) or bereavement (n = 2, 5%). Few reported implementation facilitators or barriers. Conclusions: Most included studies targeted clinician outcomes through education. Designing, testing, and implementing interventions focused on family outcomes is a critical next step. Full article

Titanium dioxide (TiO₂) is widely used in coatings, plastics, rubber, papermaking, and other industries. The microstructural characteristics of its inorganic shell largely determine the overall performance of the product, significantly affecting optical behavior, dispersibility, weather resistance, and stability. Currently, coating quality evaluation in industry still relies primarily on manual inspection, lacking objective, standardized, and reproducible quantitative methods. This study focuses on lab-prepared core–shell TiO₂ powders comprising a TiO₂ core and a thin inorganic shell enriched in alumina/silica. This study presents Titanium Dioxide U-Net (TD U-Net)—a deep learning approach for transmission electron microscopy (TEM) image segmentation and shell thickness evaluation of core–shell structured TiO₂ particles. TD U-Net employs an encoder–decoder architecture that effectively integrates multi-scale features, addressing challenges such as blurred boundaries and low contrast. We constructed a dataset of 1479 TEM images processed through a six-step workflow: image collection, data cleaning, annotation, mask generation, augmentation, and cropping. Results show that TD U-Net achieves a Dice coefficient of 0.967 for segmentation accuracy and controls shell-thickness measurement error within 5%, significantly outperforming existing image-processing models. An intelligent analysis system developed from this technology has been successfully applied to titanium dioxide product quality assessment, providing an efficient and reliable automated tool for coating-process optimization and quality control. Full article

Segmentation of white blood cells is critical for a wide range of applications. It aims to identify and isolate individual white blood cells from medical images, enabling accurate diagnosis and monitoring of diseases. In the last decade, many researchers have focused on this task using U-Net, one of the most used deep learning architectures. To further enhance segmentation accuracy and robustness, recent advances have explored the combination of U-Net with other techniques, such as attention mechanisms and aggregation techniques. However, a common challenge in white blood cell image segmentation is the similarity between the cells’ cytoplasm and other surrounding blood components, which often leads to inaccurate or incomplete segmentation due to difficulties in distinguishing low-contrast or subtle boundaries, leaving a significant gap for improvement. In this paper, we propose GAAD-U-Net, a novel architecture that integrates attention-augmented convolutions to better capture ambiguous boundaries and complex structures such as overlapping cells and low-contrast regions, followed by a gating mechanism to further suppress irrelevant feature information. These two key components are integrated in the Double U-Net base architecture. Our model achieves state-of-the-art performance on white blood cell benchmark datasets, with a 3.4% Dice score coefficient (DSC) improvement specifically on the SegPC-2021 dataset. The proposed model achieves superior performance as measured by mean the intersection over union (IoU) and DSC, with notably strong segmentation performance even for difficult images. Full article

This paper proposes a hybrid deep learning model for robust and interpretable sentiment classification of Twitter data. The model integrates Bidirectional Encoder Representations from Transformers (BERT)-based contextual embeddings, a Bidirectional Long Short-Term Memory (BiLSTM) network, and a custom attention mechanism to classify tweets into four sentiment categories: Positive, Negative, Neutral, and Irrelevant. Addressing the challenges of noisy and multilingual social media content, the model incorporates a comprehensive preprocessing pipeline and data augmentation strategies including back-translation and synonym replacement. An ablation study demonstrates that combining BERT with BiLSTM improves the model’s sensitivity to sequence dependencies, while the attention mechanism enhances both classification accuracy and interpretability. Empirical results show that the proposed model outperforms BERT-only and BERT+BiLSTM baselines, achieving F1-scores (F1) above 0.94 across all sentiment classes. Attention weight visualizations further reveal the model’s ability to focus on sentiment-bearing tokens, providing transparency in decision-making. The proposed framework is well-suited for deployment in real-time sentiment monitoring systems and offers a scalable solution for multilingual and multi-class sentiment analysis in dynamic social media environments. We also include a focused characterization of the dataset via an Exploratory Data Analysis in the Methods section. Full article

Background: Emotionally salient music may enhance attention-focused rehabilitation, yet concurrent music plus virtual-reality programs in chronic stroke are largely untested. We assessed whether personalized emotional music stimulation (EMS) layered onto a standardized virtual reality rehabilitation system (VRRS) augments cognitive, affective, physiological, and functional outcomes. Methods: In a quasi-randomized outpatient trial, 20 adults ≥ 6 months post-ischemic stroke were allocated by order of recruitment to VRRS alone (control, n = 10) or VRRS+EMS (experimental, n = 10). Both groups performed 45 min of active VRRS cognitive training (3×/week, 8 weeks), while the EMS group received approximately 60 min sessions including setup and feedback phases. Primary outcomes were cognition and global function; secondary outcomes were intrinsic motivation, depression, anxiety, and heart rate. Non-parametric tests with effect sizes and Δ-scores were used. Results: The experimental group improved across all domains: cognition (median +4.5 points), motivation (median +54 points), depression (median −3.5 points), anxiety (median −4.0 points), heart rate (median −6.35 beats per minute), and disability (median one-grade improvement), each with large effects. The control group showed smaller gains in cognition and motivation and a modest heart-rate reduction, without significant changes in mood or disability. At post-treatment, the music group outperformed controls on cognition, motivation, and disability. Change-score analyses favored the music group for every endpoint. Larger heart-rate reductions correlated with greater improvements in depression (ρ = 0.73, p < 0.001) and anxiety (ρ = 0.58, p = 0.007). Conclusions: Adding personalized emotional music to virtual-reality attention training produced coherent, clinically relevant gains in cognition, mood, motivation, autonomic regulation, and independence compared with virtual reality alone. Full article

For clinical decision support to work, medical knowledge needs to be easy to find quickly and accurately. Retrieval-Augmented Generation (RAG) systems use big language models and document retrieval to help with diagnostic reasoning, but they could cause hallucinations and have strict privacy rules in healthcare. We tested twelve different types of RAG, such as dense, sparse, hybrid, graph-based, multimodal, self-reflective, adaptive, and security-focused pipelines, on 250 de-identified patient vignettes. We used Precision@5, Mean Reciprocal Rank, nDCG@10, hallucination rate, and latency to see how well the system worked. The best retrieval accuracy (P@5 ≥ 0.68, nDCG@10 ≥ 0.67) was achieved by a Haystack pipeline (DPR + BM25 + cross-encoder) and hybrid fusion (RRF). Self-reflective RAG, on the other hand, lowered hallucinations to 5.8%. Sparse retrieval gave the fastest response (120 ms), but it was not as accurate. We also suggest a single framework for reducing hallucinations that includes retrieval confidence thresholds, chain-of-thought verification, and outside fact-checking. Our findings emphasize pragmatic protocols for the secure implementation of RAG on premises, incorporating encryption, provenance tagging, and audit trails. Future directions encompass the incorporation of clinician feedback and the expansion of multimodal inputs to genomics and proteomics for precision medicine. Full article

Background/Objectives: Eating disorders (EDs) frequently emerge during critical stages of childhood and adolescence, when identity development and emotional regulation are still maturing. Disturbances in self-concept clarity and identity integration may transform the body into a symbolic battlefield for autonomy, belonging, and self-worth. This review synthesizes developmental, psychosocial, neurocognitive, and therapeutic perspectives on the role of identity disturbance in EDs. Methods: A narrative review was conducted (2010–2025) using combinations of terms related to identity, self-concept clarity, self-discrepancy, objectification, interoception, and eating disorders (anorexia nervosa, bulimia nervosa, and binge-eating disorder). Results: Findings indicate that identity vulnerability (expressed as low self-concept clarity, heightened self-discrepancies, and self-objectification) mediates the association between early adversity, sociocultural pressures, and ED symptoms. Neurocognitive studies reveal altered self-referential processing, default mode network connectivity, and interoceptive signaling. Clinically, comorbid borderline personality features further exacerbate identity disturbance and complicate recovery. Evidence-based treatments such as enhanced cognitive-behavioral therapy (CBT-E) effectively target core maintaining mechanisms, while adjunctive interventions (mentalization-based therapy, schema therapy, narrative approaches, and compassion- or acceptance-based methods) show promise in addressing identity-related processes and improving outcomes. Conclusions: Identity disturbance provides a unifying framework for understanding why ED symptoms become entrenched despite adverse consequences. Integrating identity-focused approaches with nutritional and medical care may enhance recovery and reduce chronicity in youth. Future research should adopt longitudinal and mechanistic designs to clarify pathways linking identity change to clinical improvement and test identity-specific augmentations to standard ED treatments. Full article

The detection of abnormal behavior has consistently garnered significant attention. Conventional methods employ vision-based dual-stream networks or 3D convolutions to represent spatio-temporal information in video sequences to identify normal and pathological behaviors. Nonetheless, these methodologies generally employ datasets balanced across data categories and consist solely of two classifications. In actuality, anomalous behaviors frequently display multi-category characteristics, with each category’s distribution demonstrating a pronounced long-tail phenomenon. This paper presents a video-based technique for detecting multi-category abnormal behavior, termed the Spatio-Temporal Fusion–Temporal Difference Network (STF-TDN). The system first employs a temporal difference network (TDN) model to encapsulate movie temporal dynamics via local and global modeling. To enhance recognition performance, this study develops a feature fusion module—Spatial-Temporal Fusion (STF)—which augments the model’s representational capacity by amalgamating spatial and temporal data. Furthermore, given the long-tailed distribution characteristics of the datasets, this study employs focused loss rather than the conventional cross-entropy loss function to enhance the model’s recognition capability for under-represented categories. We perform comprehensive experiments and ablation studies on two datasets. Precision is 96.3% for the Violence5 dataset and 87.5% for the RWF-2000 dataset. The results of the experiment indicate the enhanced efficacy of the proposed strategy in detecting anomalous behavior. Full article

Background: Focused cardiac ultrasound (FOCUS) can augment the cardiac exam in primary care but is operator-dependent. We evaluated the diagnostic performance of artificial intelligence-assisted FOCUS (AI-FOCUS) performed by family physicians against cardiologist-performed echocardiography. Methods: This research is a prospective cross-sectional study in primary care; family physicians performed conventional FOCUS and AI-FOCUS, with cardiologist-performed echocardiography within 24 h as the reference standard. The primary outcomes were accuracy, sensitivity/specificity, and agreement (κ). Results: AI-FOCUS achieved 94.33% accuracy (95% CI 93.15–95.35), 89.91% sensitivity, and 96.49% specificity, with excellent agreement compared to cardiologists (κ = 0.88). Among the confirmed abnormalities (32.9% of participants), valvular disease was most frequent (42%), followed by reduced LVEF < 50% (28%) and pericardial effusion (12%). In multivariable analysis, AI-assisted LVEF < 50% (OR = 6.05, p < 0.0001) and valvular abnormalities (OR = 4.05, p < 0.0001) were strong predictors of cardiac pathology. Conclusions: AI-FOCUS performed by trained family physicians showed high diagnostic accuracy and excellent agreement with blinded cardiologist-performed echocardiography for detecting LVEF < 50%, screening-level valvular abnormalities, and pericardial effusion, supporting its use for early detection and triage in primary care. Its ease of use and reproducibility suggest value in settings with limited access to cardiology. Full article

Pediatric diffuse midline gliomas in the brainstem (bDMGs) are malignant primary brain neoplasms with poor prognosis. Conventional dogma cites that biopsy procedures have risks of devastating injury to the eloquent brainstem and have no direct benefit to affected patients. In recent years, the use of augmented reality (AR) adjuncts has demonstrated potential in providing excellent intraoperative three-dimensional (3D) visualization of intracranial structures. Put together, we hypothesize that the application of AR will be useful as a training tool for brainstem biopsy procedures. Anatomical models of bDMG tumors are created and uploaded to an AR application. The processed data is transferred into designated AR head-mounted devices. Briefly, individual 3D-rendered bDMG images are overlaid with an age-matched, life-sized child mannequin in prone position. A virtual stereotactic brain biopsy needle is deployed by the user into the lesion. At the end of the exercise, each user evaluates their trajectory of choice to assess its accuracy. Overall, the participants reported that the AR platform was useful in reviewing technical nuances for brainstem biopsy in a safe environment. This focused, proof-of-concept study adds to the growing body of literature that AR platforms demonstrate feasibility for neurosurgeons in the understanding of challenging operative neuroanatomy. Full article

This paper presents a protocol-driven hybrid architecture that integrates Retrieval-Augmented Generation (RAG) with two complementary protocols—A Model Context Protocol (MCP) and an Agent Communication Protocol (ACP)—to deliver adaptive, transparent, and interoperable intelligent tutoring for higher-education STEM courses. MCP stores, fuses, and exposes session-, task- and course-level context (learning goals, prior errors, instructor flags, and policy constraints), while ACP standardizes multipart messaging and orchestration among specialized tutor agents (retrievers, context managers, pedagogical policy agents, execution tools, and generators). A Python prototype indexes curated course materials (two course corpora: a text-focused PDF and a multimodal PDF/transcript corpus) into a vector store and applies MCP-mediated re-ranking (linear fusion of semantic similarity, MCP relevance, instructor tags, and recency) before RAG prompt assembly. In a held-out evaluation (240 annotated QA pairs) and human studies (36 students, 12 instructors), MCP-aware re-ranking improved Recall@1, increased citation fidelity, reduced unsupported numerical claims, and raised human ratings for factuality and pedagogical appropriateness. Case studies demonstrate improved context continuity, scaffolded hinting under instructor policies, and useful multimodal grounding. The paper concludes that the ACP–MCP–RAG combination enables more trustworthy, auditable, and pedagogically aligned tutoring agents and outlines directions for multimodal extensions, learned re-rankers, and large-scale institutional deployment. Full article