Healthcare organizations face a "Triple Threat" of low analytics maturity, high workforce instability, and semantic technical barriers that together produce a crisis of "Institutional Amnesia." Leadership turnover, workforce shortages, and widespread intent to leave among informatics specialists systematically erase the tacit knowledge required to navigate complex clinical data schemas, trapping organizations in a cycle where knowledge loss outpaces knowledge capture. Viewed through Nonaka's SECI (Socialization, Externalization, Combination, Internalization) model of knowledge creation, the root cause is a "Socialization Failure": high turnover fractures the social networks required for mentorship, rendering the traditional apprenticeship model of informatics unsustainable. To address this failure, we employ a Design Science Research (DSR) approach, synthesizing evidence from healthcare informatics, knowledge management, and natural language processing to develop a socio-technical framework: Human-in-the-Loop Knowledge Governance (HITL-KG). HITL-KG is designed to shift the locus of organizational knowledge from volatile human memory to durable semantic artifacts called "Validated Query Triples," each comprising a natural language intent, executable SQL, and rationale metadata. By embedding knowledge capture into the daily query workflow, the framework aims to convert ephemeral analytics into permanent institutional assets. The accompanying Three-Pillar Assessment Rubric enables organizations to identify compounding vulnerabilities across analytics maturity, workforce agility, and technical enablement. The "Validator Paradox" (who validates the AI when experts leave?) is addressed by reframing validation through Lean "Standard Work": each validated query establishes the current known standard rather than eternal truth, functioning as a "knowledge ratchet" that prevents regression. Decoupling analytical capability from individual tenure lets analytics maturity advance even as the workforce evolves. This paper proposes and theoretically motivates the framework; empirical validation is deferred to a companion study.
Public health emergencies place sustained pressure on health systems, particularly where effective coordination across sectors and levels of governance is required. Although Ethiopia has established policy frameworks and coordination mechanisms aligned with global initiatives, evidence on how these arrangements function in practice remains limited. This study assessed multisector governance for public health emergency preparedness and response, focusing on accountability, coordination, institutional roles, and implementation. A sequential mixed-methods study was conducted in Ethiopia between July and September 2025. Quantitative data were collected through a cross-sectional survey of 74 stakeholders from institutions involved in emergency preparedness and response across the health, disaster risk management, agriculture, livestock, academic, and partner sectors. Governance was assessed across seven predefined domains using 5-point Likert-scale measures, with domain scores summarized using descriptive institutional maturity categories. Qualitative data from purposively selected key informants were analyzed thematically to contextualize and explain quantitative findings. The survey achieved a response rate of 92.5%. The overall perceived governance maturity score was 3.31, corresponding to the predefined functional category. Most governance domains were rated as functional, whereas policy framework implementation remained at the emerging level. Qualitative findings indicated that national guidelines and coordination platforms were widely recognized, but their implementation varied across administrative levels. Persistent challenges included role ambiguity, inconsistent coordination and reporting practices, weak feedback mechanisms, and uneven implementation of policy and legal frameworks, particularly at subnational levels, where governance frequently relied on adaptive and informal practices. Overall, Ethiopia has established a functional institutional foundation for multisector emergency governance; however, important policy-to-practice gaps remain. Strengthening operational guidance, coordination, accountability, and implementation across administrative levels will be essential to improve emergency preparedness and response.
Machine learning has become an emerging paradigm for microrobotics, enabling autonomous micro-/nanorobot navigation in complex and highly disturbed environments without requirements of precise models. However, the state-of-the-art learning-based methods adopt "black-box" neural networks (NNs), which raises critical concerns about the trustworthiness of the generated navigation policies, especially for biomedical scenarios. Motivated to address this issue, we propose a physics-constrained learning framework that embeds deterministic physical laws into network architectures to achieve trustworthy microrobot navigation. Instead of learning from scratch, we explicitly encode the monotonic relationships of kinematics and safety rules into NNs. These constraints serve as structural inductive biases, theoretically guaranteeing that navigation policies operate strictly within the trustworthy action domains. Experiments demonstrate that microrobots can autonomously reach random targets without collision with obstacles during long-time and long-distance navigation, proving our framework's reliability owing to its explainable nature. This work can bridge the gap between data-driven performance and rigorous safety requirements, constituting a meaningful step toward trustworthy artificial intelligence-empowered microrobotics.
Understanding and predicting the behavior of liquid matter across length scales-using only the microscopic interactions encoded in the Schrödinger equation-remains a central challenge in the physical sciences. Achieving this goal requires not only an accurate and efficient description of intermolecular forces but also a consistent framework that bridges the micro-, meso-, and macroscales. Here, by combining machine-learned interatomic potentials (MLIPs) with neural classical density functional theory (cDFT), we present such a framework. MLIPs trained on quantum-mechanical energies and forces are used to generate inhomogeneous density profiles, which then serve as the training data for neural cDFT. The resulting ab initio neural cDFT is more computationally efficient than molecular simulations and provides a conceptually transparent route to the thermodynamics of both homogeneous and planar inhomogeneous systems. We demonstrate the approach for both water and carbon dioxide using several exchange-correlation functionals. Beyond accurately reproducing-at the level of the underlying approximate electronic structure-bulk equations of state and liquid-vapor phase diagrams, ab initio neural cDFT predicts, from first principles, how confinement modifies liquid-vapor coexistence in water. It also captures complex behavior in supercritical carbon dioxide such as the Fisher-Widom and Widom lines. While current applications are limited to bulk fluids and planar geometries, this approach establishes a general first-principles route to multiscale modeling of fluids by unifying two independently developed machine-learning paradigms. This work represents an important step toward generalizing cDFT beyond simple empirical potentials to chemically complex systems.
Understanding how links form and predicting future link states is of great importance in social, traffic, and many other complex temporal networks. These temporal networks are typically governed by multiple evolutionary mechanisms. However, existing dynamic graph representation methods, especially dynamic graph neural networks (DGNNs), are constrained by their single-evolutionary-path architecture, where spatial and temporal dynamics are either sequentially stacked or integrated into a monolithic module in a fixed manner. This design not only restricts the exploration of diverse evolutionary paths arising from rich evolutionary mechanisms but also confines spatiotemporal interactions to passive and implicit modeling, resulting in compromised performance and weak interpretability. To address these issues, we propose a novel theoretical model, namely the multiway autoregressive (MARS) model, which characterizes multiple evolutionary paths by capturing dependencies within and across two core factors underlying diverse evolutionary mechanisms. Based on this theoretical foundation, we develop a general DGNN framework, a multiway autoregressive network (MAN), by transforming the network architecture into a 2-D diagram that characterizes evolutionary dependencies in the spatiotemporal domain. Each node encodes an evolving state of the dynamic graph representation, while each edge denotes an evolutionary transition from one state to another. Moreover, three elementary evolutionary operators are incorporated into edges along distinct directions, capturing spatialwise, temporalwise, and cross-spatiotemporal evolutionary dynamics, respectively. This enables researchers to develop a variety of DGNNs by configuring the evolutionary operators in different ways. To validate the effectiveness of this framework, we design a new DGNN, which employs a graph convolutional network (GCN), a gated recurrent unit (GRU), and our proposed time-delayed GCN (TD-GCN) as core components. Promising experimental results demonstrate that the proposed approach achieves state-of-the-art temporal link prediction performance on both synthetic and real-world temporal networks across diverse domains.
Ultrafast ultrasound blood flow imaging is a state-of-the-art technique for depiction of complex blood flow dynamics in vivo through thousands of full-view image data (or, timestamps) acquired per second. Physics-informed Neural Network (PINN) is one of the most preeminent solvers of the Navier-Stokes equations, widely used as the governing equation of blood flow. However, the current approaches which implement time-dependent Navier-Stokes equations within the loss function are impractical for ultrafast ultrasound. We hereby propose an accelerated PINN training framework for solving the Navier-Stokes equations. It involves discretizing the time domain in Navier-Stokes equations and sequentially solving them with test-time adaptation. The novel training framework is coined as SeqPINN. Upon its success, we propose a parallel training scheme for all timestamps based on averaged constant stochastic gradient descent as initialization. Uncertainty estimation through Stochastic Weight Averaging Gaussian is then used as an indicator of generalizability of the initialization. This algorithm, named SP-PINN, further expedites PINN training while achieving comparable accuracy to SeqPINN. The performance of SeqPINN and SP-PINN was evaluated through finite-element simulations and in vitro phantoms of single-branch and trifurcate blood vessels. Results show that both algorithms were manyfold faster than the original design of PINN, while respectively achieving Root Mean Square Errors of 0.63 cm/s and 0.81 cm/s on the straight vessel and 1.07 cm/s and 1.41 cm/s on the trifurcate vessel when recovering blood flow velocities. The successful implementation of SeqPINN and SP-PINN opens the gate for real-time training of PINNs for Navier-Stokes equations and subsequently reliable imaging-based blood flow assessment in clinical practice.
Spirituality has long been recognized as a fundamental dimension of human health and well-being, yet its integration into mental health frameworks remains limited. This essay explores how spirituality is experienced as a coping resource for mental well-being, highlighting mechanisms through which spiritual and religious practices support emotional regulation, resilience, and a sense of purpose. Drawing on theoretical perspectives, empirical research, and lived experiences from diverse cultural contexts including Nepal, it highlights practices such as surrender, mindfulness, detachment, yoga, dietary and hygiene rituals, and selfless service contribute to coping with stress and adversity. The essay also considers the context-dependent and non-linear nature of the spirituality-mental health relationship, potential challenges, including maladaptive interpretations of spiritual beliefs. It underscores the need to balance spirituality with evidence-based mental health care. Recognizing spirituality as a complementary pathway to mental well-being may enhance holistic approaches to mental health and inform interventions that are culturally relevant and personally meaningful.
暂无摘要（点击查看详情）
Distinguishing vitiligo from postinflammatory hypopigmentation (PIH) is clinically challenging because both conditions may present with similar depigmented lesions. Although deep learning has shown strong potential for dermatologic image classification, limited interpretability remains a barrier to clinical adoption. This study aimed to develop an interpretable deep learning framework for accurate differentiation between vitiligo and PIH using a lightweight convolutional neural network and an ensemble of explainability methods. A total of 332 clinical images (176 vitiligo and 156 PIH) were collected from King Abdullah University Hospital and publicly available online sources. Images were preprocessed and evaluated using patient-wise 5-fold cross-validation to eliminate patient-level data leakage. A pretrained MobileNetV2 model was fine-tuned by unfreezing the final 30 layers. To enhance interpretability, gradient-weighted class activation mapping (Grad-CAM), integrated gradients, and smooth gradients (SmoothGrad) were combined into an equal-weight ensemble explanation framework. Performance was assessed using accuracy, precision, recall, F1-score, and the area under the receiver operating characteristic curve (AUC). The proposed model achieved an overall accuracy of 94.88%, macroaveraged precision of 94.88%, recall of 94.84%, F1-score of 94.86%, and an AUC of 0.9885 across the 5 validation folds. The ensemble framework produced clinically meaningful explanations in 98.48% of a representative 66-image validation subset used for interpretability analysis. The proposed framework combines high diagnostic performance with robust interpretability for distinguishing vitiligo from PIH. By integrating multiple complementary explanation methods, the approach enhances clinical transparency and may support dermatologists in the differential diagnosis of pigmentary disorders.
Current point cloud simplification methods for complex ground objects face a persistent challenge: balancing noise robustness and geometric detail preservation. To resolve this, we present a lightweight simplification framework that integrates multi-scale adaptive filtering, entropy-driven spatial partitioning, and an enhanced medial axis transform (MAT). This framework incorporates three targeted technical innovations: (1) An adaptive sliding window polynomial fitting filter with multi-resolution weight adjustment, which achieves coordinated noise suppression and sharp feature preservation; (2) A curvature-weighted enhanced MAT algorithm that reduces skeletal artifacts and topological fractures; (3) An entropy-driven adaptive recursive axis-aligned bounding box (AABB) partitioning strategy, which mitigates the inherent trade-off of conventional uniform partitioning: memory waste in sparse regions and feature loss in dense areas. We validated this framework using self-collected datasets of buildings, vegetation, and roads, and further verified its generalization performance on the public STPLS3D benchmark. Our method achieves an average noise removal rate of 87.76%, representing an average improvement of 11.99% over the baseline method; edge retention is 83.3%, an average improvement of 7.35%; the topological integrity and branch accuracy of the skeleton extraction reached 0.93 and 0.95, respectively, both of which were the best among the tested algorithms; the mean error in normal estimation was as low as 3.34%, and the average point-to-surface distance and fracture rate in 3D reconstruction were the lowest. This method provides a point cloud processing solution that combines accuracy and efficiency for fields such as 3D geographic information modeling and scene reconstruction.
This study examines how personality and emotion are linguistically realized in Tan Twan Eng's The Garden of Evening Mists, drawing on the Big Five Model of Personality (BFM) and Systemic Functional Linguistics (SFL) to construct an integrated theoretical-qualitative framework for the analysis of character language in literary narrative. Emotional polarity is understood as the systematic orientation of affective meaning, positive, neutral, or negative, realized through lexicogrammatical choices at the interpersonal, ideational, and textual levels of language. Through close discourse analysis of selected passages, the study traces how patterns of modality, process type, and cohesion correspond to shifts in emotional polarity across three narrative phases: trauma recollection, apprenticeship in Yugiri, and reflective closure. The findings indicate that emotional polarity in the novel can be read as a form of psychological adaptation rather than static sentiment: negative polarity, realized through obligation modals and syntactic compression, reflects a high-Conscientiousness, high-Neuroticism configuration associated with trauma-related self-regulation; neutral polarity, marked by evaluative balance and descriptive cohesion, signals stabilized Conscientiousness and emerging Openness; positive polarity, expressed through relational processes and permissive modality, corresponds to increased Openness and diminished Neuroticism. The interpersonal dimensions of Agreeableness and Extraversion are realized at lower density but trace parallel arcs across the three phases, with Agreeableness providing the linguistic site at which the novel's central thematic question of inherited hatred is settled. By integrating Big Five personality theory with systemic-functional analysis, the study offers a theoretically grounded and interpretively transparent framework for investigating how emotional language encodes personality transformation in narrative discourse. The findings contribute to ongoing dialogues between personality psychology, functional linguistics, and literary stylistics and illustrate how this framework can be applied to postcolonial narrative fiction.
Driven by the rising prevalence of mental health issues, this study proposes an automated stress monitoring system using wearable ECG signals. Designed for personal medical applications, the framework overcomes individual physiological variability by combining subject-specific normalization with a multi-domain feature extraction strategy encompassing HRV, P-Q-R-S-T morphology, and ECG-derived respiration (EDR). The classification model utilizes a Stacking Ensemble architecture integrating XGBoost, Random Forest, and SVM. Evaluated on the standard WESAD dataset, the proposed framework achieves high accuracy, significantly outperforming baseline single models. These findings demonstrate its robust capability for real-time stress monitoring on personal wearable devices.
Medical information extraction requires automatically identifying disease names and related terms in text. This task, known as named entity recognition (NER), relies on expert-annotated data that are costly to produce and often available only in limited quantities. Data augmentation (DA) aims to expand available training data; however, standard techniques such as synonym replacement and back-translation may introduce inappropriate substitutions or fail to preserve entity-label alignment, which is critical for sequence-labeling tasks. Although large language models can generate fluent text, their outputs may also contain factual inconsistencies or unintended changes if not carefully controlled. This study investigated whether persona-driven, document-level DA using a large language model could improve biomedical disease NER performance by generating diverse rephrasings of medical documents while preserving annotated entities. We designed a DA framework using multiple personas that varied in medical expertise, personality, tone, and narrative style. Using prompting constrained by XML tags, each persona rephrased training documents while aiming to preserve annotated entity spans. We evaluated the framework on 2 biomedical disease NER datasets with complementary roles: RareDis, a low-resource rare disease corpus, and National Center for Biotechnology Information (NCBI) disease, a more general disease benchmark. Semantic fidelity and lexical diversity were measured using BERTScore and Bilingual Evaluation Understudy (BLEU-4), respectively, and personas were grouped into high-, balanced-, and low-fidelity subsets. Biomedical pretrained BioBERT models were fine-tuned and evaluated under multiple settings, including gold-standard (GS) data only, synonym replacement, single-persona augmentation, curated persona subsets, and all-persona augmentation. Performance was assessed using microaveraged entity-level precision, recall, and F1-score, and results were examined at both the overall and individual entity-type levels. Performance values are reported as mean (SD). Persona-driven DA improved NER performance over GS-only training in both datasets, with the strongest gains obtained by combining multiple persona-generated variants with GS data. In RareDis, the best result was achieved by the low-fidelity subset (mean F1-score 73.35, SD 0.19 vs baseline 71.22, SD 0.45), while in NCBI disease, the all-personas setting performed best (mean F1-score 89.32, SD 0.26 vs baseline 87.82, SD 0.18). In low-resource experiments, the all-personas and high-fidelity persona settings in NCBI disease exceeded the performance of the model trained on 100% GS data using only 60% of the training data, whereas gains in RareDis were more modest. Entity-level analysis showed improvements across RareDis categories, particularly for symptom, and confusion analysis indicated reduced symptom-sign confusion under augmentation. Persona-driven DA improved biomedical disease NER by introducing controlled linguistic variation while largely preserving annotated entities. The strongest gains were obtained when multiple persona-generated variants were combined with GS data, although the benefit varied across datasets. These findings suggest that this approach is a promising strategy for low-resource biomedical NER.
To understand families' perspectives on caring for children with chronic conditions in the pre- and trans-pandemic contexts of COVID-19 in light of the Family Management Style Framework. Longitudinal qualitative study conducted with 24 family caregivers of children aged 2 to 4 years. Data were collected through a socioeconomic questionnaire and semi-structured interviews in two stages: pre-pandemic (October 2019 to May 2020) and trans-pandemic (June 2020 to January 2021). Data were subjected to thematic analysis guided by the theoretical framework. The family perspective guided care management over time and was modified by the pandemic. In the pre-pandemic context, it was centered on children's strengths, directing actions toward developmental promotion and health maintenance. In the trans-pandemic context, the perspective focused on children's frailties and vulnerabilities associated with the perceived threat of COVID-19, resulting in intensified surveillance and hygiene care. The family perspective guides care management and is sensitive to contextual changes, providing support for the qualification of professional practice in health crisis contexts.
Long-term garment-type wearable Holter electrocardiographic (ECG) monitoring is frequently affected by noise contamination, which complicates automated atrial fibrillation (AF) detection in real-world recordings. Although deep learning has shown high performance for AF detection, relatively few studies have evaluated explicit strategies for handling noise-included wearable ECG data. An alternative representation using the R-R interval (RRI) time series may reduce the dependence on waveform morphology and provide an alternative pathway for AF screening in noisy recordings. This study aimed to develop and evaluate a 3-class, noise-aware RRI-based AF screening framework that explicitly separated AF, non-AF, and uninterpretable noise windows, and to assess the impact of analysis window length on model performance. Single-lead garment-type wearable Holter ECG data from 117 patients at the University of Osaka Hospital were analyzed after exclusion of patients with documented atrial tachycardia, flutter, or paced rhythm according to the predefined task definition. R-peaks were automatically detected, and the resulting RRI segments were converted into 2D histogram images, with time on the x-axis and RRI-derived heart rate on the y-axis, for 1.5-, 3-, and 6-minute windows. A ResNet-34-based 2D convolutional neural network was trained for 3-class classification. Model performance was evaluated using 5-fold interpatient cross-validation on the institutional dataset and independent external testing on the MIT-BIH (Massachusetts Institute of Technology-Beth Israel Hospital) AF Database (AFDB). In the external validation, atrial flutter-annotated intervals were excluded to match the training task definition. Patient-level AF burden was evaluated by comparing reference AF burden with model-estimated AF burden using Pearson and Spearman correlation coefficients, and linear regression. Of 129 monitored patients between March 1, 2023, and November 20, 2025, 117 were analyzed. In the internal validation, the 3-class model (non-AF, AF, and noise) showed similarly high performance for the 1.5- and 3-minute windows, both with an accuracy of 96.6%. In independent external validation, the 3-minute window showed numerically the highest overall performance (accuracy: 97.3%; AF sensitivity: 96.9%; and AF specificity: 97.7%), although the differences across window lengths were modest. At the patient level, AF burden correlation was high across all window lengths, with Pearson r of 0.995, 0.991, and 0.989 and Spearman ρ of 0.988, 0.982, and 0.979 for the 1.5-, 3-, and 6-minute models, respectively. The RRI-based 2D convolutional neural network achieved high AF classification accuracy and strong patient-level correlation with reference AF burden. Using RRI features and a 3-class framework, which explicitly separated noise from AF and non-AF rhythms, a 3-minute RRI window provided a favorable balance of performance for AF screening in a garment-type Holter ECG.
As tick-borne disease cases continue to increase over time, there is a growing need to understand the ecological and epidemiological factors that contribute to disease risk. North Carolina is currently experiencing a rise in cases, yet tick-borne research in the state remains limited. In this study, we developed a framework for a community science program in collaboration with 22 county public health agencies to recruit participants to submit incidentally encountered ticks. As part of kit submissions, participants also completed a form describing where and when they encountered the tick, as well as the behaviors that led to the encounters. Submitted ticks were tested for several putative bacterial pathogens, including Borrelia burgdorferi, Ehrlichia spp., and Rickettsia amblyommatis. We additionally assessed how advertising methodology and frequency by county health agencies influenced tick kit submissions and evaluated correlations between agency perceptions of project performance and submission rates. Over two years, we received 444 ticks across 319 unique submissions. While species distributions were largely consistent with prior observations, we report the first published instance of Amblyomma americanum within the mountainous region of NC, indicating a potential range expansion into cooler, higher-altitude areas. Relative risk modeling identified recreational activities as consistently associated with higher likelihood of tick kit submissions. Multiple advertising types appeared to influence the number of kits submitted; however, these results should be interpreted cautiously due to small sample sizes. Overall, our study demonstrates a successful collaboration with state and county health agencies to engage the community in tick-borne disease research. Future studies should build on this framework to further optimize participation.
As digital technologies become embedded in everyday life, interpersonal violence increasingly takes place in cyberspace. Existing studies have focused on how states, experts, and the media define what counts as cyber violence and abuse. Less attention has been paid to how ordinary people interpret and contest these definitions and how their interpretations are linked to broader contexts. Drawing on focus group discussions among university students in Beijing, Hong Kong, and Taipei, this paper examines how young people collectively draw boundaries between love, harm, and justice when interpreting cyber dating abuse (CDA). We identify four frames through which participants evaluated behaviours defined by scholars as CDA. We show that framing varies across gender categories, acts and cities, reflecting differences in gendered dating scripts, legal frameworks and sociolegal contexts. Our findings demonstrate that individuals do not simply adopt expert or legal definitions of abuse. They actively negotiate competing understandings by situating acts within particular domains, emphasizing causes or consequences, and invoking different moral principles. Our analysis develops comparative framing from below as a systematic framework connecting individual and interactional framing with sociolegal processes, and for explaining how ordinary people collectively negotiate meanings of contested behaviours in everyday life.
Electrocardiograms (ECGs) are commonly stored in PDF, particularly as vector-based files generated by ECG management systems. Previous studies have demonstrated that ECG signals can be extracted through PDF-to-Scalable Vector Graphics (SVG) conversion, highlighting the potential to reconstruct waveform signals from vector graphics. These reconstructed signals further enable the derivation and prediction of clinically relevant ECG parameters. This study aimed to develop an integrated framework for direct reconstruction of high-fidelity ECG signals from vector-based PDF files and for simultaneous estimation of multiple clinically relevant ECG parameters using deep learning. In this retrospective methodological study, 50,000 twelve-lead ECG PDFs generated by a MUSE system (2015-2024) were analyzed. A direct PDF parsing pipeline was developed to extract vector path objects and reconstruct time-series signals without intermediate format conversion. Reconstruction accuracy was evaluated against the original system-exported signals. Two deep learning models, DualECGFormer and DualResNetECG, were developed to estimate 8 specific ECG parameters. The reference standards for these parameters-including ventricular rate; PR interval; QRS duration; QT interval; corrected QT interval (QTc); and the electrical axes of the P wave, QRS complex, and T wave-were derived from machine-generated values within the MUSE system database. Performance was compared with a rule-based approach (NeuroKit2). The proposed method achieved high reconstruction fidelity, with mean absolute errors (MAEs) below 1×10-3 mV across all leads. Compared with an SVG-based workflow, the direct parsing approach reduced processing time by approximately 4.6-fold. For parameter estimation, deep learning models outperformed the rule-based method for most parameters. DualResNetECG achieved the best overall performance, with MAEs of 1.11 bpm for ventricular rate, 7.27 milliseconds for PR interval, 9.89 milliseconds for QRS duration, 11.87 milliseconds for QT, and 13.71 milliseconds for QTc. For electrical axis estimation, MAEs ranged from 8.0° to 16.31°. The model also demonstrated reliable detection of physiologically undefined parameters (PR interval and P-wave axis), achieving an area under the receiver operating characteristic curve of up to 0.978. This study presents an efficient and scalable framework for direct extraction of ECG signals from MUSE-generated vector-based PDFs and integrated multiparameter estimation using deep learning. The approach achieves high reconstruction accuracy and competitive predictive performance, supporting its potential utility for large-scale retrospective MUSE ECG analysis.
Accurate segmentation of multiple on-orbit space-craft remains difficult in deep-space imagery because the scenes contain large uniform backgrounds, fine structural details, and limited labeled data. To address this problem, we propose SpaceSeg, a segmentation framework that adapts a vision foundation model to the spacecraft domain. The framework introduces a Multi-Scale Hierarchical Attention Refinement Decoder (MSHARD) to improve cross-scale feature decoding, a Spatial Domain Adaptation Transform (SDAT) training strategy to improve robustness to representative space-imaging disturbances, and a task-oriented objective that jointly optimizes segmentation accuracy and IoU-prediction quality. A lightweight connected-component-analysis module is also integrated into the pipeline for instance-aware target organization in multi-spacecraft scenes. We further construct SpaceES, a multi-scale on-orbit multi-spacecraft semantic segmentation dataset covering four space backgrounds and 17 spacecraft types. On SpaceES, SpaceSeg achieves 89.87% mIoU and 99.98% mAcc, setting a new state of the art among all evaluated baselines, surpassing the strongest competing method by 1.38 percentage points in mIoU with 59.6% fewer parameters, and exceeding the vanilla SAM2 baseline by 5.71 percentage points. Hardware-in-the-loop simulation and real satellite-to-satellite imagery experiments further support the practical relevance of the proposed method. Dataset and code are publicly available at https://github.com/Akibaru/SpaceSeg.
Parkinson's disease (PD) can affect vocal production before severe motor symptoms become apparent. This study presents a device-oriented dual-route voice-screening framework that combines interpretable acoustic descriptors with spectrogram-based transfer learning for PD assessment support. Publicly available sustained-vowel /a/ recordings from the Parkinson Speech Dataset with Multiple Types of Sound Recordings were used for algorithmic evaluation. The proposed handheld device, dashboard, and connectivity workflow were treated as a conceptual deployment scenario. The acoustic route generated an integrated acoustic-feature vector and grouped representation-level summaries of selected descriptor families. The spectrogram route generated RGB color-mapped and grayscale time-frequency representations, which were processed using an EfficientNet-B0 transfer-learning convolutional neural network. Among conventional acoustic-feature classifiers, the Ensemble Boosting Classifier achieved the highest AUC point estimate for the integrated acoustic-feature route. Grouped representation-level evidence showed weak-to-moderate discrimination for acoustic descriptor families. RGB and grayscale spectrograms processed through the EfficientNet-B0 route produced reported AUC point estimates of 0.93 and 0.89, respectively, within the available public-dataset evaluation. The proposed framework combines interpretable acoustic descriptors and spectrogram-based transfer learning within a device-oriented screening-support workflow. Integrated acoustic features support interpretable reporting, while spectrogram-based transfer learning provides higher representation-level discrimination. Future validation with physical prototypes, standardized acquisition protocols, and independent clinical cohorts is required before clinical deployment.