Search Results (176)

This work aims to leverage Shapley values to explain the decisions of convolutional neural networks trained to predict glaucoma. Although Shapley values offer a mathematically sound approach rooted in game theory, they require evaluating all possible combinations of features, which can be computationally intensive. To address this challenge, we introduce a novel strategy that discretizes the input by dividing the image into standard regions or sectors of interest, significantly reducing the number of features while maintaining clinical relevance. Moreover, applying Shapley values in a machine learning context necessitates the ability to selectively exclude features to evaluate their combinations. To achieve this, we propose a method involving the occlusion of specific sectors and re-training only the non-convolutional portion of the models. Despite achieving strong predictive performance, our findings reveal limited alignment with medical expectations, particularly the unexpected dominance of the background sector in the model’s decision-making process. This highlights potential concerns regarding the interpretability of convolutional neural network-based glaucoma diagnostics. Full article

Near-field acoustic holography (NAH) provides an effective way to achieve wide-band, high-resolution visualization measurement of the sound insulation performance of building components. However, based on Green’s function, the microphone array’s inherent amplitude and phase mismatch errors will exponentially amplify the sound field inversion process, significantly reducing the measurement accuracy. To systematically evaluate this problem, this study combines numerical simulation with actual measurements in a soundproof room that complies with the ISO 10140 standard, quantitatively analyzes the influence of array system errors on NAH reconstructed sound insulation and acoustic images, and proposes an error correction strategy based on channel transfer function normalization. The research results show that when the array amplitude and phase mismatch mean values are controlled within 5% and 5°, respectively, the deviation of the weighted sound insulation measured by NAH can be controlled within 1 dB, and the error in the key frequency band of building sound insulation (200–1.6k Hz) does not exceed 1.5 dB; when the mismatch mean value increases to 10% and 10°, the deviation of the weighted sound insulation can reach 2 dB, and the error in the high-frequency band (≥1.6k Hz) significantly increases to more than 2.0 dB. The sound image shows noticeable spatial distortion in the frequency band above 250 Hz. After applying the proposed correction method, the NAH measurement results of the domestic microphone array are highly consistent with the weighted sound insulation measured by the standard method, and the measurement difference in the key frequency band is less than 1.0 dB, which significantly improves the reliability and applicability of low-cost equipment in engineering applications. In addition, the study reveals the inherent mechanism of differential amplification of system errors in the propagating wave and evanescent wave channels. It provides quantitative thresholds and operational guidance for instrument selection, array calibration, and error compensation of NAH technology in building sound insulation detection. Full article

(1) Background: Suicide is a worldwide leading cause of death among people with mental disorders. Psychoeducation is an integral component of mental health care that may offer patients valuable tools to understand their conditions, develop coping strategies, and engage more effectively in the treatment process. In the present scoping review, we aimed to summarize the evidence on the implementation of psychoeducational interventions in inpatient settings after suicide attempts. (2) Methods: In August 2024, we searched the Web of Knowledge (all databases), PsycINFO, and CINAHL databases following the PRISMA-ScR guidelines. We included original articles evaluating the effects of psychoeducational interventions for patients hospitalized in psychiatric settings after a suicide attempt. We provided a narrative synthesis of the study characteristics and the main findings of the included studies. (3) Results: We included five papers reporting the results of six studies, of which two were randomized controlled trials. Participants were diagnosed with diverse mental disorders, and interventions were generally short in the hospitalization phase, with follow-ups in the short or long term. Outcomes were focused on suicidal ideation, depressive symptoms, and general functioning, along with feasibility and acceptability of the intervention. Psychoeducational interventions were generally well accepted, but more evidence is needed to determine their efficacy. (4) Conclusions: Psychoeducational intervention in an inpatient psychiatric setting may be important for the prevention of future suicide attempts. Nevertheless, research on the topic is still scarce. Methodologically sound randomized controlled trials evaluating the long-term efficacy of psychoeducational interventions on suicide prevention are needed. Future research should also investigate the utility of psychoeducation in non-psychiatric inpatient settings. Full article

Assessing teacher candidates’ self-efficacy in using reading strategies is essential for understanding their academic development. This study developed and validated the Teacher Candidates’ Self-Efficacy Scale for Informational Reading Strategies (TCSES-IRS) using a mixed-methods sequential exploratory design. Initial qualitative data from interviews with 33 candidates and a literature review guided item generation. Lawshe’s method confirmed content validity. The scale was administered to 1176 teacher candidates. Exploratory (n = 496) and confirmatory factor analyses (n = 388) supported a five-factor structure—cognitive, note-taking, exploration and preparation, physical and process-based, and reflective and analytical strategies—explaining 63.71% of total variance, with acceptable fit indices (χ²/df = 2.64, CFI = 0.912, TLI = 0.900, RMSEA = 0.069). Internal consistency was high (α = 0.899 total; subscales α = 0.708–0.906). An additional sample of 294 participants was used for nomological network validation. Convergent validity was demonstrated by significant item-total correlations and strong factor loadings. Discriminant validity was evidenced by moderate inter-factor correlations. Criterion-related validity was confirmed via significant group differences and meaningful correlations with an external self-efficacy measure. The TCSES-IRS emerges as a psychometrically sound tool for assessing informational reading self-efficacy, supporting research and practice in educational psychology. Full article

The rapid spread of COVID-19 increased the need for speedy diagnostic tools, which led scientists to conduct extensive research on deep learning (DL) applications that use chest imaging, such as chest X-ray (CXR) and computed tomography (CT). This review examines the development and performance of DL architectures, notably convolutional neural networks (CNNs) and emerging vision transformers (ViTs), in identifying COVID-19-related lung abnormalities. Individual ResNet architectures, along with CNN models, demonstrate strong diagnostic performance through the transfer protocol; however, ViTs provide better performance, with improved readability and reduced data requirements. Multimodal diagnostic systems now incorporate alternative methods, in addition to imaging, which use lung ultrasounds, clinical data, and cough sound evaluation. Information fusion techniques, which operate at the data, feature, and decision levels, enhance diagnostic performance. However, progress in COVID-19 detection is hindered by ongoing issues stemming from restricted and non-uniform datasets, as well as domain differences in image standards and complications with both diagnostic overfitting and poor generalization capabilities. Recent developments in COVID-19 diagnosis involve constructing expansive multi-noise information sets while creating clinical process-oriented AI algorithms and implementing distributed learning protocols for securing information security and system stability. While deep learning-based COVID-19 detection systems show strong potential for clinical application, broader validation, regulatory approvals, and continuous adaptation remain essential for their successful deployment and for preparing future pandemic response strategies. Full article

This paper presents a robust hardware system designed for future detection of faults in wind turbines by analyzing audible noise signals. Predictive maintenance strategies have increasingly relied on acoustic monitoring as a non-invasive method for identifying anomalies that may indicate component wear, misalignment, or impending mechanical failures. The proposed device captures and processes sound signals in real-time using strategically positioned microphones, ensuring high-fidelity data acquisition without interfering with turbine operation. Signal processing techniques are applied to extract relevant acoustic features, facilitating future identification of abnormal sound patterns that may indicate mechanical issues. The system’s effectiveness was validated through rigorous field tests, demonstrating its capability to enhance the reliability and efficiency of wind turbine maintenance. Experimental results showed an average transmission latency of 131.8 milliseconds, validating the system’s applicability for near real-time audible noise monitoring in wind turbines operating under limited connectivity conditions. Full article

Growing concern over anthropogenic underwater noise, highlighted by initiatives like the Marine Strategy Framework Directive (MSFD) and its Technical Group on Underwater Noise (TG Noise), emphasizes regions like the Western Black Sea, where increasing activities threaten marine habitats. This region is experiencing rapid growth in maritime traffic and resource exploitation, which is intensifying concerns over the noise impacts on its unique marine habitats. While machine learning offers promising solutions, a research gap persists in comprehensively evaluating diverse ML models within an integrated framework for complex underwater acoustic data, particularly concerning real-world data limitations like class imbalance. This paper addresses this by presenting a multi-faceted framework using passive acoustic monitoring (PAM) data from fixed locations (50–100 m depth). Acoustic data are processed using advanced signal processing (broadband Sound Pressure Level (SPL), Power Spectral Density (PSD)) for feature extraction (Mel-spectrograms for deep learning; PSD statistical moments for classical/unsupervised ML). The framework evaluates Convolutional Neural Networks (CNNs), Random Forest, and Support Vector Machines (SVMs) for noise event classification, alongside Gaussian Mixture Models (GMMs) for anomaly detection. Our results demonstrate that the CNN achieved the highest classification accuracy of 0.9359, significantly outperforming Random Forest (0.8494) and SVM (0.8397) on the test dataset. These findings emphasize the capability of deep learning in automatically extracting discriminative features, highlighting its potential for enhanced automated underwater acoustic monitoring. Full article

Efficient prediction of shallow-water acoustic transmission loss (TL) is crucial for underwater detection, recognition, and communication systems. Traditional physical modeling methods require repeated calculations for each new scenario in practical waveguide environments, leading to low computational efficiency. Deep learning approaches, based on data-driven principles, enable accurate input–output approximation and batch processing of large-scale datasets, significantly reducing computation time and cost. To establish a rapid prediction model mapping sound speed profiles (SSPs) to acoustic TL through controllable generation, this study proposes a hybrid framework that integrates a variational autoencoder (VAE) and a normalizing flow (Flow) through a two-stage training strategy. The VAE network is employed to learn latent representations of TL data on a low-dimensional manifold, while the Flow network is additionally used to establish a bijective mapping between the latent variables and underwater physical parameters, thereby enhancing the controllability of the generation process. Combining the trained normalizing flow with the VAE decoder could establish an end-to-end mapping from SSPs to TL. The results demonstrated that the VAE–Flow network achieved higher computational efficiency, with a computation time of 4 s for generating 1000 acoustic TL samples, versus the over 500 s required by the KRAKEN model, while preserving accuracy, with median structural similarity index measure (SSIM) values over 0.90. Full article

The Flexible Combined Imager (FCI), developed as the next-generation imager for the European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT) Meteosat Third Generation (MTG) satellite series, represents a significant advancement over its predecessor, SEVIRI, on the Meteosat Second Generation (MSG) satellites. FCI offers more spectral bands, higher spatial resolution, and faster imaging capabilities, supporting a wide range of applications in weather forecasting, climate monitoring, and environmental analysis. On 13 January 2024, the FCI onboard MTG-I1 (renamed Meteosat-12 in December 2024) experienced a critical anomaly involving the failure of its onboard Calibration and Obturation Mechanism (COM). As a result, the use of the COM was discontinued to preserve operational safety, leaving the instrument dependent on alternative calibration methods. This loss of onboard calibration presents immediate challenges, particularly for the infrared channels, including image artifacts (e.g., striping), reduced radiometric accuracy, and diminished stability. To address these issues, EUMETSAT implemented an external calibration approach leveraging algorithms from the Global Space-based Inter-Calibration System (GSICS). The inter-calibration algorithm transfers stable and accurate calibration from the Infrared Atmospheric Sounding Interferometer (IASI) hyperspectral instrument aboard Metop-B and Metop-C satellites to FCI’s infrared channels daily, ensuring continued data quality. Comparisons with Cross-track Infrared Sounder (CrIS) data from NOAA-20 and NOAA-21 satellites using a similar algorithm is then used to validate the radiometric performance of the calibration. This confirms that the external calibration method effectively compensates for the absence of onboard blackbody calibration for the infrared channels. For the visible and near-infrared channels, slower degradation rates and pre-anomaly calibration ensure continued accuracy, with vicarious calibration expected to become the primary source. This adaptive calibration strategy introduces a novel paradigm for in-flight calibration of geostationary instruments and offers valuable insights for satellite missions lacking onboard calibration devices. This paper details the COM anomaly, the external calibration process, and the broader implications for future geostationary satellite missions. Full article

Background: Traditionally, auditory rehabilitation in people with hearing loss has sought training in auditory skills to achieve an understanding of sound messages for communication. Assistive or supportive technology is limited to hearing aids that transmit sound through the air or bone to be used by the individual, and only in recent times have technologies for rehabilitation, of high cost and difficult access, begun to be used, employed by audiology professionals. Objective: The objective of this study was to compile the evidence reported in the literature on the use of technology in auditory rehabilitation for the improvement of hearing skills in people with hearing loss, beyond hearing aids and cochlear implants. Method: A systematic review of the literature was conducted between 2018 and 2024 in PubMed, Scopus, and Web of Science databases, using as search terms Technology AND “Auditory Rehabilitation” validated in DeCS and MeSH thesauri; the PICO method was used to propose the research question, and the PRISMA strategy was used for the inclusion or exclusion of the articles to be reviewed. Results: In the first search, 141 documents were obtained. Subsequently, inclusion criteria, such as development with vibrotactile stimulation, Information and Communication Technologies (ICTs), among others, and exclusion criteria, such as those related to cochlear implants and air conduction hearing aids, were applied, and finally, articles related to natural language processing, and other systematic reviews were excluded so that the database was reduced to 14 documents. To this set, due to their relevance, two papers were added, for a total of sixteen analyzed. Conclusions: There are solutions ranging from the use of smartphones for telehealth to solutions with multiple technologies, such as the development of virtual environments with vibrotactile feedback. Hearing-impaired people and even professionals in this area of healthcare have a high level of acceptance of the use of technology in rehabilitation. Finally, this article highlights the crucial role of technology in auditory rehabilitation, with solutions that improve hearing skills and the positive acceptance of these tools by patients and audiology professionals. Full article

The separation of respiratory and cardiac sounds is a significant challenge in biomedical signal processing due to their overlapping frequency and time characteristics. Traditional methods struggle with accurate extraction in noisy or diverse clinical environments. This study explores the application of machine learning, particularly convolutional neural networks (CNNs), to overcome these obstacles. Advanced machine learning models, denoising algorithms, and domain adaptation strategies address challenges such as frequency overlap, external noise, and limited labeled datasets. This study presents a robust methodology for detecting heart and lung diseases from audio signals using advanced preprocessing, feature extraction, and deep learning models. The approach integrates adaptive filtering and bandpass filtering as denoising techniques and variational autoencoders (VAEs) for feature extraction. The extracted features are input into a CNN, which classifies audio signals into different heart and lung conditions. The results highlight the potential of this combined approach for early and accurate disease detection, contributing to the development of reliable diagnostic tools for healthcare. Full article

Traditionally, the microcellular foaming process has aimed to generate uniform cell structures by applying heat uniformly to all surfaces of a polymer. Homogeneous cell distribution is known to enhance the mechanical properties and durability of the final product. However, the ability to engineer a gradient in cell density offers potential advantages for specific functional applications, such as improved sound absorption and thermal insulation. In this study, a controlled thermal gradient was introduced by heating only one side of a fully CO₂-saturated poly(methyl methacrylate) (PMMA) specimen. This approach allowed for the formation of a cell density gradient across the sample thickness. The entire process was conducted using a solid-state batch foaming technique, commonly referred to as the microcellular foaming process. A one-sided heating strategy successfully induced a spatial variation in cell morphology. Furthermore, a coalescence function was developed to account for cell merging behavior, enabling the construction of a predictive model for local cell density. The proposed model accurately captured the evolution of cell density gradients under asymmetric thermal conditions and was validated through experimental observations, demonstrating its potential for precise control over foam structure in saturated PMMA systems. Full article

With the increasing accumulation of plastic residues in agricultural ecosystems, microplastics (MPs) have emerged as a novel and pervasive environmental risk factor threatening sustainable agriculture. Compared to aquatic systems, our understanding of MP dynamics in agricultural soils—particularly their transport mechanisms, bioavailability, plant uptake pathways, and ecological impacts—remains limited. These knowledge gaps impede accurate risk assessment and hinder the development of effective mitigation strategies. This review critically synthesises current knowledge in the study of MPs within soil–plant systems. It examines how MPs influence soil physicochemical properties, plant physiological processes, toxicological responses, and rhizosphere interactions. It further explores the transport dynamics of MPs in soil–plant systems and recent advances in analytical techniques for their detection and quantification. The role of plant functional traits in mediating species-specific responses to MP exposure is also discussed. In addition, the review evaluates the ecological relevance of laboratory-based findings under realistic agricultural conditions, highlighting the methodological limitations imposed by pollution heterogeneity, complex exposure scenarios, and detection technologies. It also examines existing policy responses at both regional and global levels aimed at addressing MP pollution in agriculture. To address these challenges, we propose future research directions that include the integration of multi-method detection protocols, long-term and multi-site field experiments, the development of advanced risk modelling frameworks, and the establishment of threshold values for MP residues in edible crops. Additionally, we highlight the need for future policies to regulate the full life cycle of agricultural plastics, monitor soil MP residues, and integrate MP risks into food safety assessments. This review provides both theoretical insights and practical strategies for understanding and mitigating MP pollution in agroecosystems, supporting the transition toward more sustainable, resilient, and environmentally sound agricultural practices. Full article

The recycling of proton exchange membrane water electrolyzer (PEMWE) raw materials is imperative due to their scarcity, cost, complexity and environmental impact. This is particularly true in the context of expanding electrolyzer manufacturing and reducing production costs. Developing comprehensive recycling strategies requires the creation of a model stack due to the diversity in stack design, structure and materials. The review-derived model presented here provides a sound basis and summarizes the variety of approaches found in the literature and industry. The holistically developed recycling chain, including dismantling, mechanical processing, hydrometallurgical processes and carbon reuse, is characterized by the complete recycling of materials, the reduced application of energy-intensive process steps and the avoidance of environmentally harmful processes. Emphasis is placed on demonstrating the non-destructive disassembly of joined components, the dry mechanical decoating of catalyst-coated membranes, membrane dissolution, the separation of anode and cathode particles and the environmentally friendly hydrometallurgical processing of platinum. Full article

Objective: The development of a normal sleep–wake rhythm in the first weeks of life depends on the physiological sensory needs of the newborn as well as the environment surrounding them. This includes, for example, avoiding pain, exposure to bright light at night and high noise levels. In high-risk newborns, this process can be influenced by immaturity of the central and peripheral nervous systems, therapeutic strategies and the work organization of an intensive care unit. Methods: This study used a narrative review to examine the literature on the interrelationship of sensory modalities on sleep–wake behavior in the context of neonatal intensive care. The current Cochrane reviews on cycled lighting’s effect on premature infants’ circadian rhythm development and noise or sound management in the neonatal intensive care unit, as well as the World Health Organization (WHO) global position paper on kangaroo mother care, were included. Results: An extensive body of literature relates to fetal and neonatal development of the five sensory modalities: touch, taste, smell, hearing and sight. In contrast, there is a lack of evidence regarding the choice of optimal lighting and suitable measures for noise reduction. Since 2023, the WHO has recommended that, from the moment of birth, every “small and sick” newborn should remain in skin-to-skin contact (SSC) with their mother. Developmental support pursues a multimodal approach with the goal of fostering early parent–child bonding, including the child’s needs and environmental conditions. Discussion: The implementation of early SSC and attention to the sleep–wake cycle require systemic changes in both the obstetric and neonatal settings to ensure seamless perinatal management and subsequent neonatal intensive care. Since there is a lack of evidence on the optimal sensory environment, well-designed, well-conducted and fully reported randomized controlled trials are needed that analyze short-term effects and long-term neurodevelopmental outcomes. Full article