Search Results (8,364)

Automated surface defect detection in steel manufacturing is pivotal for ensuring product quality, yet it remains an open challenge owing to the extreme heterogeneity of defect morphologies—ranging from hairline cracks and microscopic pores to elongated scratches and shallow dents. Existing approaches, whether classical vision pipelines or recent deep-learning paradigms, struggle to simultaneously satisfy the stringent demands of industrial scenarios: high accuracy on sub-millimeter flaws, insensitivity to texture-rich backgrounds, and real-time throughput on resource-constrained hardware. Although contemporary detectors have narrowed the gap, they still exhibit pronounced sensitivity–robustness trade-offs, particularly in the presence of scale-varying defects and cluttered surfaces. To address these limitations, we introduce MBY (MBDNet-Attention-YOLO), a lightweight yet powerful framework that synergistically couples the MBDNet backbone with the YOLO detection head. Specifically, the backbone embeds three novel components: (1) HGStem, a hierarchical stem block that enriches low-level representations while suppressing redundant activations; (2) Dynamic Align Fusion (DAF), an adaptive cross-scale fusion mechanism that dynamically re-weights feature contributions according to defect saliency; and (3) C2f-DWR, a depth-wise residual variant that progressively expands receptive fields without incurring prohibitive computational costs. Building upon this enriched feature hierarchy, the neck employs our proposed MultiSEAM module—a cascaded squeeze-and-excitation attention mechanism operating at multiple granularities—to harmonize fine-grained and semantic cues, thereby amplifying weak defect signals against complex textures. Finally, we integrate the Inner-SIoU loss, which refines the geometric alignment between predicted and ground-truth boxes by jointly optimizing center distance, aspect ratio consistency, and IoU overlap, leading to faster convergence and tighter localization. Extensive experiments on two publicly available steel-defect benchmarks—NEU-DET and PVEL-AD—demonstrate the superiority of MBY. Without bells and whistles, our model achieves 85.8% mAP@0.5 on NEU-DET and 75.9% mAP@0.5 on PVEL-AD, surpassing the best-reported results by significant margins while maintaining real-time inference on an NVIDIA Jetson Xavier. Ablation studies corroborate the complementary roles of each component, underscoring MBY’s robustness across defect scales and surface conditions. These results suggest that MBY strikes an appealing balance between accuracy, efficiency, and deployability, offering a pragmatic solution for next-generation industrial quality-control systems. Full article

This study investigates the seasonal effectiveness of high-efficiency particulate air (HEPA) purifiers and window-opening behaviors in three London nurseries, using continuous indoor and outdoor PM_2.5 monitoring, window state and air purifier use, and occupant questionnaire data collected from March 2021 to February 2022. Of the approximately 40–50 nurseries contacted, only three agreed to participate. Results show that HEPA purifiers substantially reduced indoor particulate matter (PM_2.5), with the greatest effect observed during the heating season when windows remained closed for longer periods. Seasonal and behavioral analysis indicated more frequent and longer window opening in the non-heating season (windows were open 41.5% of the time on average, compared to 34.2% during the heating season) driven by both ventilation needs and heightened COVID-19 concerns. Predictive modeling identified indoor temperature as the main driver of window opening, while carbon dioxide (CO₂) had a limited effect. In addition, window opening often increased indoor PM_2.5 under prevailing outdoor air quality conditions, with mean concentrations rising from 2.73 µg/m³ (closed) to 3.45 µg/m³ (open), thus reducing the apparent benefit of air purifiers. These findings underscore the complex interplay between mechanical purification and occupant-controlled ventilation, highlighting the need to adapt indoor air quality (IAQ) strategies to both seasonal and behavioral factors in educational settings. Full article

Background/Objectives: Historically, echocardiographic imaging of the right heart has been challenging because its abnormal geometry is not conducive to reproducible anatomical and functional assessment. With the development of advanced echocardiographic techniques, it is now possible to complete an integrated assessment of the right heart that has fewer assumptions, resulting in increased accuracy and precision. Echocardiography continues to be the first-line imaging modality for diagnostic analysis and the management of acute and chronic right heart failure because of its portability, versatility, and affordability compared to cardiac computed tomography, magnetic resonance imaging, nuclear scintigraphy, and positron emission tomography. Virtually all echocardiographic parameters have been well-validated and have demonstrated prognostic significance. The goal of this narrative review of the echocardiographic parameters of the right heart chambers and hemodynamic alterations associated with right ventricular dysfunction is to present information that must be acquired during each examination to deliver a comprehensive assessment of the right heart and to discuss their clinical significance in right heart failure. Methods: Using a literature search in the PubMed database from 1985 to 2025 and the Cochrane database, which included but was not limited to terminology that are descriptive of right heart anatomy and function, disease states involving acute and chronic right heart failure and pulmonary hypertension, and the application of conventional and advanced echocardiographic modalities that strive to elucidate the pathophysiology of right heart failure, we reviewed randomized control trials, observational retrospective and prospective cohort studies, societal guidelines, and systematic review articles. Conclusions: In addition to the conventional 2-dimensional echocardiography and color, spectral, and tissue Doppler measurements, a contemporary echocardiographic assessment of a patient with suspected or proven right heart failure must include 3-dimensional echocardiographic-derived measurements, speckle-tracking echocardiography strain analysis, and hemodynamics parameters to not only characterize the right heart anatomy but to also determine the underlying pathophysiology of right heart failure. Complete and point-of-care echocardiography is available in virtually all clinical settings for routine care, but this imaging tool is particularly indispensable in the emergency department, intensive care units, and operating room, where it can provide an immediate assessment of right ventricular function and associated hemodynamic changes to assist with real-time management decisions. Full article

Recent advancements in generative artificial intelligence have significantly progressed the field of text-to-music generation, enabling users to create music from natural language descriptions. Despite the success of various models, such as MusicLM, MusicGen, and AudioLDM, the current approaches struggle to capture fine-grained genre-specific characteristics, precisely control musical attributes, and handle underrepresented cultural data. This paper introduces a novel, lightweight fine-tuning method for the AudioLDM framework using low-rank adaptation (LoRA). By updating only selected attention and projection layers, the proposed method enables efficient adaptation to musical genres with limited data and computational cost. The proposed method enhances controllability over key musical parameters such as rhythm, emotion, and timbre. At the same time, it maintains the overall quality of music generation. This paper represents the first application of LoRA in AudioLDM, offering a scalable solution for fine-grained, genre-aware music generation and customization. The experimental results demonstrate that the proposed method improves the semantic alignment and statistical similarity compared with the baseline. The contrastive language–audio pretraining score increased by 0.0498, indicating enhanced text-music consistency. The kernel audio distance score decreased by 0.8349, reflecting improved similarity to real music distributions. The mean opinion score ranged from 3.5 to 3.8, confirming the perceptual quality of the generated music. Full article

This paper presents a comparative study of data-driven modeling approaches for vanadium redox flow batteries (VRFBs), utilizing Multiple Linear Regression (MLR) and Random Forest (RF) algorithms. Experimental voltage–capacity datasets from a 1 kW/1 kWh VRFB system were digitized, processed, and used for model training, validation, and testing. The MLR model, built using eight optimized features, achieved a mean error (ME) of 0.0204 V, a residual sum of squares (RSS) of 8.87, and a root mean squared error (RMSE) of 0.1796 V on the test data, demonstrating high predictive performance in stationary operating regions. However, it exhibited limited accuracy during dynamic transitions. Optimized through out-of-bag (OOB) error minimization, the Random Forest model achieved a training RMSE of 0.093 V and a test RMSE of 0.110 V, significantly outperforming MLR in capturing dynamic behavior while maintaining comparable performance in steady-state regions. The accuracy remained high even at lower current densities. Feature importance analysis and partial dependence plots (PDPs) confirmed the dominance of current-related features and SOC dynamics in influencing VRFB terminal voltage. Overall, the Random Forest model offers superior accuracy and robustness, making it highly suitable for real-time VRFB system monitoring, control, and digital twin integration. This study highlights the potential of combining machine learning algorithms with electrochemical domain knowledge to enhance battery system modeling for future energy storage applications. Full article

Cervical cancer represents a significant global health challenge. Photodynamic therapy (PDT) appears to be a promising, minimally invasive alternative to standard treatments. However, the clinical efficacy of PDT is sometimes limited by the low solubility and aggregation of photosensitizers, their non-selective distribution in the body, hypoxia in the tumor microenvironment, and limited light penetration. Recent advances in nanoparticle and nanocomposite platforms have addressed these challenges by integrating multiple functional components into a single delivery system. By encapsulating or conjugating photosensitizers in biodegradable matrices, such as mesoporous silica, organometallic structures and core–shell construct nanocarriers increase stability in water and extend circulation time, enabling both passive and active targeting through ligand decoration. Up-conversion and dual-wavelength responsive cores facilitate deep light conversion in tissues, while simultaneous delivery of hypoxia-modulating agents alleviates oxygen deprivation to sustain reactive oxygen species generation. Controllable “motor-cargo” constructs and surface modifications improve intratumoral diffusion, while aggregation-induced emission dyes and plasmonic elements support real-time imaging and quantitative monitoring of therapeutic response. Together, these multifunctional nanosystems have demonstrated potent cytotoxicity in vitro and significant tumor suppression in vivo in mouse models of cervical cancer. Combining targeted delivery, controlled release, hypoxia mitigation, and image guidance, engineered nanoparticles provide a versatile and powerful platform to overcome the current limitations of PDT and pave the way toward more effective, patient-specific treatments for cervical malignancies. Our review of the literature summarizes studies on nanoparticles and nanocomposites used in PDT monotherapy for cervical cancer, published between 2023 and July 2025. Full article

Background/Objectives: Cognitive impairment is one of the most common and debilitating clinical features of Multiple Sclerosis (MS). Neuropsychological assessment, however, is time-consuming and requires personal resources, so, due to limited resources in daily clinical practice, information on cognitive profiles is often lacking, despite its high prognostic relevance. Time-saving and effective tools are required to bridge this gap. This study evaluates the sensitivity and specificity of a revised version of TRACK-MS (TRACK-MS-R), a recently published screening tool to identify cognitive impairment in MS in a fast and reliable way, offering a balance between efficiency and diagnostic yield for the individual patient. Methods: In this prospective cross-sectional study, 102 MS patients and 94 age-, sex-, and education-matched healthy controls (HC) completed an extensive neuropsychological assessment, including TRACK-MS-R, to test for cognitive processing speed (Symbol Digit Modalities Test, SDMT) and verbal fluency (Regensburger Word Fluency Test, RWT). Sensitivity of TRACK-MS-R was assessed by using the BICAMS-M battery as a reference, and specificity was determined by comparing MS patients to HC. Results: TRACK-MS-R demonstrated high sensitivity (97.44%) when compared to the gold standard as represented by BICAMS-M for early and accurately detecting cognitive impairment in MS patients. Additionally, as a potential cognitive marker, TRACK-MS-R showed a specificity of 82.98% in distinguishing MS patients from healthy controls. Conclusions: TRACK-MS-R proves to be a highly sensitive and time-efficient screening tool for detecting cognitive impairment in patients with MS, while demonstrating good specificity compared to HC. Whereas high sensitivity is a prerequisite for a valid screening tool, its relatively modest specificity compared to BICAMS-M (62.9%) calls for caution in interpreting standalone results but instead indicates more extensive neuropsychological testing. Its briefness and diagnostic accuracy support its implementation in routine clinical practice, particularly in time-constrained settings. Full article

Kinesthetic motor imagery (KMI), the mental rehearsal of a motor task without its actual performance, constitutes one of the most common techniques used for brain–computer interface (BCI) control for movement-related tasks. The effect of neural injury on motor cortical activity during execution and imagery remains under investigation in terms of activations, processing of motor onset, and BCI control. The current work aims to conduct a post hoc investigation of the event-related potential (ERP)-based processing of KMI during BCI control of anthropomorphic robotic arms by spinal cord injury (SCI) patients and healthy control participants in a completed clinical trial. For this purpose, we analyzed 14-channel electroencephalography (EEG) data from 10 patients with cervical SCI and 8 healthy individuals, recorded through Emotiv EPOC BCI, as the participants attempted to move anthropomorphic robotic arms using KMI. EEG data were pre-processed by band-pass filtering (8–30 Hz) and independent component analysis (ICA). ERPs were calculated at the sensor space, and analysis of variance (ANOVA) was used to determine potential differences between groups. Our results showed no statistically significant differences between SCI patients and healthy control groups regarding mean amplitude and latency (p < 0.05) across the recorded channels at various time points during stimulus presentation. Notably, no significant differences were observed in ERP components, except for the P200 component at the T8 channel. These findings suggest that brain circuits associated with motor planning and sensorimotor processes are not disrupted due to anatomical damage following SCI. The temporal dynamics of motor-related areas—particularly in channels like F3, FC5, and F7—indicate that essential motor imagery (MI) circuits remain functional. Limitations include the relatively small sample size that may hamper the generalization of our findings, the sensor-space analysis that restricts anatomical specificity and neurophysiological interpretations, and the use of a low-density EEG headset, lacking coverage over key motor regions. Non-invasive EEG-based BCI systems for motor rehabilitation in SCI patients could effectively leverage intact neural circuits to promote neuroplasticity and facilitate motor recovery. Future work should include validation against larger, longitudinal, high-density, source-space EEG datasets. Full article

Background: Standard room temperature housing (~22 °C) represents a stress for laboratory mice, resulting in an increased metabolic rate, calorie consumption, heart rate, and catecholamine levels compared to thermoneutral conditions (29–32 °C). Using a recently established two-hit model of heart failure with preserved ejection fraction (HFpEF) (Angiotensin II + High-fat diet for 28 days; MHS), we investigated how housing temperature modulates cardiac remodelling and function in male and female C57Bl/6J mice. Methods: Using the MHS mouse model, we investigated cardiac remodelling and function in 8-week-old C57BL/6J mice of both sexes housed at 10 °C, 22 °C, and 30 °C for four weeks. Control mice were analyzed in parallel. Before the MHS, the animals were allowed to acclimate for a week before the MHS started. Results: Mice housed at 10 °C consumed more food and had increased fat mass compared to those at 22 °C or 30 °C. This was accompanied by increased heart weight, stroke volume, heart rate, and cardiac output. Mice housed at 22 °C and 30 °C were similar for these cardiac parameters. Following MHS, mice at 10 °C and 22 °C developed marked cardiac hypertrophy, whereas thermoneutral housing attenuated this response and reduced left atrial enlargement. Cold-exposed females showed more diastolic dysfunction after MHS (increased E’ wave, E/E’, and isovolumetric relaxation time) than those at 22 °C or 30 °C. Ejection fraction and cardiac output declined significantly at 10 °C after MHS but were preserved at 22 °C and 30 °C in females. Conclusions: Cold housing exacerbates cardiac dysfunction in mice subjected to HFpEF-inducing stress, with pronounced effects in females. In contrast, thermoneutrality limits the cardiac hypertrophic response. Full article

Automated quality control plays a critical role in modern industries, particularly in environments that handle large volumes of packaged products requiring fast, accurate, and consistent inspections. This work presents an anomaly detection system for candle jars commonly used in industrial and commercial applications, where obtaining labeled defective samples is challenging. Two anomaly detection strategies are explored: (1) a baseline model using convolutional neural networks (CNNs) as an end-to-end classifier and (2) a hybrid approach where features extracted by CNNs are fed into One-Class classification (OCC) algorithms, including One-Class SVM (OCSVM), One-Class Isolation Forest (OCIF), One-Class Local Outlier Factor (OCLOF), One-Class Elliptic Envelope (OCEE), One-Class Autoencoder (OCAutoencoder), and Support Vector Data Description (SVDD). Both strategies are trained primarily on non-defective samples, with only a limited number of anomalous examples used for evaluation. Experimental results show that both the pure CNN model and the hybrid methods achieve excellent classification performance. The end-to-end CNN reached 100% accuracy, precision, recall, F1-score, and AUC. The best-performing hybrid model CNN-based feature extraction followed by OCIF also achieved 100% across all evaluation metrics, confirming the effectiveness and robustness of the proposed approach. Other OCC algorithms consistently delivered strong results, with all metrics above 95%, indicating solid generalization from predominantly normal data. This approach demonstrates strong potential for quality inspection tasks in scenarios with scarce defective data. Its ability to generalize effectively from mostly normal samples makes it a practical and valuable solution for real-world industrial inspection systems. Future work will focus on optimizing real-time inference and exploring advanced feature extraction techniques to further enhance detection performance. Full article

Target detection is essential in intelligent transportation and autonomous control of unmanned aerial vehicles (UAVs), with single-stage detection algorithms used widely due to their speed. However, these algorithms face limitations in detecting small targets, especially in aerial photography from unmanned aerial vehicles (UAVs), where small targets are often occluded, multi-scale semantic information is easily lost, and there is a trade-off between real-time processing and computational resources. Existing algorithms struggle to effectively extract multi-dimensional features and deep semantic information from images and to balance detection accuracy with model complexity. To address these limitations, we developed FFG-YOLO, a lightweight small-target detection method for UAVs based on YOLOv8. FFG-YOLO incorporates three modules: a feature enhancement block (FEB), a feature concat block (FCB), and a global context awareness block (GCAB). These modules strengthen feature extraction from small targets, resolve semantic bias in multi-scale feature fusion, and help differentiate small targets from complex backgrounds. We also improved the positioning accuracy of small targets using the Wasserstein distance loss function. Experiments showed that FFG-YOLO outperformed other algorithms, including YOLOv8n, in small-target detection due to its lightweight nature, meeting the stringent real-time performance and deployment requirements of UAVs. Full article

This study proposes the SMR Smart Net-Zero City (SSNC) framework—a scalable model for achieving carbon neutrality by integrating Small Modular Reactors (SMRs), renewable energy sources, and sector coupling within a microgrid architecture. As deploying renewables alone would require economically and technically impractical energy storage systems, SMRs provide a reliable and flexible baseload power source. Sector coupling systems—such as hydrogen production and heat generation—enhance grid stability by absorbing surplus energy and supporting the decarbonization of non-electric sectors. The core contribution of this study lies in its real-time data emulation framework, which overcomes a critical limitation in the current energy landscape: the absence of operational data for future technologies such as SMRs and their coupled hydrogen production systems. As these technologies are still in the pre-commercial stage, direct physical integration and validation are not yet feasible. To address this, the researchers leveraged real-time data from an existing commercial microgrid, specifically focusing on the import of grid electricity during energy shortfalls and export during solar surpluses. These patterns were repurposed to simulate the real-time operational behavior of future SMRs (ProxySMR) and sector coupling loads. This physically grounded simulation approach enables high-fidelity approximation of unavailable technologies and introduces a novel methodology to characterize their dynamic response within operational contexts. A key element of the SSNC control logic is a day–night strategy: maximum SMR output and minimal hydrogen production at night, and minimal SMR output with maximum hydrogen production during the day—balancing supply and demand while maintaining high SMR utilization for economic efficiency. The SSNC testbed was validated through a seven-day continuous operation in Busan, demonstrating stable performance and approximately 75% SMR utilization, thereby supporting the feasibility of this proxy-based method. Importantly, to the best of our knowledge, this study represents the first publicly reported attempt to emulate the real-time dynamics of a net-zero city concept based on not-yet-commercial SMRs and sector coupling systems using live operational data. This simulation-based framework offers a forward-looking, data-driven pathway to inform the development and control of next-generation carbon-neutral energy systems. Full article

Quantum finance, as a key application scenario of quantum computing, showcases multiple significant advantages of quantum machine learning over traditional machine learning methods. This paper first aims to overcome the limitations of the hidden quantum Markov model (HQMM) in handling continuous data and proposes an innovative method to convert continuous data into discrete-time sequence data. Second, a hybrid quantum computing model is developed to forecast stock market trends. The model was used to predict 15 stock indices from the Shanghai and Shenzhen Stock Exchanges between June 2018 and June 2021. Experimental results demonstrate that the proposed quantum model outperforms classical algorithmic models in handling higher complexity, achieving improved efficiency, reduced computation time, and superior predictive performance. This validation of quantum advantage in financial forecasting enables the practical deployment of quantum-inspired prediction models by investors and institutions in trading environments. This quantum-enhanced model empowers investors to predict market regimes (bullish/bearish/range-bound) using real-time data, enabling dynamic portfolio adjustments, optimized risk controls, and data-driven allocation shifts. Full article

The extensive use of conventional pesticides has been a fundamental strategy in modern agriculture for controlling pests and increasing crop productivity; however, their improper application poses significant risks to human health and environmental sustainability. This review compiles scientific evidence linking pesticide exposure to oxidative stress and genotoxic damage, particularly affecting rural populations and commonly consumed foods, even at levels exceeding the maximum permissible limits in fruits, vegetables, and animal products. Additionally, excessive pesticide use has been shown to alter soil microbiota, negatively compromising long-term agricultural fertility. In response to these challenges, recent advances in nanotechnology offer promising alternatives. This review highlights the development of nanopesticides designed for controlled release, improved stability, and targeted delivery of active ingredients, thereby reducing environmental contamination and increasing efficacy. Moreover, emerging nanobiosensor technologies, such as e-nose and e-tongue systems, have shown potential for real-time monitoring of pesticide residues and soil health. Although pesticides are still necessary, it is crucial to implement stricter laws and promote sustainable solutions that ensure safe and responsible agricultural practices. The need for evidence-based public policy is emphasized to regulate pesticide use and protect both human health and agricultural resources. Full article

Mushrooms are eukaryotic organisms with absorptive heterotrophic nutrition, capable of feeding on organic matter rich in cellulose and lignocellulose. Since ancient times, they have been considered allies and, in certain cultures, they were seen as magical beings or food of the gods. Of the great variety of edible mushrooms identified worldwide, less than 2% are traded on the market. Although mushrooms have been valued for their multiple nutritional and healing benefits, some cultures perceive them as toxic and do not accept them in their culinary practices. Despite the existing skepticism, several researchers are promoting the potential of edible mushrooms. There are two main methods of mushroom cultivation: solid-state fermentation and submerged fermentation. The former is the most widely used and simplest, since the fungus grows in its natural environment; in the latter, the fungus grows suspended without developing a fruiting body. In addition, submerged fermentation is easily monitored and scalable. Both systems are important and have their limitations. This article discusses the main methods used to increase the performance of submerged fermentation with emphasis on the modes of operation used, types of bioreactors and application of morphological bioengineering of filamentous fungi, and especially the use of intelligent automatic control technologies and the use of non-invasive monitoring in fermentation systems thanks to the development of machine learning (ML), neural networks, and the use of big data, which will allow more accurate decisions to be made in the fermentation of filamentous fungi in submerged environments with improvements in production yields. Full article