Search Results (110,633)

In computer vision, human action recognition is a hot topic, popularized by the development of deep learning. Deep learning models typically accept video input without prior processing and train them to achieve recognition. However, conducting preliminary motion analysis can be beneficial in directing the model training to prioritize the motion of individuals with less priority for the environment in which the action occurs. This paper puts forth a novel methodology for human action recognition based on motion information that employs transfer-learning techniques. The proposed method comprises four stages: (1) human detection and tracking, (2) motion estimation, (3) feature extraction, and (4) action recognition using a two-stream model. In order to develop this work, a customized dataset was utilized, comprising videos of diverse actions (e.g., walking, running, cycling, drinking, and falling) extracted from multiple public sources and websites, including Pexels and MixKit. This realistic and diverse dataset allowed for a comprehensive evaluation of the proposed method, demonstrating its effectiveness in different scenarios and conditions. Furthermore, the performance of seven pre-trained models for feature extraction was evaluated. The models analyzed were Inception-v3, MobileNet-v2, MobileNet-v3-L, VGG-16, VGG-19, Xception, and ConvNeXt-L. The results demonstrated that the ConvNeXt-L model yielded the most optimal outcomes. Furthermore, using pre-trained models for feature extraction facilitated the training process on a personal computer with a single graphics processing unit, achieving an accuracy of 94.9%. The experimental findings and outcomes suggest that integrating motion information enhances action recognition performance. Full article

The dynamic range of an image represents the difference between its darkest and brightest areas, a crucial concept in digital image processing and computer vision. Despite display technology advancements, replicating the broad dynamic range of the human visual system remains challenging, necessitating high dynamic range (HDR) synthesis, combining multiple low dynamic range images captured at contrasting exposure levels to generate a single HDR image that integrates the optimal exposure regions. Recent deep learning advancements have introduced innovative approaches to HDR generation, with the cycle-consistent generative adversarial network (CycleGAN) gaining attention due to its robustness against domain shifts and ability to preserve content style while enhancing image quality. However, traditional CycleGAN methods often rely on unpaired datasets, limiting their capacity for detail preservation. This study proposes an improved model by incorporating a switching map (SMap) as an additional channel in the CycleGAN generator using paired datasets. The SMap focuses on essential regions, guiding weighted learning to minimize the loss of detail during synthesis. Using translated images to estimate the middle exposure integrates these images into HDR synthesis, reducing unnatural transitions and halo artifacts that could occur at boundaries between various exposures. The multilayered application of the retinex algorithm captures exposure variations, achieving natural and detailed tone mapping. The proposed mutual image translation module extends CycleGAN, demonstrating superior performance in multiexposure fusion and image translation, significantly enhancing HDR image quality. The image quality evaluation indices used are CPBDM, JNBM, LPC-SI, S3, JPEG_2000, and SSEQ, and the proposed model exhibits superior performance compared to existing methods, recording average scores of 0.6196, 15.4142, 0.9642, 0.2838, 80.239, and 25.054, respectively. Therefore, based on qualitative and quantitative results, this study demonstrates the superiority of the proposed model. Full article

Background/Objectives: High-grade gliomas are the most common primary malignant brain tumors in adults. These neoplasms remain predominantly incurable due to the genetic diversity within each tumor, leading to varied responses to specific drug therapies. With the advent of new targeted and immune therapies, which have demonstrated promising outcomes in clinical trials, there is a growing need for image-based techniques to enable early prediction of treatment response. This study aimed to evaluate the potential of radiomics and artificial intelligence implementation in predicting progression-free survival (PFS) in patients with highest-grade glioma (CNS WHO 4) undergoing a standard treatment plan. Methods: In this retrospective study, prediction models were developed in a cohort of 51 patients with pathologically confirmed highest-grade glioma (CNS WHO 4) from the authors’ institution and the repository of the Cancer Imaging Archive (TCIA). Only patients with confirmed recurrence after complete tumor resection with adjuvant radiotherapy and chemotherapy with temozolomide were included. For each patient, 109 radiomic features of the tumor were obtained from a preoperative magnetic resonance imaging (MRI) examination. Four clinical features were added manually—sex, weight, age at the time of diagnosis, and the lobe of the brain where the tumor was located. The data label was the time to recurrence, which was determined based on follow-up MRI scans. Artificial intelligence algorithms were built to predict PFS in the training set (n = 75%) and then validate it in the test set (n = 25%). The performance of each model in both the training and test datasets was assessed using mean absolute percentage error (MAPE). Results: In the test set, the random forest model showed the highest predictive performance with 1-MAPE = 92.27% and a C-index of 0.9544. The decision tree, gradient booster, and artificial neural network models showed slightly lower effectiveness with 1-MAPE of 88.31%, 80.21%, and 91.29%, respectively. Conclusions: Four of the six models built gave satisfactory results. These results show that artificial intelligence models combined with radiomic features could be useful for predicting the progression-free survival of high-grade glioma patients. This could be beneficial for risk stratification of patients, enhancing the potential for personalized treatment plans and improving overall survival. Further investigation is necessary with an expanded sample size and external multicenter validation. Full article

Background: Infant growth trajectories reflect current health status and may predict future obesity and metabolic diseases. Human milk is tailored to support optimal infant growth. However, nutrient intake rather than milk composition more accurately predicts growth outcomes. Although the role of protein leverage in infant growth is unclear, protein intake is important for early infancy growth. Materials and methods: This study of exclusively breastfeeding mothers with allergies (n = 161) from the Infant Fish Oil Supplementation Study assessed relationships between intake of human milk macronutrients and infant growth. Human milk fat, protein and lactose concentrations were measured at 3 months postpartum, and infant daily intakes were estimated using an average milk intake of 800 mL/day. Results: Higher human milk protein:energy ratio was associated with higher weight-for-age z-score at 2.5 years compared to 3 months and higher body mass index-for-age z-score change (6 months to 1 year compared to 3–6 months). Maternal atopy and birth season (summer) were negatively associated with human milk lactose concentration. Passive smoke exposure was associated with reduced energy and fat concentrations and increased lactose:energy ratio. Conclusions: Our results indicate that intake of human milk macronutrients may impact early childhood growth. Full article

As the use of electrochemical batteries, especially lithium–ion (Li-Ion) batteries, increases due to emerging applications and expanding markets, the accurate and fast estimation of their state of health (SOH) is becoming increasingly important. The accuracy of the SOH estimation is highly dependent on the correlation strength between the used indicator and SOH and the accuracy of the SOH indicator measurement. This paper presents a new differential indicator which has a strong and consistent correlation with the SOH of Li-Ion batteries, based on a new Electrochemical Impedance Spectrum (EIS) Phase–Magnitude relationship. It is shown in this paper that the EIS Phase–Magnitude relationship exhibits a phase-based differential impedance magnitude SOH indicator between a first-phase peak point and a last-phase valley point. Because of the differential nature of this SOH indicator and because the two impedance values are measured at a phase peak point and a valley phase point regardless of the phase absolute values, the effect of impedance measurement shift/offset (error) on SOH estimation is reduced. This supports the future development of more accurate and faster online and offline SOH estimation algorithms and systems that have a higher immunity to impedance measurement shift/offset (error). Furthermore, in this work, the EIS was measured for a lithium–ion battery that was down to a ~15% SOH, which was not only used to support the conclusions of this paper, but also helped in filling a gap in the literature for EIS data under deep/high degradation levels. Full article

Sepsis is an exaggerated immune response resulting from systemic inflammation, which can damage tissues and organs. Acute kidney injury has been detected in at least one-third of patients with sepsis. Sepsis-associated acute kidney injury increases the risk of a secondary infection. Rapid diagnosis and appropriate initiation of antibiotics can significantly reduce mortality and morbidity. However, microorganisms are known to develop resistance to antibiotics. Estimations indicate that the annual casualties caused by microbial resistance will surpass cancer fatalities by 2050. The prevalence of bacterial infections and their growing antibiotic resistance has brought immediate attention to the search for novel treatments. Plant-derived supplements contain numerous bioactive components with therapeutic potential against a variety of conditions, including infections. Punica granatum peel is rich in phenolic compounds. The purpose of this study was to determine the anti-inflammatory and anti-bacterial properties of P. granatum peel extract (PGPE) on lipopolysaccharide (LPS)-induced acute kidney injury. Experimental groups were Control, LPS (10 mg/kg LPS, intraperitoneally), PGPE100, and PGPE300 (100 and 300 mg/mL PGPE via oral gavage, respectively, for 7 days). According to biochemical results, serum blood urea nitrogen (BUN), creatinine (Cr) and C-reactive protein (CRP), kidney tissue thiobarbituric acid reactive substances (TBARS), and reduced glutathione (GSH) levels significantly decreased in the PGPE groups compared to the LPS group. Histopathological and immunohistochemical findings revealed that toll-like receptor 4 (TLR4) level and nuclear factor kappa B (NF-κB) expression increased in the LPS group compared to the Control group. In addition, the anti-Gram-negative activity showed a dose-dependent effect on Acinetobacter baumannii, Escherichia coli, and Pseudomonas aeruginosa with the agar well diffusion method and the minimal inhibitory concentration (MIC). The MIC value was remarkable, especially on A. baumannii. We conclude that PGPE has the potential to generate desirable anti-bacterial and anti-inflammatory effects on LPS-induced acute kidney injury in rats. Full article

Deep-learning-based (DL-based) seismic inversion has emerged as one of the state-of-the-art research areas in exploration geophysics with the development of artificial intelligence technology. Due to its good portability and high computational efficiency, this method has emerged as a data-driven approach for estimating subsurface properties. However, most of the current DL-based methods rely solely on seismic data, lacking the incorporation of prior information. In addition, these methods are usually performed trace-by-trace, resulting in insufficient horizontal constraints. These limitations make traditional methods less robust, particularly when dealing with high noise levels or limited data. To address these challenges, we propose a multi-input deep learning network for pre-stack inversion, which combines data-driven and model-driven approaches for optimization. The proposed method separately extracts features from the model and data, merging them to improve feature utilization. Moreover, by adopting a 2-D training unit, rather than a trace-by-trace approach, the method improves the horizontal continuity of the results. Tests on synthetic and real seismic data confirmed the robustness and improved stability of the proposed method, even under challenging conditions. This dual-driven approach significantly enhances the reliability of seismic inversion. Full article

Electrochemical DNA sensors for DNA damage detection based on electroactive polymer poly(proflavine) (PPFL) that was synthesized at screen-printed carbon electrodes (SPCEs) from phosphate buffer (PB) and two natural deep eutectic solvents (NADESs) consisting of citric or malonic acids, D-glucose, and a certain amount of water (NADES1 and NADES2) were developed. Poly(proflavine) coatings obtained from the presented media (PPFL_PB, PPFL_NADES1, and PPFL_NADES2) were electrochemically polymerized via the multiple cycling of the potential or potentiostatic accumulation and used for the discrimination of thermal and oxidative DNA damage. The electrochemical characteristics of the poly(proflavine) coatings and their morphology were assessed using cyclic voltammetry (CV), electrochemical impedance spectroscopy (EIS), and scanning electron microscopy (SEM). The working conditions for calf thymus DNA implementation and DNA damage detection were estimated for all types of poly(proflavine) coatings. The voltammetric approach made it possible to distinguish native and chemically oxidized DNA while the impedimetric approach allowed for the successful recognition of native, thermally denatured, and chemically oxidized DNA through changes in the charge transfer resistance. The influence of different concentrations of conventional antioxidants and pharmaceutical preparations on oxidative DNA damage was characterized. Full article

The saturation of natural gas hydrates is a key parameter for characterizing hydrate reservoirs, estimating hydrate reserves, and developing hydrate as an energy resource. Several methods have been proposed to estimate hydrate saturation, although most of these studies rely on logging and seismic data. However, the methods for estimating hydrate saturation from recovered core sediments have not been thoroughly reviewed, which hinders a deeper understanding, proper application, and the use of these experimental data to integrate geophysical and numerical model results with the actual geological conditions. In this paper, the methods widely used for estimating natural gas hydrate saturation from core sediments, including those based on pore water chemistry (Cl⁻ concentration, δD, and δ¹⁸O values), gas volumetric analysis, and temperature anomaly, have been summarized in terms of the principle, estimation strategy, and issues to be considered of each method. The applicability, advantages and disadvantages, and scope of application of each method are also compared and discussed. All methods for estimating gas hydrate saturation have certain limitations. A comprehensive application of results from multiple methods could lead to a better understanding of the amount of gas hydrate in sediments, although the chlorinity of pore water is the most commonly used method of estimation. Full article

Objective: We aimed to describe the clinical characteristics associated with severity in children hospitalized with COVID-19. Method: This was an epidemiological cohort study conducted in two hospitals, one of which was a reference center for the treatment of COVID-19 cases. Data were collected from the reports generated by the hospital epidemiology centers and the medical records of patients aged between 0 and 14 years with a diagnosis of COVID-19, hospitalized between March 2020 and June 2021. To analyze the association between the clinical profile and severity, the cases were classified as severe (severe and critical) and non-severe (asymptomatic, mild, and moderate). Results: Of the 191 children followed up in the cohort, 73.3% developed the severe form. The percentage of children with oxygen saturation below 95% was 46.6%. In the multivariate analysis, a higher risk of severity was estimated among children with uncontrolled asthma (RR = 13.2), who were overweight or obese (RR = 3.21), who had cough symptoms (RR = 2.72), and those aged under one year (RR = 3.23). Conclusions: This result underscores the need to improve healthcare at every level for children and for the management of asthma and nutrition when considering children with this clinical profile who are diagnosed with COVID-19. Full article

INTRODUCTION: This study investigates autism spectrum disorders (ASD) in Greece, focusing on estimating prevalence and identifying regional disparities in children aged 4 to 7 years. MATERIALS AND METHODS: Utilizing a quantitative, descriptive, and exploratory methodology, the research employed a structured questionnaire to gather extensive maternal and child health data. RESULTS: The sample consisted of 517 mothers of children diagnosed with ASD from all over Greece, contributing to a nuanced understanding of ASD predictors. This study aims to elucidate the role of prenatal factors in the likelihood of an ASD diagnosis and their impact on the subsequent functionality of children with ASD. The study identified significant predictors of lower functionality in children with ASD, including higher maternal age, delayed ASD diagnosis, lower family income, and higher birth order. Prenatal health issues, such as vaginal bleeding and infections, also influenced functional outcomes. Notably, a family history of neurological or psychiatric conditions appeared protective. DISCUSSION: The regression model demonstrated robust predictive power, underscoring the complexity of genetic, environmental, and socioeconomic factors in ASD development. The findings advocate for early diagnosis and intervention, systematic screening, and addressing socioeconomic disparities to improve functional outcomes. The results support evidence-based service development and policy adjustments to enhance early identification, intervention, and rehabilitation for children with ASD. CONCLUSIONS: Establishing standardized case-recording procedures and an ASD register at national and regional levels is recommended for systematic monitoring and resource evaluation. Full article

Background/Objectives: In paediatric liver transplantation, donor–recipient compatibility depends on graft size. We explored whether the graft weight can be predicted using the donor’s biometric parameters. Methods: We used seven easily available biometric variables in 142 anonymised paediatric and adult donors, with data collected between 2016 and 2022. The whole or partial liver was transplanted in our hospital from these donors. We identified the variables that had the strongest correlation to our response variable: whole liver graft weight. Results: In child donors, we determined two linear models: using donor weight and height on the one hand and using donor weight and right liver span on the other hand. Both models had a coefficient of determination R² = 0.86 and p-value < 10⁻⁵. We also determined two models in adult donors using donor weight and height (R² = 0.33, p < 10⁻⁴) and donor weight and sternal height (R² = 0.38, p < 10⁻⁴). The models proved valid based on our external dataset of 245 patients from two institutions. Conclusions: In clinical practise, our models could provide rapidly accessible estimates to determine whole graft dimension compatibility in liver transplantation in children and adults. Determining similar models predicting the left lobe and lateral segment weight could prove invaluable in paediatric transplantation. Full article

Modeling agricultural systems, from the point of view of saving and optimizing water, is a challenging task, because it may require multiple soil physical and hydraulic measurements to investigate the entire crop cycle. The Beerkan method was proposed as a quick and easy approach to estimate the saturated soil hydraulic conductivity, K_s. In this study, a new complete three-dimensional model for Beerkan experiments recently proposed was used. It consists of thirteen different calculation approaches that differ in estimating the macroscopic capillary length, initial (θ_i) and saturated (θ_s) soil water contents, use transient or steady-state infiltration data, and different fitting methods to transient data. A steady-state version of the simplified method based on a Beerkan infiltration run (SSBI) was used as the benchmark. Measurements were carried out on five sampling dates during a single growing season (from November to June) in a long-term experiment in which two soil management systems were compared, i.e., minimum tillage (MT) and no tillage (NT). The objectives of this work were (i) to test the proposed new model and calculation approaches under real field conditions, (ii) investigate the impact of MT and NT on soil properties, and (iii) obtain information on the seasonal variability of K_s and other main soil physical properties (θ_i, soil bulk density, ρ_b, and water retention curve) under MT and NT. The results showed that the model always overestimated K_s compared to SSBI. Indeed, the estimated K_s differed by a factor of 11 when the most data demanding (A1) approach was considered by a factor of 4–8, depending on the transient or steady-state phase use, when A3 was considered and by a practically negligible factor of 1.0–1.9 with A4. A relatively higher seasonal variability was detected for θ_i at the MT than NT system. Under both MT and NT, ρ_b did not change between November and April but increased significantly until the end of the season. The selected calculation approaches provided substantially coherent information on K_s seasonal evolution. Regardless of the approach, the results showed a temporal stability of K_s at least from early April to June under NT; conversely, the MT system was, overall, more affected by temporal changes with a relative stability at the beginning and middle of the season. These findings suggest that a common sampling time for determining K_s could be set at early spring. Soil management affected the soil properties, because the NT system was significantly wetter and more compact than MT on four out of five dates. However, only NT showed a significantly increasing correlation between K_s and the modal pore diameter, suggesting the presence of a relatively smaller and better interconnected pore network in the no-tilled soil. This study confirms the need to test infiltration models under real field conditions to evaluate their pros and cons. The Beerkan method was effective for intensive soil sampling and accurate field investigations on the temporal variability of K_s. Full article

Numerous studies have utilized remote sensing techniques to analyze seismic data in active areas. Point density techniques, widely used in remote sensing, examine the spatial distribution of point clouds related to specific variables. Applying these techniques to complex tectonic settings, such as the East Anatolian Fault Zone, helps identify major active fractures using both surface and deep information. This study employed kernel density estimation (KDE) to compare two distinct point-cloud populations from the seismic event along the Türkiye–Syria border on 6 February 2023, providing insights into the main active orientations supporting the Global Tectonics framework. This study considered two populations of seismic foci point clouds containing over 40,000 events, recorded by the Turkish Disaster and Emergency Management Authority (AFAD) and Kandilli Observatory and Earthquake Research Institute (KOERI). These populations were divided into two datasets: crude and relocated-filtered. Kernel density analysis demonstrated that both datasets yielded similar geological interpretations. The high-density cores of both datasets perfectly matched, exhibiting identical structures consistent with geological knowledge. Areas with a minimal concentration of earthquakes at depth were also identified, separating different crustal strength levels. Full article

The range resolution of FMCW radar is critical for accurately distinguishing between multiple objects. Higher range resolution allows for better object separation and more precise location determination. While increasing the bandwidth can improve range resolution, it also raises costs and may be subject to regulatory constraints on the available frequency spectrum. This paper proposes an approach to enhance range resolution by increasing the effective bandwidth through signal processing, specifically by introducing the Estimated Signal Padding (ESP) method. ESP extends the bandwidth by lengthening the signal in the time domain, creating a synthesized signal for each target. The proposed algorithm can improve range resolution by a factor of at least 2.25. Unlike existing methods such as Zero Padding and Neural Network Padding, the proposed algorithm can distinguish targets of varying sizes rather than only targets of the same size. The performance of the algorithm has been validated through MATLAB simulations and FMCW radar experiments. Full article