Search Results (25)

Electric submersible progressive-cavity pumps (ESPCPs) deliver high lifting efficiency but are prone to failure in the high-temperature, high-pressure, and multiphase down-hole environment, leading to production losses and elevated maintenance costs. To achieve reliable condition recognition under these noisy and highly imbalanced data constraints, we fuse deep residual feature learning, ensemble decision-making, and generative augmentation into a unified diagnosis pipeline. A class-aware TimeGAN first synthesizes realistic minority-fault sequences, enlarging the training pool derived from 360 field records. The augmented data are then fed to a CNN backbone equipped with ResNet blocks, and its deep features are classified by a Random-Forest head (CNN-ResNet-RF). Across five benchmark architectures—including plain CNN, CNN-ResNet, GRU-based, and hybrid baselines—the proposed model attains the highest overall validation accuracy (≈97%) and the best Macro-F1, while the confusion-matrix diagonal confirms marked reductions in the previously dominant misclassification between tubing-leakage and low-parameter states. These results demonstrate that residual encoding, ensemble voting, and realistic data augmentation are complementary in coping with sparse, noisy, and class-imbalanced ESPCP signals. The approach therefore offers a practical and robust solution for the real-time down-hole monitoring and preventive maintenance of ESPCP systems. Full article

The application potential of unmanned aerial vehicles (UAVs) in marine search and rescue is especially of concern for the ongoing advancement of visual recognition technology and image processing technology. Limited computing resources, insufficient pixel representation for small objects in high-altitude images, and challenging visibility conditions hinder UAVs’ target recognition performance in maritime search and rescue operations, highlighting the need for further optimization and enhancement. This study introduces an innovative detection framework, CFSD-UAVNet, designed to boost the accuracy of detecting minor objects within imagery captured from elevated altitudes. To improve the performance of the feature pyramid network (FPN) and path aggregation network (PAN), a newly designed PHead structure was proposed, focusing on better leveraging shallow features. Then, structural pruning was applied to refine the model and enhance its capability in detecting small objects. Moreover, to conserve computational resources, a lightweight CED module was introduced to reduce parameters and conserve the computing resources of the UAV. At the same time, in each detection layer, a lightweight CRE module was integrated, leveraging attention mechanisms and detection heads to enhance precision for small object detection. Finally, to enhance the model’s robustness, WIoUv2 loss function was employed, ensuring a balanced treatment of positive and negative samples. The CFSD-UAVNet model was evaluated on the publicly available SeaDronesSee maritime dataset and compared with other cutting-edge algorithms. The experimental results showed that the CFSD-UAVNet model achieved an mAP@50 of 80.1% with only 1.7 M parameters and a computational cost of 10.2 G, marking a 12.1% improvement over YOLOv8 and a 4.6% increase compared to DETR. The novel CFSD-UAVNet model effectively balances the limitations of scenarios and detection accuracy, demonstrating application potential and value in the field of UAV-assisted maritime search and rescue. Full article

Electrocardiogram (ECG) signals contain complex and diverse features, serving as a crucial basis for arrhythmia diagnosis. The subtle differences in characteristics among various types of arrhythmias, coupled with class imbalance issues in datasets, often hinder existing models from effectively capturing key information within these complex signals, leading to a bias towards normal classes. To address these challenges, this paper proposes a method for arrhythmia classification based on a multi-branch, multi-head attention temporal convolutional network (MB-MHA-TCN). The model integrates three convolutional branch layers with different kernel sizes and dilation rates to capture features across varying temporal scales. A multi-head self-attention mechanism dynamically allocates weights, integrating features and correlations from different branches to enhance the recognition capability for difficult-to-classify samples. Additionally, the temporal convolutional network employs multi-layer dilated convolutions to progressively expand the receptive field for extracting long-term dependencies. To tackle data imbalance, a novel data augmentation strategy is implemented, and focal loss is utilized to increase the weight of minority classes, while Bayesian optimization is employed to fine-tune the model’s hyperparameters. The results from five-fold cross-validation on the MIT-BIH Arrhythmia Database demonstrate that the proposed method achieves an overall accuracy of 98.75%, precision of 96.60%, sensitivity of 97.21%, and F1 score of 96.89% across five categories of ECG signals. Compared to other studies, this method exhibits superior performance in arrhythmia classification, significantly improving the recognition rate of minority classes. Full article

The small area of a printed circuit board (PCB) results in densely distributed defects, leading to a lower detection accuracy, which subsequently impacts the safety and stability of the circuit board. This paper proposes a new YOLO-BFRV network model based on the improved YOLOv8 framework to identify PCB defects more efficiently and accurately. First, a bidirectional feature pyramid network (BIFPN) is introduced to expand the receptive field of each feature level and enrich the semantic information to improve the feature extraction capability. Second, the YOLOv8 backbone network is refined into a lightweight FasterNet network, reducing the computational load while improving the detection accuracy of minor defects. Subsequently, the high-speed re-parameterized detection head (RepHead) reduces inference complexity and boosts the detection speed without compromising accuracy. Finally, the VarifocalLoss is employed to enhance the detection accuracy for densely distributed PCB defects. The experimental results demonstrate that the improved model increases the mAP by 4.12% compared to the benchmark YOLOv8s model, boosts the detection speed by 45.89%, and reduces the GFLOPs by 82.53%, further confirming the superiority of the algorithm presented in this paper. Full article

One of the most basic issues in fluid mechanics is the description of flow in closed flows; more precisely, the calculation of pressure drops and the description of the flow form. Therefore, in this paper, the numerical simulation of the flow through the elbow was presented. This case was used to comprehensively describe the most important phenomena that should be taken into account during closed flows. The elbow was chosen as one of the most frequently used fittings in practice. The simulation was made with ANSYS Fluent, with the use of the turbulent model k-ω, SIMPLE simulation method, and at Reynolds number $Re=500-100,000$. The minor and major pressure loss were presented and discussed in the paper. The minor loss coefficient at the high Reynolds number was equal to around 0.2, which is close to the value of 0.22 used in engineering calculations. The influence of the Reynolds number on the shift of the stream separation point in the elbow was described. The secondary flow in the elbow was observed and the vortex structure was discussed and shown with the use of the Q-criterion (Q iso surface for level 0.005). This analysis allowed us to better visualize and describe the complex flow structure observed in the investigated case. Full article

Ensuring the integrity of aviation plug components is crucial for maintaining the safety and functionality of the aerospace industry. Traditional methods for detecting surface defects often show low detection probabilities, highlighting the need for more advanced automated detection systems. This paper enhances the YOLOv5 model by integrating the Generalized Efficient Layer Aggregation Network (GELAN), which optimizes feature aggregation and boosts model robustness, replacing the conventional Convolutional Block Attention Module (CBAM). The upgraded YOLOv5 architecture, incorporating GELAN, effectively aggregates multi-scale and multi-layer features, thus preserving essential information across the network’s depth. This capability is vital for maintaining high-fidelity feature representations, critical for detecting minute and complex defects. Additionally, the Focal EIOU loss function effectively tackles class imbalance and concentrates the model’s attention on difficult detection areas, thus significantly improving its sensitivity and overall accuracy in identifying defects. Replacing the traditional coupled head with a lightweight decoupled head improves the separation of localization and classification tasks, enhancing both accuracy and convergence speed. The lightweight decoupled head also reduces computational load without compromising detection efficiency. Experimental results demonstrate that the enhanced YOLOv5 architecture significantly improves detection probability, achieving a detection rate of 78.5%. This improvement occurs with only a minor increase in inference time per image, underscoring the efficiency of the proposed model. The optimized YOLOv5 model with GELAN proves highly effective, offering significant benefits for the precision and reliability required in aviation component inspections. Full article

The standard multi-dose nasal spray pump features an integrated actuator and nozzle, which inevitably causes a retraction of the nozzle tip during application. The retraction stroke is around 5.5 mm and drastically reduces the nozzle’s insertion depth, which further affects the initial nasal spray deposition and subsequent translocation, potentially increasing drug wastes and dosimetry variability. To address this issue, we designed a new spray pump that separated the nozzle from the actuator and connected them with a flexible tube, thereby eliminating nozzle retraction during application. The objective of this study is to test the new device’s performance in comparison to the conventional nasal pump in terms of spray generation, plume development, and dosimetry distribution. For both devices, the spray droplet size distribution was measured using a laser diffraction particle analyzer. Plume development was recorded with a high-definition camera. Nasal dosimetry was characterized in two transparent nasal cavity casts (normal and decongested) under two breathing conditions (breath-holding and constant inhalation). The nasal formulation was a 0.25% w/v methyl cellulose aqueous solution with a fluorescent dye. For each test case, the temporospatial spray translocation in the nasal cavity was recorded, and the final delivered doses were quantified in five nasal regions. The results indicate minor differences in droplet size distribution between the two devices. The nasal plume from the new device presents a narrower plume angle. The head orientation, the depth at which the nozzle is inserted into the nostril, and the administration angle play crucial roles in determining the initial deposition of nasal sprays as well as the subsequent translocation of the liquid film/droplets. Quantitative measurements of deposition distributions in the nasal models were augmented with visualization recordings to evaluate the delivery enhancements introduced by the new device. With an extension tube, the modified device produced a lower spray output and delivered lower doses in the front, middle, and back turbinate than the conventional nasal pump. However, sprays from the new device were observed to penetrate deeper into the nasal passages, predominantly through the middle-upper meatus. This resulted in consistently enhanced dosing in the middle-upper turbinate regions while at the cost of higher drug loss to the pharynx. Full article

To solve the problem that the common long-tailed classification method does not use the semantic features of the original label text of the image, and the difference between the classification accuracy of most classes and minority classes are large, the long-tailed image classification method based on enhanced contrast visual language trains the head class and tail class samples separately, uses text image to pre-train the information, and uses the enhanced momentum contrastive loss function and RandAugment enhancement to improve the learning of tail class samples. On the ImageNet-LT long-tailed dataset, the enhanced contrasting visual language-based long-tailed image classification method has improved all class accuracy, tail class accuracy, middle class accuracy, and the F₁ value by 3.4%, 7.6%, 3.5%, and 11.2%, respectively, compared to the BALLAD method. The difference in accuracy between the head class and tail class is reduced by 1.6% compared to the BALLAD method. The results of three comparative experiments indicate that the long-tailed image classification method based on enhanced contrastive visual language has improved the performance of tail classes and reduced the accuracy difference between the majority and minority classes. Full article

Age is variably described as a minor or major risk factor for traumatic intracranial lesions after head injury. However, at present, no specific CT decision rule is available for elderly patients with minor head injury (MHI). The aims of this prospective multicenter cohort study were to assess the performance of existing CT decision rules for elderly MHI patients and to compare the clinical and CT characteristics of elderly patients with the younger MHI population. Thirty-day mortality between two age groups (cutoff ≥ 60 years), along with clinical and CT characteristics, was evaluated with four CT decision rules: the National Institute for Health and Care Excellence (NICE) guideline, the Canadian CT Head Rule (CCHR), the New Orleans Criteria (NOC), and the CT Head Injury Patients (CHIP) rule. Of the 5517 MHI patients included, 2310 were aged ≥ 60 years. Elderly patients experienced loss of consciousness (17% vs. 32%) and posttraumatic amnesia (23% vs. 31%) less often, but intracranial lesions (13% vs. 10%), neurological deterioration (1.8% vs. 0.2%), and 30-day mortality (2.0% vs. 0.1%) were more frequent than in younger patients (all p < 0.001). Elderly patients with age as their only risk factor showed intracranial lesions in 5% (NOC and CHIP) to 8% (CCHR and NICE) of cases. The sensitivity of decision rules in the elderly patients was 60% (CCHR) to 97% (NOC) when age was excluded as a risk factor. Current risk factors considered when evaluating elderly patients show lower sensitivity to identify intracranial abnormalities, despite more frequent intracranial lesions. Until age-specific CT decision rules are developed, it is advisable to scan every elderly patient with an MHI. Full article

Skin exposure to high-dose irradiation, as commonly practiced in radiotherapy, affects the different skin layers, causing dry and wet desquamation, hyperkeratosis fibrosis, hard to heal wounds and alopecia and damaged hair follicles. Fetal tissue mesenchymal stromal cells (f-hPSC) were isolated from excised human fetal placental tissue, based on their direct migration from the tissue samples to the tissue dish. The current study follows earlier reports on for the mitigation of acute radiation syndrome following whole body high-dose exposure with remotely injected f-hPSC. Both the head only and a back skin flap of mice were irradiated with 16 &18 Gy, respectively, by 6MeV clinical linear accelerator electron beam. In both locations, the irradiated skin areas developed early and late radiation induced skin damages, including cutaneous fibrosis, lesions, scaring and severe hair follicle loss and reduced hair pigmentation. Injection of 2 × 10⁶ f-hPSC, 3 and 8 weeks following 16 Gy head irradiation, and 1 and 4 weeks following the 18 Gy back skin only irradiation, resulted in significantly faster healing of radiation induced damages, with reduction of wet desquamation as measured by surface moisture level and minor recovery of the skin viscoelasticity. Detailed histological morphometry showed a clear alleviation of radiation induced hyperkeratosis in f-hPSC treated mice, with significant regain of hair follicles density. Following 16 Gy head irradiation, the hair follicles density in the scalp skin was reduced significantly by almost a half relative to the controls. A nearly full recovery of hair density was found in the f-hPSC treated mice. In the 18 Gy irradiated back skin, the hair follicles density dropped in a late stage by ~70% relative to naïve controls. In irradiated f-hPSC treated mice, it was reduced by only ~30% and was significantly higher than the non-treated group. Our results suggest that local injections of xenogeneic f-hPSC could serve as a simple, safe and highly effective non-autologous pro-regenerative treatment for high-dose radiation induced skin insults. We expect that such treatment could also be applied for other irradiated organs. Full article

Introduction: Only a minority of patients with platinum refractory head and neck squamous cell carcinoma (PR/HNSCC) gain some lasting benefit from immunotherapy. Methods: The combined role of the comprehensive genomic (through the FoundationOne Cdx test) and immune profiles of 10 PR/HNSCC patients treated with the anti-PD-1 nivolumab was evaluated. The immune profiles were studied both at baseline and at the second cycle of immunotherapy, weighing 20 circulating cytokines/chemokines, adhesion molecules, and 14 soluble immune checkpoints dosed through a multiplex assay. A connectivity map was obtained by calculating the Spearman correlation between the expression profiles of circulating molecules. Results: Early progression occurred in five patients, each of them showing TP53 alteration and three of them showing a mutation/loss/amplification of genes involved in the cyclin-dependent kinase pathway. In addition, ERB2 amplification (1 patient), BRCA1 mutation (1 patient), and NOTCH1 genes alteration (3 patients) occurred. Five patients achieved either stable disease or partial response. Four of them carried mutations in PI3K/AKT/PTEN pathways. In the only two patients, with a long response to immunotherapy, the tumor mutational burden (TMB) was high. Moreover, a distinct signature, in terms of network connectivity of the circulating soluble molecules, characterizing responder and non-responder patients, was evidenced. Moreover, a strong negative and statistically significant (p-value ≤ 0.05) correlation with alive status was evidenced for sE-selectin at T1. Conclusions: Our results highlighted the complexity and heterogeneity of HNSCCs, even though it was in a small cohort. Molecular and immune approaches, combined in a single profile, could represent a promising strategy, in the context of precision immunotherapy. Full article

Mild traumatic brain injury (mTBI, or concussion), results from direct and indirect trauma to the head (i.e. a closed injury of transmitted forces), with or without loss of consciousness. The current method of diagnosis is largely based on symptom assessment and clinical history. There is an urgent need to identify an objective biomarker which can not only detect injury, but inform prognosis and recovery. Ocular motor impairment is argued to be ubiquitous across mTBI subtypes and may serve as a valuable clinical biomarker with the recent advent of more affordable and portable eye tracking technology. Many groups have positively correlated the degree of ocular motor impairment to symptom severity with a minority attempting to validate these findings with diffusion tract imaging and functional MRI. However, numerous methodological issues limit the interpretation of results, preventing any singular ocular biomarker from prevailing. This review will comprehensively describe the anatomical susceptibility, clinical measurement, and current eye tracking literature surrounding saccades, smooth pursuit, vestibulo-ocular reflex, vergence, pupillary light reflex, and accommodation in mTBI. Full article

Rusts and powdery mildew are diseases that have a major effect on yield loss in barley. Adult Plant Resistance (APR) is a post-seedling resistance mechanism and its expression is influenced by many factors, including host susceptibility and weather conditions, as well as the timing and severity of disease outbreaks. There are two mechanisms associated with APR: non-hypersensitive and minor gene APR. In this study, 431 European barley accessions were evaluated phenotypically over 2 years (2018–2019) under field conditions, scoring APR to powdery mildew (PM), barley brown rust (BBR), and stem rust (SR), and genotypically using DArTseq. Accessions were grouped into sub-collections by cultivation period (group A—cultivated prior 1985, B—cultivated after 1985, and C—Polish landraces) and by European country of origin or European region. GWAS was conducted for PM, BBR, and SR, and scored at the heading (HA) and milky-waxy (MW) seed stages in 2019 and maximum scores across all replicates were obtained 2018–2019. Disease severity was sufficient to differentiate the collection according to cultivation time and country of origin and to determine SNPs. Overall, the GWAS analysis identified 73 marker–trait associations (MTAs) with these traits. For PM resistance, we identified five MTAs at both the HA stage and when considering the maximal disease score across both growth stages and both years. One marker (3432490-28-T/C) was shared between these two traits; it is located on chromosome 4H. For BBR resistance, six MTAs at HA and one MTA at the MW stage in 2019 and seven MTAs, when considering the maximal disease score across both growth stages and both years, were identified. Of the 48 markers identified as being associated with SR resistance, 12 were on chromosome 7H, 1 was in the telomeric region of the short arm, and 7 were in the telomeric region of the long arm. Rpg1 has previously been mapped to 7HS. The results of this study will be used to create a Polish Gene Bank platform for precise breeding programs. The resistant genotypes and MTA markers will serve as a valuable resource for breeding for PM, BBR, and SR resistance in barley. Full article

Predator impacts on prey diversity are often studied among higher organisms over short periods, but microbial predator-prey systems allow examination of prey-diversity dynamics over evolutionary timescales. We previously showed that Escherichia coli commonly evolved minority mucoid phenotypes in response to predation by the bacterial predator Myxococcus xanthus by one time point of a coevolution experiment now named MyxoEE-6. Here we examine mucoid frequencies across several MyxoEE-6 timepoints to discriminate between the hypotheses that mucoids were increasing to fixation, stabilizing around equilibrium frequencies, or heading to loss toward the end of MyxoEE-6. In four focal coevolved prey populations, mucoids rose rapidly early in the experiment and then fluctuated within detectable minority frequency ranges through the end of MyxoEE-6, generating frequency dynamics suggestive of negative frequency-dependent selection. However, a competition experiment between mucoid and non-mucoid clones found a predation-specific advantage of the mucoid clone that was insensitive to frequency over the examined range, leaving the mechanism that maintains minority mucoidy unresolved. The advantage of mucoidy under predation was found to be associated with reduced population size after growth (productivity) in the absence of predators, suggesting a tradeoff between productivity and resistance to predation that we hypothesize may reverse mucoid vs non-mucoid fitness ranks within each MyxoEE-6 cycle. We also found that mucoidy was associated with diverse colony phenotypes and diverse candidate mutations primarily localized in the exopolysaccharide operon yjbEFGH. Collectively, our results show that selection from predatory bacteria can generate apparently stable sympatric phenotypic polymorphisms within coevolving prey populations and also allopatric diversity across populations by selecting for diverse mutations and colony phenotypes associated with mucoidy. More broadly, our results suggest that myxobacterial predation increases long-term diversity within natural microbial communities. Full article

The segmentation of power lines (PLs) from aerial images is a crucial task for the safe navigation of unmanned aerial vehicles (UAVs) operating at low altitudes. Despite the advances in deep learning-based approaches for PL segmentation, these models are still vulnerable to the class imbalance present in the data. The PLs occupy only a minimal portion (1–5%) of the aerial images as compared to the background region (95–99%). Generally, this class imbalance problem is addressed via the use of PL-specific detectors in conjunction with the popular class balanced cross entropy (BBCE) loss function. However, these PL-specific detectors do not work outside their application areas and a BBCE loss requires hyperparameter tuning for class-wise weights, which is not trivial. Moreover, the BBCE loss results in low dice scores and precision values and thus, fails to achieve an optimal trade-off between dice scores, model accuracy, and precision–recall values. In this work, we propose a generalized focal loss function based on the Matthews correlation coefficient (MCC) or the Phi coefficient to address the class imbalance problem in PL segmentation while utilizing a generic deep segmentation architecture. We evaluate our loss function by improving the vanilla U-Net model with an additional convolutional auxiliary classifier head (ACU-Net) for better learning and faster model convergence. The evaluation of two PL datasets, namely the Mendeley Power Line Dataset and the Power Line Dataset of Urban Scenes (PLDU), where PLs occupy around 1% and 2% of the aerial images area, respectively, reveal that our proposed loss function outperforms the popular BBCE loss by 16% in PL dice scores on both the datasets, 19% in precision and false detection rate (FDR) values for the Mendeley PL dataset and 15% in precision and FDR values for the PLDU with a minor degradation in the accuracy and recall values. Moreover, our proposed ACU-Net outperforms the baseline vanilla U-Net for the characteristic evaluation parameters in the range of 1–10% for both the PL datasets. Thus, our proposed loss function with ACU-Net achieves an optimal trade-off for the characteristic evaluation parameters without any bells and whistles. Our code is available at Github. Full article