Search Results (4,290)

Electroencephalogram (EEG) analysis is essential for diagnosing neurological disorders but typically requires expert interpretation and significant time. Purpose: This study aims to automate the classification of normal and abnormal EEG recordings to support clinical diagnosis and reduce manual workload. Automating the initial screening of EEGs can help clinicians quickly identify potential neurological abnormalities, enabling timely intervention and guiding further diagnostic and treatment strategies. Methodology: We utilized the Temple University Hospital EEG dataset to develop a DenseNet-based deep learning model. To enable a fair comparison of different EEG representations, we used three input types: signal images, spectrograms, and scalograms. To reduce dimensionality and simplify computation, we focused on two channels: T5 and O1. For interpretability, we applied Local Interpretable Model-agnostic Explanations (LIME) and Gradient-weighted Class Activation Mapping (Grad-CAM) to visualize the EEG regions influencing the model’s predictions. Key Findings: Among the input types, spectrogram-based representations achieved the highest classification accuracy, indicating that time-frequency features are especially effective for this task. The model demonstrated strong performance overall, and the integration of LIME and Grad-CAM provided transparent explanations of its decisions, enhancing interpretability. This approach offers a practical and interpretable solution for automated EEG screening, contributing to more efficient clinical workflows and better understanding of complex neurological conditions. Full article

Background/Objectives: Obesity is a major cause of repolarisation defects of the heart. The frontal QRS-T angle is a new parameter used for cardiac evaluation. This study aimed to evaluate the effects of a laparoscopic cholecystectomy and anaesthetic agents on the frontal QRS-T angle in individuals with obesity. Methods: A total of 91 patients who underwent a laparoscopic cholecystectomy surgery were included in this study. The patients were divided into two groups according to body mass index (BMI) < 30 (n = 68) and ≥30 (n = 23). The frontal QRS-T angle (FQRST), QT interval (QT), corrected QT, and other electrocardiography (ECG) findings were recorded at different time points. Results: In the BMI ≥ 30 group, the frontal QRS-T angle and QT interval measured during the intraoperative period were statistically higher than those of the BMI < 30 group (p < 0.001, p < 0.001). Additionally, the frontal QRS-T angle value was statistically higher in all patients postoperatively compared with the preoperative and intraoperative periods (p < 0.001). Furthermore, there was a positive correlation between the BMI and the frontal QRS-T angle. Our study found that the QRS-T angle and the QT interval duration measured during surgery in the BMI ≥ 30 group who underwent a laparoscopic cholecystectomy were significantly higher than in the BMI < 30 group. Conclusions: We recommend close haemodynamic monitoring during and after surgery for patients with obesity undergoing a laparoscopic cholecystectomy. Full article

Gear wear degrades transmission performance, necessitating highly reliable fault diagnosis methods. To address the limitations of existing approaches—where dynamic models rely heavily on prior knowledge, while data-driven methods lack interpretability—this study proposes an integrated bidirectional verification framework combining dynamic modeling and deep learning for interpretable gear wear diagnosis. First, a dynamic gear wear model is established to quantitatively reveal wear-induced modulation effects on meshing stiffness and vibration responses. Then, a deep network incorporating Gradient-weighted Class Activation Mapping (Grad-CAM) enables visualized extraction of frequency-domain sensitive features. Bidirectional verification between the dynamic model and deep learning demonstrates enhanced meshing harmonics in wear faults, leading to a quantitative diagnostic index that achieves 0.9560 recognition accuracy for gear wear across four speed conditions, significantly outperforming comparative indicators. This research provides a novel approach for gear wear diagnosis that ensures both high accuracy and interpretability. Full article

Tumor microenvironment (TME) analysis plays an extremely important role in computational pathology. Deep learning shows tremendous potential for tumor tissue segmentation on pathological images, which is an essential part of TME analysis. However, fully supervised segmentation methods based on deep learning usually require a large number of manual annotations, which is time-consuming and labor-intensive. Recently, weakly supervised semantic segmentation (WSSS) works based on the Class Activation Map (CAM) have shown promising results to learn the concept of segmentation from image-level class labels but usually have imprecise boundaries due to the lack of pixel-wise supervision. On the other hand, the Segment Anything Model (SAM), a foundation model for segmentation, has shown an impressive ability for general semantic segmentation on natural images, while it suffers from the noise caused by the initial prompts. To address these problems, we propose a simple but effective weakly supervised framework, termed as 2AM, combining CAM and SAM for tumor tissue segmentation on pathological images. Our 2AM model is composed of three modules: (1) a CAM module for generating salient regions for tumor tissues on pathological images; (2) an adaptive point selection (APS) module for providing more reliable initial prompts for the subsequent SAM by designing three priors of basic appearance, space distribution, and feature difference; and (3) a SAM module for predicting the final segmentation. Experimental results on two independent datasets show that our proposed method boosts tumor segmentation accuracy by nearly 25% compared with the baseline method, and achieves more than 15% improvement compared with previous state-of-the-art segmentation methods with WSSS settings. Full article

In this study, particles of ground grape seeds were utilized to adsorb phosphate ions from a prepared solution, aiming to reduce phosphate concentration. Through a series of adsorption experiments, the effects of the adsorbent concentration, initial phosphate ion concentration, temperature, and pH on the phosphate ion uptake were studied. The removal efficiency of the phosphate ion decreased from 77 to 61% as a 25 to 45 °C increment in temperature was observed, which indicated the exothermicity in the adsorption process. The phosphate ion movement onto the adsorbent surface that exhibited the highest uptake value favored a neutral reaction environment with a pH value of seven. The experimental results, when compared using different adsorption isotherms, showed that the best fit was exhibited by the Jovanovic isotherm, which was further confirmed owing to its high 0.974 R² value. Intraparticle diffusion and pseudo second order models describe the kinetics of phosphate adsorption onto grape seeds, with reaction constants of 8.8 × 10⁻³ (mg/g min) and 0.412 (mg/g·min^0.5), respectively. The adsorption was physiosorptive, spontaneous, exothermic, and favorable. Furthermore, the negative entropy with a value of −0.0887 kJ/mol·K revealed reduced randomness in the adsorption process system. Full article

Using the fly eye as a model system, we previously demonstrated that upregulation of the fly gene mask protects against FUS- and Tau-induced photoreceptor degeneration. Building upon this finding, we investigated whether the protective role of mask is conserved in mammals. To this end, we generated a transgenic mouse line carrying Cre-inducible ANKHD1, the human homolog of mask. Utilizing the TauP301S-PS19 mouse model for Tau-related dementia, we found that expressing ANKHD1 driven by CamK2a-Cre reduced hyperphosphorylated human Tau in 6-month-old mice. Additionally, ANKHD1 expression was associated with a trend toward reduced gliosis and preservation of the presynaptic marker Synaptophysin, suggesting a protective role of ANKHD1 against TauP301S-linked neuropathology. At 9 months of age, novel object recognition (NOR) testing revealed cognitive impairment in female, but not male, PS19 mice. Notably, co-expression of ANKHD1 restored cognitive performance in the affected female mice. Together, this study highlights the novel effect of ANKHD1 in counteracting the adverse effects induced by the mutant human Tau protein. This finding underscores ANKHD1’s potential as a unique therapeutic target for tauopathies. Full article

Background: This study investigated a new multi-acid-etching formulation for zirconia ceramics, containing hydrochloric, hydrofluoric, nitric, orthophosphoric, and sulfuric acids. The solution was tested on polycrystalline (5Y-TZP zirconia), lithium disilicate, hybrid ceramic, and feldspathic porcelain to assess compatibility, etching selectivity, and surface conditioning. Methods: Two-hundred-and-forty CAD/CAM specimens were etched for 20 s, 60 s, 30 min, or 1 h, and their surface roughness and etching patterns ware evaluated using 3D optical profilometry and scanning electron microscopy (SEM). Results: A positive correlation was observed between etching time and surface roughness (Ra values). The most pronounced changes were observed in lithium disilicate and feldspathic porcelain, with Ra values increasing from 0.733 ± 0.082 µm (Group 5) to 1.295 ± 0.123 µm (Group 8), and from 0.902 ± 0.102 µm (Group 13) to 1.480 ± 0.096 µm (Group 16), respectively. Zirconia increased from 0.181 ± 0.043 µm (Group 1) to 0.371 ± 0.074 µm (Group 4), and the hybrid ceramic from 0.053 ± 0.008 µm (Group 9) to 0.099 ± 0.016 µm (Group 12). Two-way ANOVA revealed significant effects of material and etching time, as well as a significant interaction between the two factors (p < 0.001). SEM observation revealed non-selective etching pattern for the lithium disilicate groups, indicating a risk of over-etching. Conclusions: The tested etching solution increased surface roughness, especially for the lithium disilicate and feldspathic porcelain specimens. In zirconia, one-hour etching improved surface characteristics with minimal observable damage. However, additional studies are necessary to validate the mechanical stability and bond effectives of this approach. Full article

Background/Objectives: This study aims to design and evaluate transfer learning strategies that fine-tune multiple pre-trained convolutional neural network architectures based on their characteristics to improve the accuracy and generalizability of colorectal cancer histopathological image classification. Methods: The application of transfer learning with pre-trained models on specialized and multiple datasets is proposed, where the proposed models, CRCHistoDense, CRCHistoIncep, and CRCHistoXcep, are algorithmically fine-tuned at varying depths to improve the performance of colorectal cancer classification. These models were applied to datasets of 10,613 images from public and private repositories, external sources, and unseen data. To validate the models’ decision-making and improve transparency, we integrated Grad-CAM to provide visual explanations that influence classification decisions. Results and Conclusions: On average across all datasets, CRCHistoDense, CRCHistoIncep, and CRCHistoXcep achieved test accuracies of 99.34%, 99.48%, and 99.45%, respectively, highlighting the effectiveness of fine-tuning in improving classification performance and generalization. Statistical methods, including paired t-tests, ANOVA, and the Kruskal–Wallis test, confirmed significant improvements in the proposed methods’ performance, with p-values below 0.05. These findings demonstrate that fine-tuning based on the characteristics of CNN’s architecture enhances colorectal cancer classification in histopathology, thereby improving the diagnostic potential of deep learning models. Full article

In intelligent vehicular networks, the accuracy of semantic segmentation in road scenes is crucial for vehicle-mounted artificial intelligence to achieve environmental perception, decision support, and safety control. Although deep learning methods have made significant progress, two main challenges remain: first, the difficulty in balancing global and local features leads to blurred object boundaries and misclassification; second, conventional convolutions have limited ability to perceive irregular objects, causing information loss and affecting segmentation accuracy. To address these issues, this paper proposes a global–local collaborative attention module and a spider web convolution module. The former enhances feature representation through bidirectional feature interaction and dynamic weight allocation, reducing false positives and missed detections. The latter introduces an asymmetric sampling topology and six-directional receptive field paths to effectively improve the recognition of irregular objects. Experiments on the Cityscapes, CamVid, and BDD100K datasets, collected using vehicle-mounted cameras, demonstrate that the proposed method performs excellently across multiple evaluation metrics, including mIoU, mRecall, mPrecision, and mAccuracy. Comparative experiments with classical segmentation networks, attention mechanisms, and convolution modules validate the effectiveness of the proposed approach. The proposed method demonstrates outstanding performance in sensor-based semantic segmentation tasks and is well-suited for environmental perception systems in autonomous driving. Full article

To enable the practical application of passive back-support exoskeletons employing pneumatic artificial muscles (PAMs) in tasks such as agricultural work, we evaluated their assistive effectiveness in a half-squatting posture with a staggered stance. In this context, assistive force profiles were adjusted according to body posture to achieve more effective support. The targeted assistive force profile was designed to be continuously active from the standing to the half-squatting position, with minimal variation across this range. The assistive force profile was developed based on a PAM contractile force model and implemented using a cam mechanism. The effectiveness of assistance was assessed by measuring body flexion angles and erector spinae muscle activity during lifting and carrying tasks. The results showed that the assistive effect was greater on the side with the forward leg. Compared to the condition without exoskeleton assistance, the conventional pulley-based system reduced muscle activity by approximately 20% whereas the cam-based system achieved a reduction of approximately 30%. Full article

Tea cultivation faces significant threats from red spider mite (RSM: Oligonychus coffeae) infestations, which reduce yields and economic viability in major tea-producing regions. Current automated detection methods rely on supervised deep learning models requiring extensive labeled data, limiting scalability for smallholder farmers. This article proposes AIM-Net (AI-based Infestation Mapping Network) by evaluating SwAV (Swapping Assignments between Views), a self-supervised learning framework, for classifying RSM infestation severity (Mild, Moderate, Severe) using a geo-referenced, field-acquired dataset of RSM infested tea-leaves, Cam-RSM. The methodology combines SwAV pre-training on unlabeled data with fine-tuning on labeled subsets, employing multi-crop augmentation and online clustering to learn discriminative features without full supervision. Comparative analysis against a fully supervised ResNet-50 baseline utilized 5-fold cross-validation, assessing accuracy, F1-scores, and computational efficiency. Results demonstrate SwAV’s superiority, achieving 98.7% overall accuracy (vs. 92.1% for ResNet-50) and macro-average F1-scores of 98.3% across classes, with a 62% reduction in labeled data requirements. The model showed particular strength in Mild_RSM-class detection (F1-score: 98.5%) and computational efficiency, enabling deployment on edge devices. Statistical validation confirmed significant improvements (p < 0.001) over baseline approaches. These findings establish self-supervised learning as a transformative tool for precision pest management, offering resource-efficient solutions for early infestation detection while maintaining high accuracy. Full article

The piston-type expander (PTE), as the primary output component, significantly influences the performance of an energy storage system. This paper proposes a non-cam variable valve actuation system for the PTE, supported by a mathematical model. An enhanced S-curve trajectory planning method is used to design the valve lift curve. The study investigates the effects of various valve lift design parameters on output power and efficiency at different rotational speeds, employing orthogonal design and SPSS Statistics 27 (Statistical Product and Service Solutions) simulations. A grey comprehensive evaluation method is used to identify optimal valve lift parameters for each speed. The results show that valve lift parameters influence PTE performance to varying degrees, with intake duration having the greatest effect, followed by maximum valve lift, while intake end time has the least impact. The non-cam PTE outperforms the cam-based PTE. At 800 rpm, the optimal design yields 7.12 kW and 53.5% efficiency; at 900 rpm, 8.17 kW and 50.6%; at 1000 rpm, 9.2 kW and 46.8%; and at 1100 rpm, 12.09 kW and 41.2%. At these speeds, output power increases by 18.37%, 11.42%, 11.62%, and 9.82%, while energy efficiency improves by 15.01%, 15.05%, 14.24%, and 13.86%, respectively. Full article

The accelerating integration of electric vehicles (EVs) into contemporary transportation infrastructure has underscored significant limitations in traditional charging paradigms, particularly in accommodating heterogeneous user requirements within dynamic operational environments. This study presents a differentiated optimization framework for EV charging strategies through the systematic classification of user types. A multidimensional decision-making environment is established for three representative user categories—residential, commercial, and industrial—by synthesizing time-variant electricity pricing models with dynamic carbon emission pricing mechanisms. A bi-level optimization architecture is subsequently formulated, leveraging deep reinforcement learning (DRL) to capture user-specific demand characteristics through customized reward functions and adaptive constraint structures. Validation is conducted within a high-fidelity simulation environment featuring 90 autonomous EV charging agents operating in a metropolitan parking facility. Empirical results indicate that the proposed typology-driven approach yields a 32.6% average cost reduction across user groups relative to baseline charging protocols, with statistically significant improvements in expenditure optimization (p < 0.01). Further interpretability analysis employing gradient-weighted class activation mapping (Grad-CAM) demonstrates that the model’s attention mechanisms are well aligned with theoretically anticipated demand prioritization patterns across the distinct user types, thereby confirming the decision-theoretic soundness of the framework. Full article

Understanding species composition and forest dynamics is essential for predicting biomass productivity and informing conservation in tropical montane ecosystems. We evaluated floristic, demographic, and biomass changes in eighteen 0.1 ha permanent plots in the Colombian Sub-Andean forest, including both primary (ca. 60 y old) and secondary forests (ca. 30 years old). Two censuses of individuals (DBH ≥ 2.5 cm) were conducted over 7–13 years. We recorded 516 species across 202 genera and 89 families. Floristic composition differed significantly between forest types (PERMANOVA, p = 0.001), and black oak (Trigonobalanus excelsa Lozano, Hern. Cam. & Henao) forests formed distinct assemblages. Demographic rates were higher in secondary forests, with mortality (4.17% yr), recruitment (4.51% yr), and relative growth rate (0.02% yr) exceeding those of primary forests. The mean aboveground biomass accumulation and the rate of annual change were higher in primary forests (447.5 Mg ha⁻¹ and 466.8 Mg ha⁻¹ yr⁻¹, respectively) than in secondary forests (217.2 Mg ha⁻¹ and 217.2 Mg ha⁻¹ yr⁻¹, respectively). Notably, black oak forests showed the greatest biomass accumulation and rate of change in biomass. Annual net biomass production was higher in secondary forests (8.72 Mg ha⁻¹ yr⁻¹) than in primary forests (5.66 Mg ha⁻¹ yr⁻¹). These findings highlight the ecological distinctiveness and recovery potential of secondary Sub-Andean forests and underscore the value of multitemporal monitoring to understand forest resilience and assess vulnerability to environmental change. Full article

Background/Objectives: This study aimed to assess the long-term clinical performance of full-arch fixed restorations made of fiber-reinforced composite (FRC) supported by four ultra-short implants (4.0 × 5.0 mm) in patients with edentulous, atrophic mandibles. Methods: Ten patients were treated at Sapienza University of Rome and monitored over a 10-year period. Each case involved the placement of four plateau-design implants with a pure conometric connection and a calcium phosphate-treated surface. The final prostheses were fabricated using CAD/CAM-milled Trinia^® fiber-reinforced composite frameworks. Clinical parameters included implant and prosthesis survival, marginal bone level (MBL), peri-implant probing depth (PPD), and patient-reported outcome measures (PROMs). Results: Implant and prosthesis survival reached 100% over the 10-year follow-up. MBL data showed a minor bone gain of approximately 0.11 mm per 5 years (p < 0.0001). PPD remained stable under 3 mm, with a minimal increase of 0.16 mm over the same period (p < 0.0001). PROMs reflected sustained high patient satisfaction. No technical complications, such as chipping or framework fracture, were observed. Conclusions: Rehabilitation of the edentulous mandible with ultra-short implants and metal-free FRC prostheses proved to be a minimally invasive and long-lasting treatment option. The 10-year follow-up confirmed excellent implant and prosthetic outcomes, favorable peri-implant tissue health, and strong patient satisfaction. Nonetheless, further studies with larger sample sizes are needed to confirm these encouraging results and strengthen the clinical evidence. Full article