Search Results (11,706)

In recent years, the classification of underwater images has become one of the most remarkable areas of research in computer vision due to its useful applications in marine sciences, aquatic robotics, and sea exploration. Underwater imaging is pivotal for the evaluation of marine eco-systems, analysis of biological habitats, and monitoring underwater infrastructure. Extracting useful information from underwater images is highly challenging due to factors such as light distortion, scattering, poor contrast, and complex foreground patterns. These difficulties make traditional image processing and machine learning techniques struggle to analyze images accurately. As a result, these challenges and complexities make the classification difficult or poor to perform. Recently, deep learning techniques, especially convolutional neural network (CNN), have emerged as influential tools for underwater image classification, contributing noteworthy improvements in accuracy and performance in the presence of all these challenges. In this paper, we have proposed a modified ResNet-18 model for the binary classification of underwater images into raw and enhanced images. In the proposed modified ResNet-18 model, we have added new layers such as Linear, rectified linear unit (ReLU) and dropout layers, arranged in a block that was repeated three times to enhance feature extraction and improve learning. This enables our model to learn the complex patterns present in the image in more detail, which helps the model to perform the classification very well. Due to these newly added layers, our proposed model addresses various complexities such as noise, distortion, varying illumination conditions, and complex patterns by learning vigorous features from underwater image datasets. To handle the issue of class imbalance present in the dataset, we applied a data augmentation technique. The proposed model achieved outstanding performance, with 96% accuracy, 99% precision, 92% sensitivity, 99% specificity, 95% F1-score, and a 96% Area under the Receiver Operating Characteristic Curve (AUC-ROC) score. These results demonstrate the strength and reliability of our proposed model in handling the challenges posed by the underwater imagery and making it a favorable solution for advancing underwater image classification tasks. Full article

Interpreting diagnostic imaging and identifying clinically relevant features remain challenging tasks, particularly for novice radiologists who often lack structured guidance and expert feedback. To bridge this gap, we propose an Eye-Gaze Guided Multimodal Fusion framework that leverages expert eye-tracking data to enhance learning and decision-making in medical image interpretation. By integrating chest X-ray (CXR) images with expert fixation maps, our approach captures radiologists’ visual attention patterns and highlights regions of interest (ROIs) critical for accurate diagnosis. The fusion model utilizes a shared backbone architecture to jointly process image and gaze modalities, thereby minimizing the impact of noise in fixation data. We validate the system’s interpretability using Gradient-weighted Class Activation Mapping (Grad-CAM) and assess both classification performance and explanation alignment with expert annotations. Comprehensive evaluations, including robustness under gaze noise and expert clinical review, demonstrate the framework’s effectiveness in improving model reliability and interpretability. This work offers a promising pathway toward intelligent, human-centered AI systems that support both diagnostic accuracy and medical training. Full article

This paper investigates the well-posedness of the inverse problem for the time-fractional Burgers equation, which aims to reconstruct initial conditions from terminal observations. Such equations are crucial for the modeling of hydrodynamic phenomena with memory effects. The inverse problem involves inferring initial conditions from terminal observation data, and such problems are typically ill-posed. A framework based on optimal control theory is proposed, addressing the ill-posedness via $H^1$ regularization. Three substantial results are achieved: (1) a rigorous mathematical framework transforming the ill-posed inverse problem into a well-posed optimization problem with proven existence of solutions; (2) theoretical guarantee of solution uniqueness when the regularization parameter is $\alpha >0$ and the stability is of order O($\delta$) with respect to observation noise ($\delta$); and (3) the discovery of a “super-stability” phenomenon in numerical experiments, where the actual stability index (0.046) significantly outperforms theoretical expectations (1.0). Finally, the theoretical framework is validated through comprehensive numerical experiments, demonstrating the accuracy and practical effectiveness of the proposed optimal control approach for the reconstruction of hydrodynamic initial conditions. Full article

Speaker profiling systems are often evaluated on a single corpus, which complicates reliable comparison. We present a fully reproducible evaluation pipeline that trains Convolutional Neural Networks (CNNs) and Long-Short Term Memory (LSTM) models independently on three speech corpora representing distinct recording conditions—studio-quality TIMIT, crowdsourced Mozilla Common Voice, and in-the-wild VoxCeleb1. All models share the same architecture, optimizer, and data preprocessing; no corpus-specific hyperparameter tuning is applied. We perform a detailed preprocessing and feature extraction procedure, evaluating multiple configurations and validating their applicability and effectiveness in improving the obtained results. A feature analysis shows that Mel spectrograms benefit CNNs, whereas Mel Frequency Cepstral Coefficients (MFCCs) suit LSTMs, and that the optimal Mel-bin count grows with corpus Signal Noise Rate (SNR). With this fixed recipe, EfficientNet achieves 99.82% gender accuracy on Common Voice (+1.25 pp over the previous best) and 98.86% on VoxCeleb1 (+0.57 pp). MobileNet attains 99.86% age-group accuracy on Common Voice (+2.86 pp) and a 5.35-year MAE for age estimation on TIMIT using a lightweight configuration. The consistent, near-state-of-the-art results across three acoustically diverse datasets substantiate the robustness and versatility of the proposed pipeline. Code and pre-trained weights are released to facilitate downstream research. Full article

Traditional Foreign Object Debris (FOD) detection methods face challenges such as difficulties in large-size data acquisition and the ineffective application of detection algorithms with high accuracy. In this paper, image data augmentation was performed using generative adversarial networks and diffusion models, generating images of monitoring areas under different environmental conditions and FOD images of varied types. Additionally, a three-stage image blending method considering size transformation, a seamless process, and style transfer was proposed. The image quality of different blending methods was quantitatively evaluated using metrics such as structural similarity index and peak signal-to-noise ratio, as well as Depthanything. Finally, object detection models with a similarity distance strategy (SimD), including Faster R-CNN, YOLOv8, and YOLOv11, were tested on the dataset. The experimental results demonstrated that realistic FOD data were effectively generated. The Structural Similarity Index Measure (SSIM) and Peak Signal-to-Noise Ratio (PSNR) of the synthesized image by the proposed three-stage image blending method outperformed the other methods, reaching 0.99 and 45 dB. YOLOv11 with SimD trained on the augmented dataset achieved the mAP of 86.95%. Based on the results, it could be concluded that both data augmentation and SimD significantly improved the accuracy of FOD detection. Full article

Current mainstream classification methods predominantly employ end-to-end multi-class frameworks. These approaches encounter inherent challenges including high-dimensional feature space complexity, decision boundary ambiguity that escalates with increasing class cardinality, sensitivity to label noise, and limited adaptability to dynamic model expansion. However, human beings may avoid these mistakes naturally. Research indicates that humans subconsciously employ a decision-making process favoring binary outcomes, particularly when responding to questions requiring nuanced differentiation. Intuitively, responding to binary inquiries such as “yes/no” often proves easier for humans than addressing queries of “what/which”. Inspired by the human decision-making hypothesis, we proposes a decision paradigm named the evolutionary binary decision framework (EBDF) centered around binary classification, evolving from traditional multi-classifiers in deep learning. To facilitate this evolution, we leverage the top-N outputs from the traditional multi-class classifier to dynamically steer subsequent binary classifiers, thereby constructing a cascaded decision-making framework that emulates the hierarchical reasoning of a binary decision tree. Theoretically, we demonstrate mathematical proof that by surpassing a certain threshold of the performance of binary classifiers, our framework may outperform traditional multi-classification framework. Furthermore, we conduct experiments utilizing several prominent deep learning models across various image classification datasets. The experimental results indicate significant potential for our strategy to surpass the ceiling in multi-classification performance. Full article

Background: Hearing loss is increasingly prevalent and poses a significant public health concern. While both aging and occupational noise exposure are recognized contributors, their interactive effects and gender-specific patterns remain underexplored. Methods: This cross-sectional study analyzed data from 135,251 employees in Jiangsu Province, China. Demographic information, noise exposure metrics, and hearing thresholds were obtained through field measurements, questionnaires, and audiometric testing. Multivariate logistic regression, restricted cubic spline modeling, and interaction analyses were conducted. Machine learning models were employed to assess feature importance. Results: A nonlinear relationship between age and high-frequency hearing loss (HFHL) was identified, with a critical inflection point at 37.8 years. Noise exposure significantly amplified HFHL risk, particularly in older adults (OR = 2.564; 95% CI: 2.456–2.677, p < 0.001), with consistent findings across genders. Men exhibited greater susceptibility at high frequencies, even after adjusting for age and co-exposures. Aging and noise exposure have a joint association with hearing loss (OR = 2.564; 95% CI: 2.456–2.677, p < 0.001) and an interactive association (additive interaction: RERI = 2.075, AP = 0.502, SI = 2.967; multiplicative interaction: OR = 1.265; 95% CI: 1.176–1.36, p < 0.001). And machine learning also confirmed age, gender, and noise exposure as key predictors. Conclusions: Aging and occupational noise exert synergistic effects on auditory decline, with distinct gender disparities. These findings highlight the need for integrated, demographically tailored occupational health strategies. Machine learning approaches further validate key risk factors and support targeted screening for hearing loss prevention. Full article

Erosion gullies can reduce arable land area and decrease agricultural machinery efficiency; therefore, automatic gully extraction on a regional scale should be one of the preconditions of gully control and land management. The purpose of this study is to compare the effects of the grey level co-occurrence matrix (GLCM) and topographic–hydrologic features on automatic gully extraction and guide future practices in adjacent regions. To accomplish this, GaoFen-2 (GF-2) satellite imagery and high-resolution digital elevation model (DEM) data were first collected. The GLCM and topographic–hydrologic features were generated, and then, a gully label dataset was built via visual interpretation. Second, the study area was divided into training, testing, and validation areas, and four practices using different feature combinations were conducted. The DeepLabV3+ and ResNet50 architectures were applied to train five models in each practice. Thirdly, the trainset gully intersection over union (IOU), test set gully IOU, receiver operating characteristic curve (ROC), area under the curve (AUC), user’s accuracy, producer’s accuracy, Kappa coefficient, and gully IOU in the validation area were used to assess the performance of the models in each practice. The results show that the validated gully IOU was 0.4299 (±0.0082) when only the red (R), green (G), blue (B), and near-infrared (NIR) bands were applied, and solely combining the topographic–hydrologic features with the RGB and NIR bands significantly improved the performance of the models, which boosted the validated gully IOU to 0.4796 (±0.0146). Nevertheless, solely combining GLCM features with RGB and NIR bands decreased the accuracy, which resulted in the lowest validated gully IOU of 0.3755 (±0.0229). Finally, by employing the full set of RGB and NIR bands, the GLCM and topographic–hydrologic features obtained a validated gully IOU of 0.4762 (±0.0163) and tended to show an equivalent improvement with the combination of topographic–hydrologic features and RGB and NIR bands. A preliminary explanation is that the GLCM captures the local textures of gullies and their backgrounds, and thus introduces ambiguity and noise into the convolutional neural network (CNN). Therefore, the GLCM tends to provide no benefit to automatic gully extraction with CNN-type algorithms, while topographic–hydrologic features, which are also original drivers of gullies, help determine the possible presence of water-origin gullies when optical bands fail to tell the difference between a gully and its confusing background. Full article

Underwater image enhancement (UIE) technology plays a vital role in marine resource exploration, environmental monitoring, and underwater archaeology. However, due to the absorption and scattering of light in underwater environments, images often suffer from blurred details, color distortion, and low contrast, which seriously affect the usability of underwater images. To address the above limitations, a lightweight transformer-based model (LITM) is proposed for improving underwater degraded images. Firstly, our proposed method utilizes a lightweight RGB transformer enhancer (LRTE) that uses efficient channel attention blocks to capture local detail features in the RGB domain. Subsequently, a lightweight HSV transformer encoder (LHTE) is utilized to extract global brightness, color, and saturation from the hue–saturation–value (HSV) domain. Finally, we propose a multi-modal integration block (MMIB) to effectively fuse enhanced information from the RGB and HSV pathways, as well as the input image. Our proposed LITM method significantly outperforms state-of-the-art methods, achieving a peak signal-to-noise ratio (PSNR) of 26.70 and a structural similarity index (SSIM) of 0.9405 on the LSUI dataset. Furthermore, the designed method also exhibits good generality and adaptability on unpaired datasets. Full article

To address the diverse requirements of accuracy and robustness in integrated navigation for unmanned aerial vehicles, an interactive robust filter algorithm that integrates the interactive multiple model concept and leverages the complementary applicability of the strong tracking filter and the smooth variable structure filter is proposed. The algorithm operates as follows: the strong tracking filter, along with the smooth variable structure filter, operates side by side with distinct models. During the filter process, the likelihood function is utilized to update the filter probabilities and determine the weights for each one of the filters. Input interaction, coupled with output fusion, is then carried out. The results of the experiments validate that the presented interactive filter algorithm significantly reduces estimation errors. When confronted with complex, dynamic noise environments and system uncertainties, it retains high-precision state estimation while demonstrating markedly improved robustness. The proposed interactive robust filter algorithm is compared against the strong tracking filter, smooth variable structure filter, and strong tracking smooth filter. Taking the strong tracking smooth filter, which has the highest accuracy among the three, as the reference baseline, the presented interactive robust filter algorithm achieves over 16% improvement in velocity accuracy and over 40% improvement in position accuracy. Full article

High-pitched noise generated by dental air turbine handpieces (ATHs) causes discomfort and anxiety, discouraging dental visits. Understanding the time-dependent noise generation mechanism associated with compressed airflow in ATHs is crucial for effective noise reduction. However, the direct investigation of airflow dynamics within ATHs is challenging. The transient-state modeling of computational fluid dynamics (CFD) simulations remains unexplored owing to the complexities of high rotational speeds and air compressibility. This study develops a novel CFD framework for transient (time-dependent) modeling under high-speed rotational conditions. Simulations were performed using a three-dimensional model reconstructed from a commercial ATH. Simulations were conducted at 320,000 rpm using a novel framework that combines the immersed boundary and building cube methods. A fine 0.025 mm mesh spacing near the ATH, combined with supercomputing resources, enabled the simulation of hundreds of millions of cells. The simulation results were validated using experimental noise measurements. The CFD simulation revealed transient airflow and aeroacoustic behavior inside and around the ATH that closely matched the prominent frequency peaks from the experimental data. This study is the first to simulate the transient airflow of ATHs. The proposed CFD model can accurately predict aeroacoustics, contributing to the future development of quieter and more efficient dental devices. Full article

The rapid densification of wireless networks demands efficient evaluation of special functions underpinning system-level performance metrics. To facilitate research, we introduce a computational framework tailored for the zero-balanced Gauss hypergeometric function $\Psi (x,y)\triangleq {}_{2}F_1(1,x;1+x;-y)$, a fundamental mathematical kernel emerging in Signal-to-Interference-plus-Noise Ratio (SINR) coverage analysis of non-uniform cellular deployments. Specifically, we propose a novel Reciprocal-Argument Transformation Algorithm (RTA), derived rigorously from a Mellin–Barnes reciprocal-argument identity, achieving geometric convergence with $O\left(1/y\right)$. By integrating RTA with a Pfaff-series solver into a hybrid algorithm guided by a golden-ratio switching criterion, our approach ensures optimal efficiency and numerical stability. Comprehensive validation demonstrates that the hybrid algorithm reliably attains machine-precision accuracy ($\sim {10}^{-16}$) within 1 μs per evaluation, dramatically accelerating calculations in realistic scenarios from hours to fractions of a second. Consequently, our method significantly enhances the feasibility of tractable optimization in ultra-dense non-uniform cellular networks, bridging the computational gap in large-scale wireless performance modeling. Full article

This research experimentally and numerically evaluates the effectiveness of viscoelastic (VE) and rotational friction (RF) dampers in enhancing the seismic performance of coupled shear wall (CSW) systems. This study consists of two phases: (1) element testing to characterize the hysteretic behavior and energy dissipation capacity of VE and RF dampers, and (2) shake table testing of a large-scale CSW structure equipped with these dampers under the white noise, sinusoidal and Kokuji waves. The experimental results are validated through numerical analysis using STERA 3D (version 11.5), a nonlinear finite element software, to simulate the dynamic response of the damped CSW system. Key performance indicators, including inter-story drift, base shear, and energy dissipation, are compared between experimental and numerical results, demonstrating strong correlation. The findings reveal that VE dampers effectively control high-frequency vibrations, while RF dampers provide stable energy dissipation across varying displacement amplitudes. The validated numerical model facilitates the optimization of damper configurations for performance-based seismic design. This study provides valuable insights into the selection and implementation of supplemental damping systems for CSW structures, contributing to improved seismic resilience in buildings. Full article

Compressed sensing is widely used in modern resource-constrained sensor networks. However, achieving high-quality and robust signal reconstruction under low sampling rates and noise interference remains challenging. Traditional CS methods have limited performance, so many deep learning-based CS models have been proposed. Although these models show strong fitting capabilities, they often lack the ability to handle complex noise in sensor networks, which affects their performance stability. To address these challenges, this paper proposes SBCS-Net. This framework innovatively expands the iterative process of sparse Bayesian compressed sensing using convolutional neural networks and Transformer. The core of SBCS-Net is to optimize key SBL parameters through end-to-end learning. This can adaptively improve signal sparsity and probabilistically process measurement noise, while fully leveraging the powerful feature extraction and global context modeling capabilities of deep learning modules. To comprehensively evaluate its performance, we conduct systematic experiments on multiple public benchmark datasets. These studies include comparisons with various advanced and traditional compressed sensing methods, comprehensive noise robustness tests, ablation studies of key components, computational complexity analysis, and rigorous statistical significance tests. Extensive experimental results consistently show that SBCS-Net outperforms many mainstream methods in both reconstruction accuracy and visual quality. In particular, it exhibits excellent robustness under challenging conditions such as extremely low sampling rates and strong noise. Therefore, SBCS-Net provides an effective solution for high-fidelity, robust signal recovery in sensor networks and related fields. Full article

To address the decline in measurement accuracy of magnetometers due to process errors and environmental interference, as well as the insufficient robustness of traditional calibration algorithms under strong interference conditions, this paper proposes an ellipsoid fitting algorithm based on Dynamic Adaptive Elite Particle Swarm Optimization (DAEPSO). The proposed algorithm integrates three enhancement mechanisms: dynamic stratified elite guidance, adaptive inertia weight adjustment, and inferior particle relearning via Lévy flight, aiming to improve convergence speed, solution accuracy, and noise resistance. First, a magnetometer calibration model is established. Second, the DAEPSO algorithm is employed to fit the ellipsoid parameters. Finally, error calibration is performed based on the optimized ellipsoid parameters. Our simulation experiments demonstrate that compared with the traditional Least Squares Method (LSM) the proposed method reduces the standard deviation of the total magnetic field intensity by 54.73%, effectively improving calibration precision in the presence of outliers. Furthermore, when compared to PSO, TSLPSO, MPSO, and AWPSO, the sum of the absolute distances from the simulation data to the fitted ellipsoidal surface decreases by 53.60%, 41.96%, 53.01%, and 27.40%, respectively. The results from 60 independent experiments show that DAEPSO achieves lower median errors and smaller interquartile ranges than comparative algorithms. In summary, the DAEPSO-based ellipsoid fitting algorithm exhibits high fitting accuracy and strong robustness in environments with intense interference noise, providing reliable theoretical support for practical engineering applications. Full article