Search Results (538)

Thedeep score-based diffusion (DeScoD) model performs well in electrocardiogram (ECG) denoising tasks. However, due to the theoretical error lower bound in approximating functions with linear transformations, it often lacks flexibility when fitting non-stationary noise, baseline wander, or morphologically variable features such as QRS complexes in ECG signals. In this paper, we propose a Kolmogorov–Arnold network enhanced deep score-based diffusion (KAN-DeScoD) model, which is the first to integrate Kolmogorov–Arnold network (KAN) layers into an ECG denoising diffusion model. By leveraging KAN’s adaptive activation functions, which more finely capture the complex structures within ECG signals, the model’s robustness in high-noise environments, as well as the accuracy and stability of signal reconstruction, are improved. We validate the effectiveness of the proposed method on the QT Database and the MIT-BIH Noise Stress Test Database (NSTDB). Experimental results show that under different shots and noise intensities, ours outperforms the DeScoD model across multiple metrics. The research results demonstrate the effectiveness of introducing KAN, which improves the model’s robustness in high-noise environments and the accuracy of signal reconstruction. Full article

The anomalous logging responses caused by complex geological and downhole engineering conditions, which can be the expansion of a borehole, the formation of fractures, and the mud intrusion, usually result in the absence of some important curves and undermine the accuracy of the reservoir evaluation. The strong nonlinearity and non-stationarity of the log curves remain problematic to conventional interpolation and statistical techniques; the traditional models do not take into account any sequential relationship between points along the depth axis, whereas the deep sequence models can only regress on the points, which limits their capability of ensuring the overall geological consistency. In order to resolve these difficulties, this paper introduces a Geology-Balanced Time Series Conditional Generative Adversarial Network (GC-TSGAN) in which the lithological data is converted into an initial state in the form of prior conditions and is input into both the generator and the discriminator. The model uses LSTM to learn depth-sequential dependencies and a BCE GAN-based adversarial loss to achieve distributional consistency and local morphological fidelity. Hyperparameter tuning is used with the help of random search and Bayesian optimization. The logging data of 41 wells in the B Basin, Chad, are experimented using GC-TSGAN alongside baseline models such as RF, XGBoost, LSTM and ANN; GC-TSGAN is proven to be much better than baseline models in terms of the RMSE, MAE, and squares of predicate and value. The findings confirm that the proposed model can effectively reconstruct log curves with high precision even in a complicated geological environment, thereby providing quality data for performing geological modeling and evaluating the reservoirs. Full article

The precise, real-time delineation of rice lodging areas constitutes a fundamental prerequisite for the adaptive operation of unmanned combine harvesters. However, existing deep learning methods struggle to resolve a critical limitation: achieving an optimal equilibrium between robust regional morphological perception—which is crucial for irregular lodging patterns—and the ultra-low computational overhead demanded by resource-constrained edge terminals. To address this specific constraint, StarNet-RiceSeg is proposed as a lightweight semantic segmentation network explicitly tailored for unmanned harvesters. Initially, the architecture incorporates the minimalist StarNet as its backbone. By leveraging the unique “Star Operation,” it implicitly maps features into a high-dimensional nonlinear space, thereby significantly augmenting feature discriminability while drastically curtailing computational overhead. Furthermore, to mitigate the misdetection issues stemming from the textural similarity between lodged and upright rice, the Rice Spatial Attention (RSA) module was designed. By intensifying feature interaction within the spatial dimension, this module steers the network to focus on the cohesive morphology of lodged regions while effectively suppressing background noise. Experiments conducted on a self-constructed high-resolution rice lodging dataset demonstrate that StarNet-RiceSeg achieves a mIoU of 94.42%, significantly outperforming mainstream models such as U-Net, DeepLabV3+, SegNet and HRNet. Notably, the model maintains a compact footprint with only 8.01 million parameters and a computational load as low as 9.32 GFLOPs. Following optimization with TensorRT, the system achieved a real-time inference speed of 32.51 FPS on the NVIDIA Jetson Xavier NX embedded platform. These results indicate that StarNet-RiceSeg provides a high-precision, low-latency solution for perceiving rice lodging areas in complex field environments, facilitating unmanned precision harvesting. Full article

Bone marrow cell classification and nucleus segmentation in microscopic images are fundamental tasks for computer-aided diagnosis of lymphocytic leukemia. However, bone marrow cells from different subtypes exhibit high morphological similarity, and structural information is often constrained under optical microscopic imaging, posing challenges for stable and effective feature representation. To address this issue, we propose MTLL (Multitask Model on Lymphocytic Leukemia), a novel multitask approach that performs cell classification and nucleus segmentation within a unified network to exploit their complementary information. The model constructs a hybrid backbone for shared feature representation based on a CNN-Transformer architecture, in which Fuse-MBConv modules are tightly integrated with multilayer multi-scale transformers to enable deep fusion of local texture and global semantic information. For the segmentation branch, we design an AM (Atrous Multilayer Perceptron) decoder that combines atrous spatial pyramid pooling with multilayer perceptrons to fuse multi-scale information and accurately delineate nucleus boundaries. The classification branch incorporates prior knowledge of cell nuclei structures to capture subtle variations in cellular morphology and texture, thereby enhancing the model’s ability to distinguish between leukemia subtypes. Experimental results demonstrate that the MTLL model significantly outperforms existing advanced single-task and multi-task models in both lymphocytic leukemia classification and cell nucleus segmentation. These results validate the effectiveness of the multi-task feature-sharing strategy for lymphocytic leukemia diagnosis using bone marrow microscopic images. Full article

The detection of surface defects on smartphone components is a critical step in quality assurance for industrial manufacturing. However, existing deep learning-based methods struggle with the extreme variations in defect morphology and scale, while labeled training data remains scarce due to the high cost of expert annotation. To address these challenges, we propose a twofold solution. First, we introduce MSAdaNet, a Multi-Scale Adaptive Defect Detection Network, which integrates three novel modules: a Parallel Multi-Scale Feature Aggregation (PMSFA) backbone, a Focusing Diffusion Pyramid Network (FDPN) neck, and a Scale-Adaptive Shared Detection (SASD) head. Second, to combat data scarcity, we propose a novel data generation pipeline, creating the synthetic Smartphone Camera Bezel Dataset (SCBD) of 4936 images. Extensive experiments on both real-world and synthetic datasets validate our approach. On the challenging public SSGD, MSAdaNet achieves a state-of-the-art mAP@0.5 of 54.8%, outperforming prominent frameworks and improving upon the strong YOLOv11m baseline by +10.6 points in mAP@0.5 and +18.3 points in recall. Furthermore, on our synthetic SCBD, the model achieves an impressive 94.0% mAP@0.5, confirming the quality of our data generation pipeline and the robustness of our architecture across different data distributions. Ablation studies systematically confirm the significant contribution of each proposed module, validating MSAdaNet as an effective and efficient solution for industrial defect detection. Full article

Background: The Magnetic Resonance Imaging (MRI)-based tumor segmentation remains a challenging problem in medical imaging due to tumor heterogeneity, unpredictable morphological features, and the high complexity of calculations needed to implement it in clinical practice, putting it out of the scope of real-time applications. Although neural networks have significantly improved segmentation performance, they still struggle to capture morphological tumor features while maintaining computational efficiency. This work introduces Hybrid Adaptive Attention U-Net (HAAU-Net) framework, combining context-aware morphologically stable features and spatial channel attention to achieve high-quality tumor segmentation with less computational cost. Methods: The proposed HAAU-Net framework integrates multi-scale Adaptive Attention Blocks (AAB), Context-Aware Morphological Feature Module (CAMFM) and Spatial-Channel Hybrid Attention Mechanism (SCHAM). CAMFM is used to maintain the stability of morphological features by hierarchical aggregation and dynamic normalization of features. SCHAM enhances feature representation by modelling channels and spatial regions where the strongest feature are determined to use in segmentation. On the BRaTS 2022/2023 data, the proposed HAAU-Net is evaluated using four modalities including T1, T1GD, T2 and T2-FLAIR sequences. Results: The proposed model able to obtain 96.8% segmentation accuracy with a Dice coefficient of 0.89 on the entire tumor region, outperforming the alternative U-Net (0.83) and conventional CNN methods of segmentation (0.81). The proposed HAAU-Net architecture cuts the computational complexity of the standard deep learning models by 43% and still achieve real-time inference (28 FPS on a regular GPU). The hybrid model used to predict survival has a C-Index of 0.91 which is higher than the traditional SVM-based methods (0.72). Conclusions: Spatial-channel attention, combined with morphologically stable features, can be combined to allow clinically significant interpretability in attention maps. The proposed framework significantly improves segmentation performance while maintaining computational effeciency. This broad system has a serious potential of AI-enabled clinical decision support system and early prognostic diagnosis in neuro-oncology with practical deployment capability. Full article

Urbanization significantly enhances socio-economic conditions, health, and well-being for many by improving access to services, education, and economic opportunities. However, socio-economic and public health disparities are also being exacerbated by urbanization. The reliable data required to monitor these conditions are often unavailable, outdated, or inconsistent. This study introduces a multimodal deep learning framework that integrates satellite imagery with street network datasets to predict urban socio-economic indicators and public health determinants at the sector level as a political administrative unit of public health planning in Rwanda. We extracted latent visual and topological embeddings of the urban built environment, using a Convolutional Neural Network (CNN) and Graph Neural Network (GNN). These embeddings were fused through an attentional mechanism to train a multi-task regression model that simultaneously predicts multiple socio-economic indicators and public health determinants. This framework was applied to the City of Kigali in Rwanda. Overall, the multimodal fusion model achieved the best average performance across targets, with an average correlation of 0.68 and MAE of 1.26 for socio-economic indicators, and 0.68 and 1.46 for public health determinants, demonstrating the benefit of integrating visual and topological information. The learned fused embedding space arranges socio-economic indicators and public health determinant deciles along a continuous morphological gradient from sparsely built rural settings to dense urban settings, demonstrating that the urban form encodes latent signals that capture socio-economic indicators and health determinants. Moreover, the study reveals a strong relationship between socio-economic indicators and the public health index, with education, cooking materials, and floor materials exhibiting a correlation above 0.96. This work demonstrates the utility of an integrated framework for socio-economic indicator profiling and public health planning in data-scarce urban contexts, offering a scalable approach for monitoring the indicators of Sustainable Development Goals in rapidly changing urban environments. Full article

This study introduces C-EMDNet, a nonlinear speech denoising approach that combines the adaptive decomposition capabilities of Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) and a deep convolutional architecture operating directly in the time-intrinsic mode function (IMF) domain. Unlike conventional enhancement methods that rely on fixed time–frequency representations, such as the short-time Fourier transform (STFT), the proposed approach interprets CEEMDAN IMFs as a morphological latent space that captures the multi-scale structure of speech. A U-Net-like network was trained to estimate mode-wise masks, enabling selective noise suppression while preserving the harmonic and formant structures. Experiments on standard noisy speech datasets show that C-EMDNet outperforms classical denoising algorithms and competitive deep learning baselines. These results highlight the promise of nonlinear morphological representations for an alternative framework speech enhancement. Full article

In the field of forensic document examination, accurately determining the chronological sequence of intersecting lines between seal ink and handwriting is a crucial technical step for verifying document authenticity, identifying contract tampering, and detecting forged signatures. This technique analyzes the physical superimposition relationship formed by the deposition of the two media on the paper substrate to provide objective scientific evidence for judicial practice. Although traditional methods such as microscopic imaging and mass spectrometry analysis have achieved some progress, they still suffer from common limitations including high equipment costs, complex operation, and potential damage to samples. This study proposes and validates an innovative non-destructive determination method that integrates structured light 3D reconstruction technology with deep learning algorithms. The research captures the microscopic 3D morphological features of the ink intersection area using a high-precision structured light scanning system and effectively eliminates noise interference caused by paper substrate undulation through Gaussian flattening technology. Subsequently, a multimodal fusion strategy combines 2D texture images with 3D depth information to construct a dataset rich in features. On this basis, a deep learning model based on an improved Residual Neural Network (ResNet) is designed, incorporating the ELU activation function and an EMA mechanism to enhance the model’s feature extraction capability and convergence stability. Experimental results demonstrate that the proposed method achieves a recognition accuracy of 94.39% on the test set, fully validating its effectiveness and application potential in the non-destructive determination of ink stroke sequencing. Full article

Accurate detection of cardiac arrhythmias from electrocardiograms remains challenging for rare rhythm classes due to class imbalance and morphological variability. We present a hybrid deep learning framework combining per-lead convolutional encoders with a cross-lead transformer that models relationships across different lead signals through self-attention, accepting variable lead configurations. To address minority-class scarcity, a generative adversarial network synthesizes physiologically plausible beat segments for underrepresented arrhythmias. Attention-based visualizations localize influential waveform regions aligned with clinically meaningful structures. Post-training pruning and INT8 quantization enable efficient deployment with minimal performance loss. Extensive experiments on the MIT-BIH Arrhythmia Database across sixteen heartbeat classes from two-lead recordings yield exceptional results over ten independent runs: accuracy of 99.67%, F1-score of 99.66%, and AUC of 99.8%. External validation on the ECG5000 single-lead dataset and the St Petersburg INCART twelve-lead dataset confirms robust generalizability with F1-scores of 97.6% and 98% respectively. Our framework delivers accurate, interpretable, stable, and deployable arrhythmia detection across diverse clinical settings. Full article

Liquid metal embrittlement (LME) during hot stamping of Zn-coated high-strength steels poses significant challenges to the long-term durability of automotive components. This study investigates how ~30 μm deep LME cracks affect the mechanical behavior of Zn-coated high-strength boron steel. LME-free flat specimens were compared with hat-shaped specimens containing LME cracks. While tensile strength and ductility exhibited minimal changes, the high-cycle fatigue limit (R = −1, 10⁷ cycles) decreased by 10.9% from 550 MPa to 490 MPa in hat-shaped specimens. Fractographic examination revealed distinct stress-dependent crack initiation mechanisms: at high stress amplitudes (≥690 MPa), LME cracks competed with intrinsic substrate defects but did not dominate fatigue failure. In contrast, at moderate-to-low stress amplitudes (≤630 MPa), LME cracks dominated fatigue degradation through a multi-site crack initiation tendency. El Haddad analysis positioned these cracks at the short-to-long crack transition boundary (l ≈ l₀). Preliminary fracture mechanics analysis reveals that conventional single-crack LEFM models systematically overestimate the fatigue threshold stress for LME-affected specimens, a discrepancy qualitatively attributed to the high surface density and morphological complexity of LME crack networks and to chemically assisted grain boundary weakening induced by liquid Zn infiltration—effects not captured by standard fracture mechanics frameworks. These results establish the stress-dependent mechanisms governing LME crack-induced fatigue degradation and provide a mechanistic basis for the development of more accurate fatigue life prediction methods for Zn-coated hot-stamped high-strength steels. Full article

The morphological identification of lepidopteran pest pupae has long been a difficult task. To explore automated solutions, this study established a standardized, multi-angle image dataset of pupae from 11 economically important lepidopteran pests. We then systematically evaluated six deep learning models, including both convolutional neural networks and Transformer architectures. The results show that all models successfully learned to distinguish the vast majority of species, with Vit-Small achieving the highest accuracy (98.71 ± 0.16%) and the highest F1-score (98.69 ± 0.20%). This confirms that pupal morphology provides sufficient discriminative visual information to support highly accurate automated identification. However, all models exhibited consistent, minor confusion among Helicoverpa armigera, Mythimna separata and Spodoptera exigua. Analysis revealed these errors originated from specific viewing angles of a limited number of specimens, underscoring the value of the multi-angle imaging protocol used in this study. This study transforms pupal identification from a traditional taxonomic difficulty into a solvable computer vision task, providing a dataset, methodological benchmarks, and a feasibility validation for developing image-based tools for pupal-stage pest surveillance. Full article

This study applies deep learning to classify façade features in Tel Aviv’s market-developed apartment housing (1980s–1990s), a vast landscape typically excluded from architectural history due to its non-iconic character. We constructed a curated corpus of 877 expert-labeled high-resolution façade images and evaluated whether convolutional neural networks can detect historically meaningful patterns at urban scale. Focusing on the “staggered balcony” motif—linked to national regulation 5442/1992—we show that a ConvNeXt-Tiny model achieved robust classification performance (96.6% accuracy, 90.3% F1) after rigorous dataset curation and expert relabeling. Initial experiments on noisier data produced inconsistent results, underscoring the importance of domain expertise in operationalizing historical categories. Rather than treating machine learning as definitive classification, we present an iterative workflow where architectural historians use model outputs to refine categories, test morphological hypotheses, and identify overlooked variations. The findings demonstrate how CNN-based analysis can advance empirical research on non-iconic built environments and open methodological pathways for cultural heritage studies and digital architectural humanities. Full article

Colorectal cancer (CRC) remains a leading cause of cancer mortality globally, yet current histopathological diagnostics capture only limited features. This study aimed to discover subtle, prognostically significant histomorphological patterns in CRC tissues using unsupervised deep learning. We developed a framework integrating convolutional neural networks with deep clustering, trained on 23,341 image patches from 493 patients. We identified 30 distinct histomorphological clusters from CRC tissue images. Through univariate and multivariate survival analyses, three clusters (Cluster13, Cluster19, and Cluster24) were consistently associated with patient prognosis. These clusters were integrated with clinical factors (T stage, N stage, and differentiation degree) to construct a prognostic risk model. Patients stratified into high-risk and low-risk groups based on model predictions showed significant survival differences in both the training set (N = 493) and an independent validation set (N = 2590). Furthermore, logistic regression and multivariate Cox analyses demonstrated that incorporating the three histomorphological clusters alongside clinical factors yielded a modest but statistically significant improvement in predictive performance compared to clinical factors alone, indicating their complementary value for prognosis. This work demonstrates that computational pathology can uncover novel, visually elusive morphological features with independent prognostic value, offering potential to refine CRC patient stratification and inform clinical decision-making. Full article

Background/Objectives: Biopsy examination remains the gold standard for cancer diagnosis, relying on histopathological assessment of tissue samples to identify malignant changes. However, manual interpretation of histopathological slides is time-consuming, subjective, and susceptible to inter-observer variability. The digitization of histopathological images enables automated analysis and offers opportunities to support clinicians with more consistent and objective diagnostic tools. This study aims to enhance cancer diagnosis by proposing a hybrid framework that integrates deep-learning-based histopathological image analysis with Whispering Gallery Mode (WGM) optical sensing for complementary tissue characterization. Methods: The proposed framework combines automated tumor classification from histopathological images with biochemical signal analysis obtained from WGM optical sensors. Deep learning models, including EfficientNet-B0, InceptionV3, and Vision Transformer (ViT), were employed for binary and multi-class tumor classification using the BreakHis dataset. To address class imbalance, a Deep Convolutional Generative Adversarial Network (DCGAN) was utilized to generate synthetic histopathological images alongside conventional data augmentation techniques. In parallel, WGM optical sensors were incorporated to capture subtle tissue-specific signatures, with machine learning algorithms enabling automated feature extraction and classification of the acquired signals. Results: In multi-class classification, InceptionV3 combined with DCGAN-based augmentation achieved an accuracy of 94.45%, while binary classification reached 96.49%. Fine-tuned Vision Transformer models achieved a higher classification accuracy of 98% on the BreakHis dataset. The integration of WGM optical sensing provided additional biochemical information, offering complementary insights to image-based analysis and supporting more robust diagnostic decision-making. Conclusions: The proposed hybrid framework demonstrates the potential of combining deep-learning-based histopathological image analysis with WGM optical sensing to improve the accuracy and reliability of cancer classification. By integrating morphological and biochemical information, the framework offers a promising approach for enhanced, objective, and supportive cancer diagnostic systems. Full article