Search Results (504)

Paratuberculosis is a chronic infectious disease associated with substantial economic losses in livestock production. Histopathological examination remains an important diagnostic approach but can be time-consuming and dependent on specialist expertise, motivating the development of automated and interpretable image-classification methods. This study evaluated an explainable framework integrating a pretrained Swin-Tiny Transformer, handcrafted Gray-Level Co-occurrence Matrix (GLCM) and Local Binary Pattern (LBP) texture descriptors, and XGBoost classification for paratuberculosis histopathology image analysis. Following duplicate screening, 349 unique images comprising 199 MAP-positive and 150 MAP-negative samples were evaluated using stratified image-level five-fold cross-validation. Four model configurations were compared to assess the independent and incremental contributions of the learned and handcrafted feature representations. The standalone Swin-Tiny model achieved the highest mean ROC–AUC of $0.979\pm 0.015$, while the Swin-embedding XGBoost and hybrid Swin + GLCM/LBP + XGBoost models achieved mean ROC–AUC values of $0.977\pm 0.016$ and $0.977\pm 0.017$, respectively. The GLCM/LBP-only model achieved a mean ROC–AUC of $0.934\pm 0.041$, indicating that the handcrafted texture descriptors contained independently discriminative information but provided limited incremental value when combined with the Swin embeddings. Grad-CAM and XGBoost feature-importance analyses provided image-level and feature-level insights into model predictions. These findings demonstrate the effectiveness of Swin-Tiny representations for paratuberculosis histopathology image classification while highlighting the need for external validation using larger, independently sourced datasets. Full article

Segmentation of mining-disturbed land is important for eco-geological environment monitoring. Although existing methods possess strong segmentation capabilities, mining-disturbed land exhibits irregular edges, different spatial sizes, and global texture variability, which lead to difficulties in extracting discriminative features, thereby limiting segmentation accuracy. This study built an RGB-based binary semantic segmentation dataset, covering typical mining-disturbed land in Fujian Province of China. Then a spatial prompt and wavelet Mamba-based multi-scale cross-domain feature fusion network (SWDF-Net) was proposed. (1) Wavelet Mamba-based dual-frequency collaborative enhancement module: high-low frequency information was decoupled by wavelet transform, and collaboratively enhanced by fusion of multiple local details and Mamba-guided global information. It can highlight the high-frequency edge features and low-frequency texture patterns of mining-disturbed land. (2) Cross-domain feature alignment and fusion module: cross-domain statistical calibration and channel conditional modulation were used to narrow spatial–frequency feature distribution gap, which facilitates cross-domain feature alignment and fusion. (3) Spatial prompt-based multi-scale feature weighted fusion module: pixel-level weight maps were generated by spatial prior prompt derived from a spatial decoder and edge-gated branch, which adaptively fuses former multi-scale dual-domain features. The SWDF-Net achieved the best Intersection over Union of 72.22% for mining-disturbed land and performed competitively on the ISPRS Vaihingen and Potsdam datasets. Full article

This study examines patterns of ridership recovery across Seoul subway stations following the COVID-19 pandemic and identifies the factors associated with recovery stagnation. To this end, we apply cluster analysis, Markov transition analysis, and binary logistic regression using subway ridership data from 2019 to 2025. We classify station-level recovery into four types: Entrenched Low-Recovery, Partial Recovery, Improved Recovery, and Over-Recovery. The results reveal substantial variation in recovery trajectories across stations. Entrenched Low-Recovery stations display strong path dependence and state persistence, whereas Improved Recovery and Over-Recovery stations exhibit lower state stability and higher transition probabilities. In addition, recovery in morning commuting and evening travel, restaurant density, and average sales per store reduce the likelihood of recovery stagnation, whereas stations with stronger transfer functions are more likely to remain in a low-recovery state. These findings indicate that post-pandemic urban rail recovery is not a uniform return to pre-pandemic ridership levels, but a heterogeneous and dynamic process shaped by station functions and surrounding urban conditions. From a policy perspective, this heterogeneity highlights the need for station-specific recovery strategies that account for differences in transportation and local characteristics. More broadly, by revealing the persistence and transitions of station-level recovery states, this study provides a dynamic perspective on post-pandemic urban rail recovery beyond point-in-time comparisons of ridership change. Full article

Surface defect detection across diverse surface types, such as biological shells, photovoltaic panels, and pavement infrastructure, is essential for industrial quality control and structural health monitoring. Although Deep Learning (DL) models perform well in this domain, high computational and memory requirements limit their deployment on resource-constrained edge devices. To address this, we propose a computationally efficient descriptor-level feature extraction framework combining Local Binary Pattern (LBP), Gabor filters, and Discrete Wavelet Transform (DWT) to capture complementary textural, directional, and frequency characteristics of surface defects. An AdaBoost-driven feature selection strategy reduces the high-dimensional hybrid feature pool to the 25 most discriminative attributes, and SHapley Additive exPlanations (SHAP) analysis is applied to interpret feature contributions. The proposed framework achieved 100% accuracy on the EggCrack dataset and 79.0% accuracy on the ELPV dataset, which is more challenging. On ELPV, it outperformed the MCU-optimized MCUNet-In0 (78.16%) and remained within 4.16 percentage points of EfficientNetV2-B0 (83.16%), while achieving approximately 7577× fewer floating-point operations (FLOPs) and 89× smaller model size. Deployed on a Raspberry Pi 5 CPU, the framework achieved an end-to-end inference time of 37 ms per image, enabling real-time inference. These results indicate that the proposed hybrid descriptor and selection framework offers a favorable accuracy-efficiency trade-off for real-time surface defect detection on CPU-based edge devices. Full article

Rapid reconstruction of finite-element stress distributions can support preliminary screening when many related two-dimensional geometries must be examined in a fixed mechanical setting. This study presents a revised RSP-Net for reconstructing normalized RGB renderings of von Mises stress distributions from binary geometry images. The architecture uses U-Net++ as a dense image-to-image backbone and introduces residual Inception refinement only at the input and final full-resolution stages. The training objective combines the mean squared error with a Sobel-based gradient consistency term that emphasizes local variations in the rendered stress pattern and dominates the weighted objective at the selected setting. The evaluation uses a fixed, non-overlapping train/validation/test split, three independent training seeds, closely related U-Net baselines, two trivial baselines, validation-only loss weight selection, and metrics restricted to the structural region. Across three runs, RSP-Net obtains an RGB-RMSE of $0.09727\pm 0.00057$, an RGB-MAE of $0.03799\pm 0.00029$, and an interior RGB-SSIM of $0.85616\pm 0.00099$. Its mean improvements over U-Net++ are modest and are not statistically significant after Holm correction, whereas the gradient consistency term significantly improves RGB-MAE, interior RGB-SSIM, all three high-gradient-region errors, and the Sobel gradient error while increasing the RGB-RMSE by approximately 1.02%. The official color table further yields a normalized scalar RMSE of 0.04568 and a conditional MPa RMSE of 29.83 MPa when the true per-image scale is supplied. Because the targets are independently normalized color renderings, the method reconstructs stress distribution patterns but does not independently predict absolute stresses in MPa. Full article

This paper focuses on the binary task of parking-slot occupancy detection (vacant vs. occupied) from UAV aerial imagery. Accurate parking-slot detection from UAV imagery is challenged by small target sizes, highly regular rectangular shapes, large-scale variations, complex backgrounds, and perspective distortions. Compared with fixed surveillance cameras, UAV-based detection offers flexible deployment, wide-area coverage, and no requirement for pre-installed infrastructure, making it especially suitable for large open-air parking lots and temporary parking scenarios. To address this, this study proposes a task-specific framework for UAV-based parking-slot detection, with improvements in backbone design, attention modeling, and bounding-box regression. A lightweight MnasNet-inspired backbone is used to improve multi-scale feature extraction at low computational cost. An enhanced EMA module with adaptive grouping, FFT-based frequency enhancement, and gated fusion is introduced to better model the structured patterns of parking lot scenes. In addition, a UIoU+ loss tailored to rectangular geometry is proposed to improve localization quality. Sensitivity analysis and repeated experiments show that the method is stable and statistically reliable. All main metrics are evaluated on an independent held-out test set to ensure generalization. Extensive experiments demonstrate that each component brings consistent performance gains. The proposed model achieves 99.44 ± 0.12% mAP@0.5, 90.31 ± 0.27% mAP@0.5:0.95, 99.27 ± 0.15% precision, and 99.00 ± 0.18% recall on the self-built UAV Parking Lot dataset. Its mAP@0.5:0.95 is 26.01 percentage points higher than the YOLOv11n baseline. Consistent performance improvements are also validated on two additional public benchmarks (CARPK and PKLot), confirming the generalization of the proposed method beyond the self-built dataset. Most importantly, the method supports real-time inference on embedded UAV platforms and achieves state-of-the-art performance among lightweight detectors, making it an ideal solution for practical intelligent parking management. Ablation studies further confirm the complementary synergy between the proposed backbone, attention module, and loss function. Full implementation code, pre-trained weights, and detailed reproduction guidelines are publicly available to ensure research reproducibility. Full article

This article presents an empirical comparison of four automatic face recognition methods: Principal Component Analysis (PCA), Local Binary Patterns (LBP), the deep neural network FaceNet, and an original hybrid approach proposed by the authors, referred to as FLLF (Feature-Level Late Fusion). The experiments were conducted on a subset of the VGGFace2 database, comprising approximately 480 classes in the training set and 60 classes in the validation set. For closed-set identification, a new test set (20%) was extracted from the training set. Classification accuracy, training and inference times, and prediction confidence distributions were evaluated for each method. The results show a clear advantage of the deep learning approaches: FaceNet achieved an accuracy of approximately 98% with only five training images per person, whereas the classical methods—PCA and LBP—reached only approximately 7% and 22%, respectively. The proposed FLLF method, which fuses FaceNet embeddings with PCA-whitened LBP descriptors at the feature level and classifies them with a calibrated linear SVM, further improved accuracy to approximately 98.5% and produced the highest prediction confidence values of all tested methods. However, calibration quality was not directly assessed using standard metrics such as expected calibration error or reliability diagrams, so this observation should be interpreted as a confidence-distribution shift rather than a formal calibration improvement. The article also discusses the theoretical foundations of each algorithm, their respective advantages and limitations, and the architecture of the software system implemented for this study. Full article

Screen content images (SCIs) exhibit complex structural heterogeneity, rendering traditional statistics-based natural scene image quality assessment (NR-IQA) metrics ineffective. Although deep learning models achieve high prediction accuracy, their prohibitive computational demands preclude deployment in latency-sensitive industrial scenarios. While existing handcrafted lightweight SCI-IQA metrics reduce computational overhead, most rely on unsegmented global feature pooling or holistic edge statistics (e.g., edge histograms or Fisher vector coding), thereby diluting locally critical text-edge distortions in vast homogeneous backgrounds. To address this limitation, we propose an ultra-lightweight, deep-learning-free NR-IQA framework centered on human visual masking. Unlike existing lightweight methods, our approach explicitly employs dual-scale Canny edge operators to partition SCIs into edge-sensitive and flat background regions. Guided by this visual prior, structural degradations and micro-compression textures are extracted region-wise using Sobel gradients and uniform local binary patterns (LBPs) and aggregated with global Commission Internationale de I’Eclairage L*a*b*(CIELAB) color statistics into a compact 60-dimensional descriptor. A grid-search-optimized Support Vector Regression (SVR) maps these features to subjective quality scores. Extensive cross-validation on the SIQAD and SCID datasets demonstrates that our metric outperforms existing handcrafted lightweight SCI metrics and traditional NSS models, while achieving accuracy competitive with representative full-reference metrics. Consuming only 79.3 ms per image on a standard CPU, it offers a practical accuracy–efficiency trade-off for resource-constrained periodic quality monitoring. Full article

Background/Objectives: Artificial intelligence (AI) models for glaucoma screening using colour fundus photography have shown strong internal performance; however, their external validity and calibration reliability remain uncertain. This study developed a controlled four-dataset benchmark to evaluate six glaucoma-screening models across RIM-ONE, DRISHTI-GS, the Hillel Yaffe Glaucoma Dataset, and ORIGA. Methods: The evaluated models included a hybrid deep-handcrafted random forest (RF), transfer-learning and semi-supervised VGG16 models, and a compact convolutional neural network (CNN). Performance was assessed in within-dataset and cross-dataset settings using discrimination metrics, including accuracy, area under the receiver operating characteristic curve (AUC), balanced accuracy, and Matthews correlation coefficient (MCC), as well as calibration metrics, including Brier score and expected calibration error (ECE). Threshold stability, preprocessing ablation, and repeated-seed analyses were also performed. Results: Within-dataset evaluation showed strong discrimination, with the complete hybrid CNN + histogram of oriented gradients (HOG) + local binary patterns (LBP) + minimum redundancy maximum relevance (mRMR) + RF pipeline achieving a mean accuracy of 0.8697 and an AUC of 0.8972. However, cross-dataset performance was poor, with a mean accuracy of 0.5685 and an AUC of 0.5798. The denoising autoencoder-enhanced transfer-learning model showed improved probability calibration in transfer settings. Preprocessing effects were dataset-dependent, with raw images, region of interest (ROI) cropping, and Retinex normalisation producing different external performance patterns. Conclusions: High internal accuracy did not translate into reliable cross-domain generalisation. None of the evaluated models were suitable for zero-shot deployment without site-specific validation or recalibration. Full article

Multi-modal sentiment analysis integrates linguistic, acoustic, and visual evidence, yet the reliability of these streams varies across samples because of missing observations, masking, measurement noise, and feature corruption. This paper presents a trainable reliability-aware evidential fusion framework that estimates not only sentiment predictions but also modality-specific evidence, predictive uncertainty, observable input quality, cross-modal disagreement, and normalized sample-dependent fusion weights. Each available modality is independently encoded and processed by an evidential classification head and a quality estimation head. Availability masks enforce exact exclusion of missing streams, while estimated quality, Dirichlet uncertainty, and Jensen–Shannon disagreement jointly regulate the contribution of each observed stream. The model is optimized end-to-end using fused classification, evidential regularization, clean–corrupted consistency, reliability-calibrated cross-modal alignment, and quality regression objectives. Experiments are conducted on both CMU-MOSI and CMU-MOSEI using their official speaker-independent splits. Binary classification follows the standard non-zero protocol, in which samples with sentiment score zero are excluded from Acc-2 and binary F1 evaluation; all labeled samples are retained for seven-class accuracy, mean absolute error, and correlation. The evaluation covers complete-input, every single- and double-modality missing pattern, graded and unseen corruption, combined missing-plus-corrupted conditions, calibration, selective prediction, statistical testing, and computational efficiency. All comparative values in the main tables are identified as local controlled adaptations under the common pipeline, while selected published reference values are reported separately to prevent provenance mixing. Across both datasets, the empirical results show that the proposed method preserves competitive complete-input performance while providing larger and more consistent gains as modality availability or integrity deteriorates. Full article

Automated cotton-percentage identification underpins sustainable textile recycling, but established near-infrared and ATR-FTIR spectroscopy systems cost USD 10,000–25,000 per unit and remain inaccessible to small recyclers. We address this on the CottonFabricImageBD dataset (1300 RGB originals, 13 ordinal cotton classes from 30% to 99%) and report three contributions. First, the Learnable Residual LBP stem, which retains the pretrained ResNet50 first convolution intact and adds a fully differentiable Local Binary Pattern branch as an additive contribution gated by a single learnable scalar $\alpha$ initialized to zero, ensuring the model is numerically equivalent to the baseline at initialization (verified to a maximum absolute logit difference below ${10}^{-4}$). Second, a controlled six-variant comparison (vanilla baseline, CLBP, LBP-Conv, LBP-Residual, LBP+SVM, LBP+ANN) under identical stratified five-fold cross-validation on the 1300 dataset originals. Third, the isolation of pretraining preservation as the dominant architectural variable: the 7.08 pp top-1 gap between LBP-Conv (43.77%) and LBP-Residual (50.85%), both embedding the identical learnable LBP module, is statistically significant ($p=0.004$, uncorrected paired t-test, $df=4$) and consistent across all five folds. This gap mainly reconfirms, in the LBP setting, the established cost of discarding pretrained early-layer filters; by contrast, the improvement of LBP-Residual over the vanilla baseline (1.31 pp top-1) is consistent in direction but not statistically significant at the five-fold level ($p=0.229$), so LBP-Residual, CLBP (50.23% top-1), and the baseline (49.54% top-1) are statistically tied on aggregate accuracy and the ranking among them is exploratory. Classical LBP+SVM and LBP+ANN baselines reach 31.85% and 34.46% top-1, confirming a genuine but limited cotton-density signal in hand-crafted descriptors. Compared to the concurrent triplet-architecture approach of Wiedemann et al. (2025), which achieves 48.15% top-1 accuracy on the same dataset under identical five-fold cross-validation, LBP-Residual attains 50.85% top-1 using a single lightweight backbone rather than an ensemble of three. These results support the design principle: augment, do not replace. Full article

Color texture classification is an important task in computer vision, with applications in medical imaging, industrial inspection, remote sensing, and material analysis. This paper presents a hybrid framework that integrates Principal Component Analysis (PCA), Local Binary Patterns (LBPs), Wavelet Scattering Transform, and the XGBoost classifier for color texture classification. The proposed pipeline first performs image pre-processing, including resizing and denoising, followed by channel-wise feature extraction using LBP and Wavelet Scattering Transform on the Red, Green, and Blue channels independently. Then, the obtained feature vectors were concatenated, and PCA was applied on the fused feature space for dimensionality reduction and redundancy elimination before proceeding to XGBoost classification. This method not only leverages complementary information of Chroma and texture information but also achieves reduced dimensionality and computational burden. The finally optimized features were input into the XGBoost classifier for color texture classification, which is good at fitting non-linear dependency and includes a regularization to generalize better. Our proposed framework was tested on three benchmark color texture datasets: KTH-TIPS, Outex_10, and VisTex. Experimental results have demonstrated that on these three datasets, the average performance reaches 98.0% accuracy, 0.981 precision, 0.981 recall, and 0.979 F1-score, respectively. It demonstrates that the two selected complementary feature extraction methods provide a compact yet effective representation for color texture classification on these datasets. It is expected that the proposed framework serves as an efficient combination of established methods and as a good competitive baseline for color texture analysis. Future works will consider applying it to larger color texture datasets for general verification, enhancing its computational efficiency and automating the parameter selection process. Full article

Down syndrome is associated with characteristic craniofacial features that have motivated the development of computer-vision-based facial-image-based screening systems. Recent studies have increasingly relied on deep computer vision and deep-learning approaches, but many provide limited interpretability, while earlier landmark-based methods offered transparent geometric and texture-based measurements. This creates a gap between interpretable handcrafted features and high-performing deep representations. To address this gap, this study proposes a hybrid interpretable–deep framework that combines landmark-derived geometry features, landmark-guided local binary pattern (LBP) texture descriptors, and frozen EfficientNetB0 convolutional neural network (CNN) deep features. The primary contribution of this study is the systematic integration and comprehensive evaluation of complementary interpretable and deep-feature representations within a unified facial-image screening framework. Feature fusion is followed by random forest feature ranking and SVM-RBF classification. Experiments were conducted on 2979 successfully processed facial images from an original dataset of 2999 images. Geometry-only, texture-enhanced, deep-feature, and hybrid fusion models were evaluated using repeated stratified train–test splits. The final RF Top-800 fusion model achieved strong facial-image classification screening performance, with F1 = 0.9045 +/− 0.0134 and AUC = 0.9675 +/− 0.0073 across repeated stratified train–test splits. Ablation analysis showed that removing geometry features, removing LBP features, or using only EfficientNetB0 reduced performance, supporting complementary contributions from interpretable geometry and texture feature components and deep-feature representations. Statistical comparisons and duplicate-sensitivity analyses further supported the robustness of the results. The findings demonstrate that landmark-derived geometry and texture descriptors remain valuable when integrated with modern deep representations, providing feature-level interpretability while improving screening performance through a hybrid framework that combines interpretable handcrafted features with high-performing deep representations. Full article

Parkinson’s disease (PD) diagnosis and severity assessment increasingly rely on gait analysis. To capture complex gait dynamics from limited sensor data, we propose a novel multi-branch framework integrating a one-dimensional convolutional neural network (1D-CNN), a bidirectional gated recurrent unit (Bi-GRU), and Kolmogorov–Arnold networks (KANs) (CNN-GRU-KAN) utilizing 18-channel vertical ground reaction force (VGRF) signals. Each module serves a distinct clinical purpose: the 1D-CNN branch with squeeze-and-excitation (SE) attention extracts localized spatial plantar pressure patterns, while the Bi-GRU branch with temporal attention captures long-range rhythm abnormalities. Crucially, the KAN serves as the classification head. By utilizing learnable B-spline functions instead of traditional fixed activations, KAN adaptively models the highly non-linear boundaries between healthy controls and varying PD severities, effectively mitigating overfitting. Under rigorous subject-independent cross-validation, our model achieves 98.43% accuracy for binary PD detection and 93.46% for five-class UPDRS severity grading. These results highlight the framework’s strong potential for low-cost, unobtrusive clinical tracking and home monitoring. Full article

From a Historic Urban Landscape (HUL) perspective, varying conservation intensities and statutory boundaries may create “zonal differentiation” within historic districts. However, conventional homogenized renewal strategies frequently overlook this heterogeneity, affecting physical townscapes and human perception. This study analyzes 1714 panoramic street-view images from the Core Protection Zone and Construction Control Zone of Jimei School Village. Eleven objective visual features and six model-predicted perception dimensions were examined, with locally experienced participants providing contextual validation. Integrating K-Means clustering, Random Forest modeling, and SHapley Additive exPlanations (SHAP) feature attribution, the study investigates nonlinear, model-based contribution patterns across the two zones. Results reveal significant but non-binary differences in objective features and predicted perceptions. Street-view typologies show zonal tendencies while also indicating within-zone diversity and cross-zone overlap. Feature attribution shows that color composition, greenery, building interfaces, and vehicle presence are more prominent in the Core Protection Zone, whereas greenery, spatial openness, and road-space organization play stronger roles in the Construction Control Zone. This study establishes an interpretable street-view-based framework for historic-district assessment, providing empirical support for differentiated and human-oriented zonal renewal. Full article