搜索结果：Texture

Morphological characteristics of coarse aggregates, namely shape, angularity and surface texture, are closely related to rock mineral compositions and crushing mechanisms. Aggregates with different mineral compositions should be crushed using a suitable crusher. To explore the influence of mineral and crusher types on aggregate morphology, this study investigated the effect of the mineral compositions and crushing operations on the morphologies of coarse aggregates. Six types of major rock-forming minerals (i.e., quartz, amphibole, potassium feldspar, sodium feldspar, calcite and pyroxene) were selected and two typical crushers (i.e., jaw crusher and impact crusher) were used. The morphology parameters (i.e., angularity, texture, sphericity and F&E) of coarse aggregates were measured using the Aggregate Image Measurement SystemII (AIMSII). Further, the morphologies of different aggregates produced by two crushers were compared. The results showed that the angularity values of some aggregates crushed by a jaw crusher (average 3451) were bigger than those by an impact crusher (average 3067) and the angularity of the harder mineral was less affected by the crusher. For surface texture, there was no significant difference between these two crushers, with average texture index of 333.83 for the impact crusher and 332.67 for the jaw crusher, indicating that the surface texture of aggregates was mainly affected by their compositions and barely influenced by the crushing operations. The shape results indicated that the impact crusher produced more cubical aggregate particles compared to the jaw crusher, with average sphericity of 0.666 versus 0.607, whereas the jaw crusher produced aggregates with more elongated or flat particles, with an average F and E index of 3.331 versus 2.726. This study fills the research gap that few previous investigations focused on-the crushing morphology of single mineral aggregates-and the findings help us to understand the effect of the mineral compositions and crushing operations on the morphologies of coarse aggregates, which in turn guides the selection of suitable crushers for different minerals.

Facial emotion recognition plays an important role in affective computing and human-computer interaction. Although convolutional neural network (CNN)-based methods have demonstrated remarkable performance, deep features alone may not sufficiently capture subtle geometric deformations and local texture variations, particularly under limited training data and challenging real-world conditions. To address this limitation, this study proposes a hybrid framework that integrates CNN-based deep features with handcrafted geometric and texture features. Specifically, 17 landmark-based angular features extracted from the eyebrows, eyes, nose, and mouth are combined with histogram of oriented gradients (HOG) features extracted from the nose and mouth regions through feature-level concatenation. The proposed method was extensively evaluated on three controlled datasets (JAFFE, CK+, and KDEF) and two large-scale in-the-wild datasets (RAF-DB and AffectNet). Five-fold cross-validation, leave-one-subject-out cross-validation, statistical significance analysis using paired t-tests, computational efficiency analysis, and comparisons with conventional handcrafted methods, standard CNN models, transfer learning-based methods, and recent hybrid feature-fusion methods were performed to comprehensively validate the proposed approach. Experimental results demonstrated consistent improvements across different datasets, evaluation protocols, and CNN backbone networks while maintaining a favorable balance between recognition performance and computational efficiency. These findings demonstrate that handcrafted geometric and local texture features effectively complement CNN-based deep representations, providing a robust and generalizable framework for facial emotion recognition across both controlled and large-scale in-the-wild datasets.

To address the challenges of extremely small object scales, weak texture information, and severe background interference in aerial remote sensing images, we propose MambaIR-YOLO, a feature-guided lightweight state-space framework for aerial small-object detection. Based on the YOLOv5 architecture, this method introduces systematic improvements in three key areas-fine-grained feature modeling, long-range dependency learning, and high-resolution spatial information preservation-while ensuring real-time performance. Specifically, a feature-level MambaIR_SR (feature-level Mamba-based super-resolution guidance) training auxiliary branch is designed to generate high-resolution detail-guided information by fusing shallow-level detail features with deep-level semantic features, improving the representation of small-object edges and textures during the training phase. In the main network, we introduce the Object Detail State-Space Block (ODSSBlock), driven by Lightweight Mamba. Through a channel-compressed state-space modeling mechanism, the ODSSBlock unifies local detail preservation and long-range context modeling with low computational overhead. Concurrently, a Feature Modulation Block (FMB) is constructed at the shallow feature level to enhance the representation of high-frequency structural information, thereby mitigating the irreversible detail degradation caused by multiple downsampling steps. In the feature fusion and detection stages, we introduce a Coordinate and Channel Attention (C3CA) attention enhancement module and construct a lightweight decoupled detection head based on a Lightweight Decoupled Head (LADH). This performs single-scale dense prediction on high-resolution features, improving small-object localization and reducing information loss. Notably, the MambaIR_SR branch is exclusively utilized for optimization during the training phase and is entirely discarded during inference, introducing no additional computational overhead. Experimental results on the VEDAI dataset demonstrate that the proposed method achieves an average mAP50 of 84.19 ± 0.03% over three independent runs, with only 4.46 M parameters and 19.97 GFLOPs, outperforming the baseline methods while maintaining low computational cost.

Aluminum alloys based on the Al-Ce-Mg system, with microstructure and properties weakly sensitive to elevated temperatures, are developed for applications such as thermal engine blocks and supersonic aircraft fuselage components. Cast Al-10Ce-4Mg (wt%) was produced by slab casting and further processed by rolling. Rolling with an area reduction of 54% results in a 97% increase in strength (tensile strength of 269 MPa at room temperature) and a more than a factor of two increase in strain at failure relative to the as-cast (AC) state. We show that Al-10Ce-4Mg is highly stable upon exposure to elevated temperature: 83% strength retention after exposure to 300 °C for 100 h, without reduction in the strain at failure. The yield stress (YS) follows the same trend, increasing by ~40% upon rolling and retaining 92% of its value after exposure to 300 °C for 100 h. The ratio of the strength at 300 °C to the strength at room temperature is 0.45, which is larger than the respective ratio of most commercial Al alloys. Rolling increases the dislocation density and introduces texture, while not modifying the grain size significantly and fragmenting only the largest intermetallics. Exposure to elevated temperatures up to 100 h has no measurable effect on grain size and intermetallic distribution, while slightly changing texture. This microstructural evolution is used to rationalize the observed changes in mechanical properties.

Potato leaf diseases directly reduce yield and quality, and accurate field detection is important for precision plant protection. However, potato disease lesions are often weak in deep semantic representation, easily disturbed by complex field backgrounds, and variable in multi-scale lesion texture. To address these challenges, this study proposes an improved YOLO-based potato leaf disease detection model. The proposed model enhances the detector through three task-oriented modules. Deep Symptom Enhancement is used to strengthen deep disease feature extraction. Lesion Selection Attention based on large separable kernel attention improves the spatial selection of lesion regions. Multi-Scale Refinement Adapter uses a Mona-based C2PSA structure with two stacked Mona adapters to refine multi-scale texture and lesion-boundary information. Experiments were conducted on a potato leaf disease image dataset using mAP50, average recall (AR), parameters, GFLOPs, and FPS as evaluation metrics. The baseline YOLO26s achieved 81.31% mAP50 and 77.85% AR. The proposed SLR-YOLO model achieved 88.92% mAP50 and 83.51% AR, improving mAP50 and AR by 7.61 and 5.66 percentage points, respectively, while maintaining 118.6 FPS. The results show that the proposed framework improves detection accuracy for potato leaf disease images while retaining practical real-time performance.

Traditional sweets are associated with dental caries, high blood glucose, and obesity. This study evaluated the impact of incorporating Moringa oleifera leaf powder (MOLP) into soft candies (gummies and marshmallows) formulated with a fixed base matrix of cranberry juice and low-glycemic, noncariogenic sweeteners (isomaltulose, oligofructose, and tagatose). MOLP was incorporated at 0%, 2.5%, 5.0%, and 7.5% while keeping the proportions of cranberry juice and sweeteners constant, enabling isolation of the specific contribution of MOLP. The increase in MOLP levels produced statistically significant increases in the antioxidant capacity and phenolic content of the candies, which confirms their functional potential. MOLP addition also produced statistically significant changes in color, shifting products toward darker, greenish‑yellow tones due to moringa pigments. Texture was matrix‑dependent: Gummies became progressively firmer and chewier with increasing MOLP, whereas marshmallows presented moderate structural changes attributable to their aerated nature. Sensory evaluation revealed the highest consumer acceptance for the control and 2.5% formulations, with a reduced liking at 7.5% due to bitterness, color changes, and excessive firmness. Purchase intent remained positive when health benefits were communicated. Overall, moderate MOLP inclusion (2.5%-5%) provides a practical balance between functional enhancement and sensory appeal for cranberry‑based soft candies formulated with low‑glycemic sweeteners. PRACTICAL APPLICATIONS: This study shows that gummies and marshmallows can be made with low-glycemic sweeteners and plant ingredients such as moringa to improve their nutritional quality. The results can help food producers develop sweets with added antioxidant value while keeping acceptable taste and texture. These products may be of interest to consumers seeking confectionery options with a more health-oriented composition.

Lignin deposition in fruit trees represents a fundamental physiological trade-off: essential for structural integrity and stress adaptation, yet excessive or mistimed activation compromises fruit texture, palatability, and market value. This review synthesizes advances in lignin biosynthesis and its multilayered regulation in commercial fruit species. We describe how abiotic (drought, salinity, temperature extremes) and biotic (pathogens, pests) stresses trigger lignification through transcriptional, post-transcriptional, hormonal, and epigenetic mechanisms, centered on the conserved NAC-MYB cascade. This core module integrates WRKY/ERF transcription factors (TFs), microRNA networks, and hormone signaling. Transcriptional programs are further refined by microRNAs, alternative splicing, DNA methylation, histone acetylation, and phytohormone crosstalk (abscisic acid, ABA; jasmonic acid, JA; salicylic acid, SA; and brassinosteroids, BRs). We emphasize molecular crosstalk integrating abiotic and biotic stress signaling via shared TFs, reactive oxygen species (ROS), and epigenetic memory. We critically examine lignification's dual nature during development and postharvest storage, contributing to desirable traits (stone formation and skin toughness) but also driving defects (stone cell gritty texture and chilling-induced wooliness). Finally, we propose a strategic framework leveraging molecular breeding, targeted gene editing, and precision horticulture to fine-tune lignification, enabling climate-resilient cultivars without compromising fruit quality.

Stone cells, characterized by heavily lignified secondary cell walls, severely impair the texture and eating quality of pear fruit, particularly in Asian cultivars. Although brassinosteroids (BRs) are key hormones regulating fruit development, their role in stone cell formation has remained unclear. Here, we found that exogenous BR treatment promotes stone cell lignification in pear fruit. Using co-expression network analysis across 24 diverse tissues and correlation analysis of 206 pear accessions, we identified PbrBEH3, a member of the BRASSINAZOLE-RESISTANT (BZR)/BRI1-EMS-SUPPRESSOR (BES) transcription factor family, the expression was strongly and positively correlated with stone cell content. Functional analyses demonstrated that transient overexpression (OE) of PbrBEH3 in pear fruit and stable OE in pear callus significantly increased lignin accumulation, whereas CRISPR/Cas9-mediated knockout reduced lignin levels. Consistently, PbrBEH3 OE in Arabidopsis thaliana promoted lignin deposition in stems. Mechanistically, BR treatment enhanced nuclear accumulation of PbrBEH3, which directly activated Pbr4CL1 and PbrMYB169 transcription by binding to E-box and BRRE cis-elements in their promoters, respectively. Collectively, our findings establish PbrBEH3 as a crucial transcriptional activator that integrates BR signaling to regulate lignin biosynthesis and stone cell formation in pear fruit. This work provides novel insights into the hormonal regulation of fruit quality and identifies a potential molecular target for improving fruit texture in pear breeding.

Background: The present study aimed to develop and characterize three honey- and essential oil-based structured systems intended for topical skin application. Materials and Methods: The semisolid systems were prepared as oil-in-water structured emulsions containing four types of honey (Manuka, Tualang, chestnut, and manna), three essential oils (palmarosa, cistus, and lavender), and five vegetable oils (pomegranate seed, aloe, centella, hemp seed, and calendula). Each formulation consisted of two honey types, one essential oil, and two vegetable oils, with final concentrations of 5% honey and 0.1-0.2% essential oil. The formulations were investigated through physicochemical, rheological, and in vivo skin evaluations. Results: Rheological analysis demonstrated non-Newtonian pseudoplastic behavior with shear-thinning and thixotropic characteristics, indicating that the systems are structured, semisolid, and suitable for topical application. Differences in spreadability and consistency suggested variations in the internal organization of the emulsion matrices. In vivo skin assessments were performed over four weeks using non-invasive instrumental methods. The obtained results demonstrated improvements in skin hydration, elasticity, firmness, and skin barrier function, together with reductions in transepidermal water loss and erythema. The evaluation of skin textural parameters revealed improvements in skin uniformity and a reduction in the appearance of wrinkles. Among the tested structured systems, F3 formulation exhibited the most pronounced moisturizing effect, while F1 formulation showed notable improvements in parameters associated with skin texture and wrinkle-related features. Conclusions: Overall, the results indicate that honey-based topical systems enriched with essential and vegetable oils represent promising multifunctional semisolid formulations for topical skin-conditioning and barrier-supportive applications. Their favorable rheological behavior, combined with beneficial effects on skin hydration, barrier function, and skin surface properties, supports their potential use in skin-conditioning and anti-aging-related formulations.

Industrial metal surface defect detection is essential for visual sensing-based quality inspection, where lightweight models must balance reliable defect perception and efficient deployment under limited computational resources. However, weak textures, blurred boundaries, and small defect scales often reduce the reliability of lightweight detectors, while heavy global modeling modules increase computational cost. This article proposes a lightweight You Only Look Once (YOLO)-based framework, termed Mediation-based Conflict Mitigation YOLO (MCM-YOLO), for industrial metal surface defect detection. Built upon YOLOv11, MCM-YOLO introduces a C3-Res2Lite (C3-R2L) block to enhance fine-grained local representation and an Aggregated Bidirectional Feature Pyramid Network (Agg-BiFPN) to strengthen multiscale feature aggregation. We further observe that directly coupling the enhanced backbone and strengthened neck may degrade performance, which is referred to as Feature Integration Conflict (FIC). To alleviate the potential feature incompatibility associated with FIC, this article introduces a synergistic attention block, termed MCM-SAB, at the critical interface to perform coordinated channel recalibration and spatial refinement. Experiments on NEU-DET and GC-10 show that MCM-YOLO achieves 80.4% and 65.1% mean average precision (mAP) at an intersection-over-union (IoU) threshold of 0.5, respectively, with 6.3 giga floating-point operations (GFLOPs) and 2.70 million parameters. These results indicate that MCM-YOLO provides a competitive accuracy-efficiency tradeoff for industrial visual inspection.

Per- and polyfluoroalkyl substances (PFAS) are persistent groundwater contaminants that pose long-term risks to water sources and public health. Predicting PFAS occurrence remains challenging due to high-dimensional environmental data and limited model interpretability. In this study, we benchmarked explainable machine learning models to predict PFAS occurrence in groundwater and to identify the features strongly associated with estimated PFAS occurrence. A comprehensive dataset of 12,406 groundwater characterization records collected between 2001 and 2019 was compiled with 172 explanatory features to describe PFAS source proximity, land use, hydrogeology, soil properties, meteorology, and sampling sites characteristics. Four tree-based ensemble classifiers, including Random Forest, XGBoost, LightGBM, and CatBoost, were evaluated under multiple classification schemes. Binary classification, which split total PFAS concentrations into 'low' and 'high' categories based on the median value, achieved the most robust and generalizable performance, with testing accuracy above 94% and well-established precision-recall curves. Increasing the number of classification bins degraded performance, particularly for intermediate bins, highlighting intrinsic separability limits in PFAS occurrence data rather than model deficiencies. Model interpretability was addressed using SHapley Additive exPlanations (SHAP), which revealed that sampling year and proximity to major PFAS sources were the dominant predictors across all models. Additional contributions were attributed to proximity to other PFAS sources, soil texture, hydrologic features, land use, and precipitation patterns. SHAP interaction analyses further revealed model-learned temporal variation in the attribution of source-proximity predictors. These patterns are interpreted as hypothesis generating model associations that may be related to regulatory changes, evolving monitoring strategies, analytical-era differences, and possible secondary-source influences, rather than as confirmation of specific environmental processes. Collectively, this study achieved interpretable prediction for PFAS occurrence in groundwater, and offered a transparent, data-driven framework to inform PFAS risk assessment.

Infrared thermal imaging offers objective physiological insights for Traditional Chinese Medicine (TCM), yet automated acupoint localization struggles with low texture and extreme pose variations. To address this, we constructed a multi-pose thermal facial dataset using digitized bone-proportional measurements and TCM anatomical rules. We propose T2FAL, a two-stage cascaded regression framework explicitly decoupling macroscopic face detection from fine-grained acupoint localization. Stage 1 utilizes an improved YOLOv12m-based Thermal-Aware Face Detector-integrating ICAN_C2f, MixNeck, and TAF-IoU-to mitigate domain shifts and thermal noise. Stage 2 deploys pose-specific regressors incorporating Gated Feature-Conditioned Cascade Refinement (FCCR) and a Selective GeoDeriv module, mathematically translating anatomical rules into geometric constraints. Under the stated experimental protocol, Stage 1 processed images at 87.3 frames per second on an NVIDIA RTX 4090. For Stage 2, the final frontal- and profile-view configurations achieved mAP@50-95 values of 73.19% and 86.92%, with mean pixel errors of 1.986 and 3.109 pixels, respectively. These results support the feasibility of automated reference-coordinate localization on the datasets used. However, given the partial reliance of the annotations on image registration and geometric rules, the use of the cross-domain set for model selection, and the lack of clinical or diagnostic evaluation, further validation using independently generated expert annotations, a strictly held-out external test set, and clinically labeled data is required before clinical or diagnostic use.

Real-time recognition of the transient interface opening between a nickel starting sheet and its titanium starter plate is required to stop high-frequency tamping at the correct instant. This study proposes Yolov11_tapcheck, a deployment-oriented detector based on YOLO11n, and implements a field real-time detection system for nickel starting sheet pre-stripping. To address small target scale, weak texture, reflective background interference, and motion blur, the model introduces efficient multi-scale attention (EMA) and reconstructs the neck as a BiFPN-inspired bidirectional feature-fusion path. A tapcheck dataset was established from production-line images, and the method was evaluated with image-level accuracy, latency, and field control metrics. Compared with the original YOLO11n, Yolov11_tapcheck improved mAP50 from 95.5% to 98.1% and F1 from 93.7% to 97.5%. The average model-pipeline latency was 7.1 ms; when MQTT delivery, OPCSERVER writing, PLC feedback timestamp comparison, and 1 ms polling were included, the average frame-to-PLC acknowledgement latency was 10.24 ms, with a maximum of 11.58 ms, remaining below the 13.333 ms frame interval in 1000 logged samples. Two field stages further showed that the opening success rate for cycles 3-8 increased from about 60-65% to more than 90%, while opening detection success increased from 90% to 100%. These results indicate that the proposed method can support real-time opening recognition and closed-loop control in nickel starting sheet stripping.

Vehicle-ejected debris detection is a practical but insufficiently studied problem in intelligent traffic enforcement. Unlike static road litter, objects thrown from moving vehicles are usually small, irregular, transient, and easily confused with road textures, shadows, lane markings, and light reflections. In current traffic management, such violations still rely heavily on manual video review or offline inspection, while task-specific datasets and edge-deployable detection solutions remain limited. To address this gap, this study constructs a vehicle-ejected debris dataset containing 4328 annotated image samples collected from real road scenarios. The dataset covers urban and suburban roads, daytime and nighttime illumination, near-range and distant small-object cases, and hard negative samples. To meet the coupled requirements of vehicle-mounted small-object detection and edge-side INT8 deployment, this study develops a hardware-aware lightweight detection framework based on YOLOv8m. The original CSPDarknet backbone is replaced with the convolutional variant of MobileNetV4 to reduce feature-extraction cost, while a scale-specific Channel Alignment Module is inserted between the heterogeneous MobileNetV4 backbone and the YOLOv8m PANet neck to preserve multi-scale feature compatibility. The alignment module uses only BPU-friendly convolution, batch normalization, and activation operations, thereby avoiding deployment-unfriendly operators while maintaining compatibility with INT8 quantization and edge acceleration. The trained FP32 model is quantized to INT8 and deployed on the RDK X5 BPU using the Horizon OpenExplorer toolkit. Experimental results and repeated-seed validation show that the proposed model achieves a consistent accuracy-efficiency advantage on the constructed dataset. In a representative run, the proposed model obtains 93.1% mAP50, while reducing the number of parameters from 25.9 M to 13.1 M and GFLOPs from 78.9 to 39.6 compared with the YOLOv8m baseline. After INT8 deployment, the model reaches 112.6 FPS on the RDK X5 platform with only a minor accuracy decrease. These results indicate that the proposed framework can serve as a practical edge-deployable perception module for real-time vehicle-ejected debris monitoring under vehicle-mounted traffic-enforcement scenarios. It should be noted that this work focuses on single-frame debris detection, while event-level ejection verification, temporal consistency analysis, offending-vehicle attribution, and enforcement decision-making remain beyond the scope of this study.

Accurate real-time detection of small traffic objects remains a critical challenge for onboard vision-based traffic perception, particularly under conditions of weak texture, scale variation, occlusion, and limited computational resources. To address these challenges, this paper proposes DER-YOLO, a lightweight small-object-oriented detector built upon YOLO11n, specifically designed for complex traffic scenes. DER-YOLO introduces stage-wise feature calibration across the backbone, neck, and pre-head stages to enhance small-object representation. First, a Decoupled Global Context C3k2 (DGC-C3k2) module strengthens contextual representation for weak and low-saliency traffic objects after local feature extraction. Second, an ECA-guided Cross-scale Adaptive Fusion (ECAF) module adaptively balances high-level semantic cues and shallow high-resolution details to improve multi-scale feature interaction. Third, a Refined Large Selective Kernel (RLSK) module refines high-resolution spatial responses before the P3 detection head, enhancing small-object localization. Extensive experiments on KITTI and BDD100K demonstrate that DER-YOLO improves detection accuracy while maintaining real-time inference. On KITTI, it achieves 86.95% mAP@0.5 and 60.89% mAP@0.5:0.95, with small-object AP@0.5:0.95 increasing from 27.0% to 29.3%. On BDD100K, it achieves 55.05% mAP@0.5 and 29.21% mAP@0.5:0.95, with small-object AP@0.5:0.95 increasing from 12.6% to 15.2%. With 2.744 M parameters, 7.401 GFLOPs, and over 100 FPS, DER-YOLO provides an effective and lightweight solution for real-time small-object detection in onboard traffic perception scenarios.

To evaluate the diagnostic value of pituitary magnetic resonance imaging (MRI) radiomics features combined with clinical parameters in distinguishing idiopathic central precocious puberty (ICPP) from isolated premature thelarche (PT), and to establish a practical and effective predictive model. A retrospective analysis was conducted on 316 girls presenting with breast development, including 231 with ICPP and 85 with PT, from Hebei Children's Hospital (internal dataset), and 47 girls (26 with ICPP and 21 with PT) from Shijiazhuang Children's Hospital (external validation dataset). Clinical models, pituitary MRI radiomics models, and combined clinical-radiomics models were developed. Model performance was evaluated using ROC curves, calibration analysis, and decision curve analysis. Significant differences were observed between the ICPP and isolated PT groups in body mass index (BMI), bone age index (BAI), folliclestimulating hormone (FSH), luteinizing hormone (LH), and estradiol (E2) (all P < 0.01). The pituitary gland was more frequently protruding in the ICPP group than in the PT group (80.5% vs. 49.5%, P < 0.05). In radiomics models, texture features contributed most significantly, with sagittal anterior pituitary features showing the greatest discriminative value. The clinical-radiomics fusion model outperformed individual clinical or radiomics models in predictive performance. Key clinical and radiomics factors for differentiating idiopathic central precocious puberty (ICPP) from premature thelarche (PT) were identified: baseline BMI, BAI, FSH, LH, E2, and PRL are critical clinical indicators (with baseline LH exhibiting superior specificity as an independent diagnostic marker), and multimodal pituitary MRI radiomics analysis (predominantly the RAD-Total model) highlights the anterior pituitary as a dominant predictive feature, with ICPP-related pituitary enlargement attributed to premature hypothalamic-pituitary-gonadal axis (HPGA) activation. Notably, the clinical-radiomics fusion model (Clinical+RAD-Total) outperforms standalone clinical or radiomics models in diagnostic accuracy, achieving an AUC of 0.944 and considerable clinical net benefit, despite limitations including low sensitivity due to sample imbalance and slightly diminished performance in external validation cohorts. Clinical-radiomics models demonstrated excellent predictive performance in both internal and external datasets. Among these, the Clinical+RADTotal model provided the highest diagnostic accuracy and clinical utility.