Search Results (242)

This present study investigated the microbial dynamics, physicochemical and functional properties, and sensory characteristics of kefir produced by two different approaches: traditional kefir obtained directly from grains and kefir manufactured through a double-fermentation process in cow milk. For the first fermentation, kefir grains were inoculated in milk at different levels (1%, 3%, and 5% w/v) and incubated at 30 °C for 24 h. The lowest inoculation level promoted the greatest increase in grain biomass, whereas higher inoculation levels produced more pronounced pH decreases. All products maintained stable pH values during refrigerated storage at 4 °C for 15 days. Products derived from initial fermentations with 1% and 3% inoculum were subsequently used in a second fermentation step at two inoculation levels (1% and 10% v/v) to produce double-fermentation kefir products. These products exhibited higher counts of lactic acid bacteria and reduced yeast populations compared with traditional grain kefir. After 15 days of storage, all kefir samples maintained more than 10⁸ CFU/mL of lactic acid bacteria, more than 10⁷ CFU/mL of acetic acid bacteria, and around 10⁵ CFU/mL of yeasts. Protein content was comparable among all kefir products and unfermented milk. The product obtained with 1% grains followed by 10% v/v inoculation showed enhanced biofilm formation that increased during storage and displayed the strongest antimicrobial activity, and was therefore selected for sensory evaluation, where it achieved favorable acceptance by regular kefir consumers. Full article

Crack segmentation plays a vital role in ensuring structural safety, yet practical deployment on resource-limited platforms demands models that balance accuracy with efficiency. While high-accuracy models often rely on computationally heavy designs to expand their receptive fields, recent lightweight approaches typically delay this expansion to the deepest, low-resolution layers to maintain efficiency. This design choice leaves long-range context underutilized, where fine-grained evidence is most intact. In this paper, we propose an evidence-preserving receptive-field expansion network, which integrates a multi-scale dilated block to efficiently capture long-range context from the earliest stages and an input-guided gate that leverages grayscale conversion, average pooling, and gradient extraction to highlight crack evidence directly from raw inputs. Experiments on six benchmark datasets demonstrate that the proposed network achieves consistently higher accuracy under lightweight constraints. Each of the three proposed variants—Base, Small, and Tiny—outperforms its corresponding baselines with larger parameter counts, surpassing a total of 13 models. For example, the Base variant reduces parameters by 66% compared to the second-best CrackFormer II and floating-point operations by 53% on the Ceramic dataset, while still delivering superior accuracy. Pareto analyses further confirm that the proposed model establishes a superior accuracy–efficiency trade-off across parameters and floating-point operations. Full article

We evaluated for the morphology and growth behavior of IMC grain according to number of reflows of solder cap pure Sn microbumps. In the structure of Ni barrier/Cu layer between Cu pillar and pure Sn, solder cap pure Sn on the top layer was analyzed for the behavior change of IMC grain according to the number of reflows. The height and diameter of the bumps on the wafer were designed to be 40 μm and 30 μm, respectively. The vertical structure of the microbump consisted of Ti/Cu (1000 Å/2000 Å), Cu pillar (20 µm), Ni barrier (3 µm), and Cu (1 µm). The overall height of the bump is about 40 μm. Additionally, the height of the solder cap pure Sn as the last layer is 20 μm. The diameter of the bump is 30 μm. It was formed using plating. After plating to solder cap Sn, it was finally formed for the microbump using reflow. Samples were prepared according to the number of reflows (1, 3, 5, 7, and 9). To observe the grain morphology of the IMC, the pure Sn on the upper layer (solder cap) was removed using SupraBond RO-22 etchant. In the removed state, the morphology of the IMC grain was evaluated to the inside surface of bump using SEM and a 3D scope. The average number of IMC grains decreased linearly during reflow cycles 1 to 5 and then gradually decreased during reflow cycles 7 to 10. The average surface area of IMC grains was 18.243 μm when reflow was performed once. The average surface area of IMC grains increased proportionally for reflow cycles 1 to 10. Based on the experimental results, when the count of reflow was performed more than 10 times, it was confirmed that the solder cap pure Sn was reduced by more than 50% due to the increase in the area of IMC grain. Full article

Advances in sensor technology have significantly improved the efficiency and precision of agricultural spraying. Unmanned aerial vehicles (UAVs) are widely utilized for applying plant protection products (PPPs) and fertilizers, offering enhanced spatial control and operational flexibility. This study evaluated the performance of an autonomous UAV-based precision spraying system that applies variable rates based on zone levels defined in a prescription map. The system integrates real-time kinematic global navigation satellite system positioning with a proximity-triggered spray algorithm. Field experiments on a rice field were conducted to assess spray accuracy and fertilization efficacy with liquid fertilizer. Spray deposition patterns on water-sensitive paper showed that the graded strategy distinguished among zone levels, with the highest deposition in high-spray zones, moderate in medium zones, and minimal in no-spray zones. However, entry and exit deviations—used to measure system response delays—averaged 0.878 m and 0.955 m, respectively, indicating slight lags in spray activation and deactivation. Fertilization results showed that higher application levels significantly increased the grain-filling rate and thousand-grain weight (both p < 0.001), but had no significant effect on panicle number or grain count per panicle (p > 0.05). This suggests that increased fertilization primarily enhances grain development rather than overall plant structure. Overall, the system shows strong potential to optimize inputs and yields, though UAV path tracking errors and system response delays require further refinement to enhance spray uniformity and accuracy under real-world applications. Full article

The surface quality of industrial gaskets directly impacts sealing performance, operational reliability, and market competitiveness. Inadequate or unreliable defect detection in silicone gaskets can lead to frequent maintenance, undetected faults, and security risks in downstream systems. This paper presents VST-YOLOv8, a trustworthy and secure defect detection framework built upon an enhanced YOLOv8 architecture. To address the limitations of C2F feature extraction in the traditional YOLOv8 backbone, we integrate the lightweight Mobile Vision Transformer v2 (ViT v2) to improve global feature representation while maintaining interpretability. For real-time industrial deployment, we incorporate the Gating-Structured Convolution (GSConv) module, which adaptively adjusts convolution kernels to emphasize features of different shapes, ensuring stable detection under varying production conditions. A Slim-neck structure reduces parameter count and computational complexity without sacrificing accuracy, contributing to robustness against performance degradation. Additionally, the Triplet Attention mechanism combines channel, spatial, and fine-grained attention to enhance feature discrimination, improving reliability in challenging visual environments. Experimental results show that VST-YOLOv8 achieves higher accuracy and recall compared to the baseline YOLOv8, while maintaining low latency suitable for edge deployment. When integrated with secure industrial control systems, the proposed framework supports authenticated, tamper-resistant detection pipelines, ensuring both operational efficiency and data integrity in real-world production. These contributions strengthen trust in AI-driven quality inspection, making the system suitable for safety-critical manufacturing processes. Full article

The potential of hops (Humulus lupulus L.) and β-acid extract were evaluated for improving the quality of maralfalfa grass (Cenchrus purpureus) silage (with added sorghum grain, sorghum straw, and urea) during aerobic exposure and their residual effects on in vitro ruminal fermentation characteristics. Silage samples and ground hops pellets (Galena and Chinook varieties) as well as β-acid mixtures were incubated at 37 °C for 24 h and then maintained under aerobic exposure for 12 h. The sample pH, counts of filamentous fungi, yeasts, and total coliforms, and volatile fatty acid (VFA) concentrations were determined. Subsequently, in vitro ruminal fermentation was conducted to determine total gas production and concentrations of hydrogen, methane, carbon dioxide, and VFAs. The β-acid treatment controlled yeast populations, but an increase (p < 0.05) in pH values was observed for the Galena and Chinook treatments compared to the Control. However, pH did not differ significantly (p > 0.05) between the Control and the β-acid treatment. Butyric acid concentrations in the silage were lower (p < 0.05) compared to the Control, except in the silage treatment with Galena. In the in vitro ruminal fermentation, the β-acid treatment showed higher butyric acid levels than the Chinook and Galena, but these differences were not significant (p > 0.05). There were no differences (p > 0.05) in methane between the treatments. An increase (p < 0.05) in propionic acid concentration was observed in the in vitro ruminal fermentation with β-acids. It was concluded that β-acids could help reduce silage deterioration during the aerobic phase, reducing the butyric acid and yeast populations, and their residual effect could improve ruminal fermentation, increasing propionate and acetate concentrations. Full article

The tribological behavior of ferritic ductile iron without ultrafine graphite nodules (FDI) and ferritic ductile iron with ultrafine graphite nodules (FDI-UG) was investigated in the present study. Ultrafine graphite nodules with a count of 3400 nod/mm² were introduced by annealing treatment of quenched ductile iron, which effectively reduced the friction coefficient of ferritic ductile iron from approximately 0.3 to 0.15. This improvement was attributed to the ultrafine graphite nodules, which, due to their small spacing, facilitated a more uniform distribution on the tribological surface. Additionally, the formation of ultrafine graphite nodules in ferritized ductile iron refined the grain size (15 μm) and enhanced the hardness of ferritic ductile iron (183 HV), thereby significantly reducing abrasive wear. The more uniform graphite lubrication on the tribosurface and high hardness of fine ferrite grains in FDI-UG further enhanced wear resistance between the frictional pairs, effectively suppressing adhesion wear at high loads (6 N). Consequently, the ferritic ductile iron containing ultrafine graphite nodules and fine ferrite grains exhibited a superior wear resistance (6.84 × 10⁻³ mm³ and 9.47 × 10⁻³ mm³) compared to its untreated counterpart (9.22 × 10⁻³ mm³ and 11.95 × 10⁻³ mm³). These findings suggest that the incorporation of ultrafine graphite nodules was an effective strategy to enhance the tribological properties of ductile iron. Full article

Small-target detection in Unmanned Aerial Vehicle (UAV) aerial images remains a significant and unresolved challenge in aerial image analysis, hampered by low target resolution, dense object clustering, and complex, cluttered backgrounds. In order to cope with these problems, we present AeroLight, a novel and efficient detection architecture that achieves high-fidelity performance in resource-constrained environments. AeroLight is built upon three key innovations. First, we have optimized the feature pyramid at the architectural level by integrating a high-resolution head specifically designed for minute object detection. This design enhances sensitivity to fine-grained spatial details while streamlining redundant and computationally expensive network layers. Second, a Dynamic Feature Fusion (DFF) module is proposed to adaptively recalibrate and merge multi-scale feature maps, mitigating information loss during integration and strengthening object representation across diverse scales. Finally, we enhance the localization precision of irregular-shaped objects by refining bounding box regression using a Shape-IoU loss function. AeroLight is shown to improve mAP50 and mAP50-95 by 7.5% and 3.3%, respectively, on the VisDrone2019 dataset, while reducing the parameter count by 28.8% when compared with the baseline model. Further validation on the RSOD dataset and Huaxing Farm Drone dataset confirms its superior performance and generalization capabilities. AeroLight provides a powerful and efficient solution for real-world UAV applications, setting a new standard for lightweight, high-precision object recognition in aerial imaging scenarios. Full article

We have developed a generalisable machine learning framework for reservoir quality prediction in deeply buried clastic systems. Applied to the Lower Jurassic deltaic sandstones of the Tilje Formation (Halten Terrace, North Sea), the approach integrates sedimentological facies modelling with mineralogical and petrophysical prediction in a single workflow. Using supervised Extreme Gradient Boosting (XGBoost) models, we classify reservoir facies, predict permeability directly from standard wireline log parameters and estimate the abundance of porosity-preserving grain coating chlorite (gamma ray, neutron porosity, caliper, photoelectric effect, bulk density, compressional and shear sonic, and deep resistivity). Model development and evaluation employed stratified K-fold cross-validation to preserve facies proportions and mineralogical variability across folds, supporting robust performance assessment and testing generalisability across a geologically heterogeneous dataset. Core description, point count petrography, and core plug analyses were used for ground truthing. The models distinguish chlorite-associated facies with up to 80% accuracy and estimate permeability with a mean absolute error of 0.782 log(mD), improving substantially on conventional regression-based approaches. The models also enable prediction, for the first time using wireline logs, grain-coating chlorite abundance with a mean absolute error of 1.79% (range 0–16%). The framework takes advantage of diagnostic petrophysical responses associated with chlorite and high porosity, yielding geologically consistent and interpretable results. It addresses persistent challenges in characterising thinly bedded, heterogeneous intervals beyond the resolution of traditional methods and is transferable to other clastic reservoirs, including those considered for carbon storage and geothermal applications. The workflow supports cost-effective, high-confidence subsurface characterisation and contributes a flexible methodology for future work at the interface of geoscience and machine learning. Full article

Traditional educational assessment systems suffer from inefficient question selection strategies that fail to optimally probe student knowledge while requiring extensive testing time. We present a novel hierarchical probabilistic neural framework that integrates Bayesian inference with symmetric deep neural architectures to enable adaptive, efficient knowledge assessment. Our method models student knowledge as latent representations within a graph-structured concept dependency network, where probabilistic mastery states, updated through variational inference, are encoded by symmetric graph properties and symmetric concept representations that preserve structural equivalences across similar knowledge configurations. The system employs a symmetric dual-network architecture: a concept embedding network that learns scale-invariant hierarchical knowledge representations from assessment data and a question selection network that optimizes symmetric information gain through deep reinforcement learning with symmetric reward structures. We introduce a novel uncertainty-aware objective function that leverages symmetric uncertainty measures to balance exploration of uncertain knowledge regions with exploitation of informative question patterns. The hierarchical structure captures both fine-grained concept mastery and broader domain understanding through multi-scale graph convolutions that preserve local graph symmetries and global structural invariances. Our symmetric information-theoretic method ensures balanced assessment strategies that maintain diagnostic equivalence across isomorphic concept subgraphs. Experimental validation on large-scale educational datasets demonstrates that our method achieves 76.3% diagnostic accuracy while reducing the question count by 35.1% compared to traditional assessments. The learned concept embeddings reveal interpretable knowledge structures with symmetric dependency patterns that align with pedagogical theory. Our work generalizes across domains and student populations through symmetric transfer learning mechanisms, providing a principled framework for intelligent tutoring systems and adaptive testing platforms. The integration of probabilistic reasoning with symmetric neural pattern recognition offers a robust solution to the fundamental trade-off between assessment efficiency and diagnostic precision in educational technology. Full article

Low plant stands at early crop establishment contribute to yield reduction in legumes. Gap-filling has been suggested as a mechanism to mitigate these losses. There is, however, limited understanding of a yield-compensation mechanism in the cowpea. This study aimed to investigate the mechanism of yield-compensation in some accessions of cowpeas at Minjibir and Shika in Northern Nigeria. The experiment was laid out in a split plot fitted into a randomized complete block design in three replicates. The main plot consisted of four plant densities (33,333; 66,666; 99,999; 133,333 plants ha⁻¹) while the subplots consisted of six cowpea accessions (IT89KD-288, IT93K-452-1, IT99K-537-1-1, IT98K-205-8, IT08K-150-27, and DANILA). Results showed that plant density and environment affected grain yield. Total grain yield increased with increasing plant density and was higher at Minjibir than at Shika. The highest total grain yield of 1793.3 kg ha⁻¹ was observed in the accession DANILA at 99,999 plants ha⁻¹, while the lowest (1100 kg ha⁻¹) was observed in the accession IT98K-205-8 at 33,333 plants ha⁻¹. Leaf area index, stand count at harvest, and intercepted photosynthetically active radiation were positively correlated with total grain yield at both locations, suggesting that these traits could be considered for cowpea improvement. Cowpea growers and breeders could consider the erect (IT93K-452-1 and IT98K-205-8) and semi-erect accessions (IT99K-573-1-1 and IT08K-150-27) for cultivation at 133,333 plants ha⁻¹. Prostrate accessions (IT89KD-288 and DANILA) could be planted at 99,999 plants ha⁻¹ at Minjibir. The accessions IT93K-452-1-1, IT98-205-8, IT99K-573-1-1, and IT08K-150-27 could be considered for cultivation at Shika irrespective of plant density. Full article

This study presents an innovative unmanned aerial vehicle (UAV)-based intelligent detection method utilizing an improved Faster Region-based Convolutional Neural Network (Faster R-CNN) architecture to address the inefficiency and inaccuracy inherent in manual wheat spike counting. We systematically collected a high-resolution image dataset (2000 images, 4096 × 3072 pixels) covering key growth stages (heading, grain filling, and maturity) of winter wheat (Triticum aestivum L.) during 2022–2023 using a DJI M300 RTK equipped with multispectral sensors. The dataset encompasses diverse field scenarios under five fertilization treatments (organic-only, organic–inorganic 7:3 and 3:7 ratios, inorganic-only, and no fertilizer) and two irrigation regimes (full and deficit irrigation), ensuring representativeness and generalizability. For model development, we replaced conventional VGG16 with ResNet-50 as the backbone network, incorporating residual connections and channel attention mechanisms to achieve 92.1% mean average precision (mAP) while reducing parameters from 135 M to 77 M (43% decrease). The GFLOPS of the improved model has been reduced from 1.9 to 1.7, an decrease of 10.53%, and the computational efficiency of the model has been improved. Performance tests demonstrated a 15% reduction in missed detection rate compared to YOLOv8 in dense canopies, with spike count regression analysis yielding R² = 0.88 (p < 0.05) against manual measurements and yield prediction errors below 10% for optimal treatments. To validate robustness, we established a dedicated 500-image test set (25% of total data) spanning density gradients (30–80 spikes/m²) and varying illumination conditions, maintaining >85% accuracy even under cloudy weather. Furthermore, by integrating spike recognition with agronomic parameters (e.g., grain weight), we developed a comprehensive yield estimation model achieving 93.5% accuracy under optimal water–fertilizer management (70% ETc irrigation with 3:7 organic–inorganic ratio). This work systematically addresses key technical challenges in automated spike detection through standardized data acquisition, lightweight model design, and field validation, offering significant practical value for smart agriculture development. Full article

Salinity is a major abiotic stress limiting black gram (Vigna mungo) productivity, particularly in arid and semi-arid regions. Saline soils negatively impact plant growth, nodulation, nitrogen fixation, and yield. This study evaluated the efficacy of co-inoculating salt-tolerant plant growth-promoting bacteria Paenibacillus sp. SPR11 and Bradyrhizobium yuanmingense PR3 on black gram performance under saline field conditions (EC: 8.87 dS m⁻¹; pH: 8.37) with low organic carbon (0.6%) and nutrient deficiencies. In vitro assays demonstrated the biocontrol potential of SPR11, inhibiting Fusarium oxysporum and Macrophomina phaseolina by 76% and 62%, respectively. Germination assays and net house experiments under 300 mM NaCl stress showed that co-inoculation significantly improved physiological traits, including germination rate, root length (61.39%), shoot biomass (59.95%), and nitrogen fixation (52.4%) in nitrogen-free media. Field trials further revealed enhanced stress tolerance markers: chlorophyll content increased by 54.74%, proline by 50.89%, and antioxidant enzyme activities (SOD, CAT, PAL) were significantly upregulated. Electrolyte leakage was reduced by 55.77%, indicating improved membrane stability. Agronomic performance also improved, with co-inoculated plants showing increased root length (7.19%), grain yield (15.55 q ha⁻¹; 77.04% over control), total biomass (26.73 q ha⁻¹; 57.06%), and straw yield (8.18 q ha⁻¹). Pod number, seed count, and seed weight were also enhanced. Nutrient analysis showed elevated uptake of nitrogen, phosphorus, potassium, and key micronutrients (Zn, Fe) in both grain and straw. To the best of our knowledge, this is the very first field-based report demonstrating the synergistic benefits of co-inoculating Paenibacillus sp. SPR11 and Bradyrhizobium yuanmingense PR3 in black gram under saline, nutrient-poor conditions without external nitrogen inputs. The results highlight a sustainable strategy to enhance legume productivity and resilience in salt-affected soils. Full article

This study evaluates the applicability and accuracy of coarse-grain modeling (CGM) in discrete-element method (DEM)–based simulations, focusing on particle-mixing efficiency in five representative industrial mixers: the tumbling V mixer, ribbon-blade mixer, paddle-blade mixer, vertical-blade mixer, and conical-screw mixer. Although the DEM is widely employed for particulate system simulations, the high computational cost associated with fine particles significantly hinders large-scale applications. CGM addresses these issues by scaling up particle sizes, thereby reducing particle counts and allowing longer simulation timesteps. We utilized the Lacey mixing index (LMI) as a statistical measure to quantitatively assess mixing uniformity across various CGM scaling factors. Based on the results, CGM significantly reduced computational time (by over 90% in certain cases) while preserving acceptable accuracy levels in terms of LMI values. The mixing behaviors remained consistent under various CGM conditions, based on both visually inspected particle distributions and the temporal LMI trends. Although minor deviations occurred in early-stage mixing, these discrepancies diminished with time, with the final LMI errors remaining below 5% in most scenarios. These findings indicate that CGM effectively enhances computational efficiency in DEM simulations without significantly compromising physical accuracy. This research provides practical guidelines for optimizing industrial-scale particle-mixing processes and conducting computationally feasible, scalable, and reliable DEM simulations. Full article

To enhance the effectiveness of sustainable preservation of modified corn wet distillers’ grains plus solubles (mcWDGS), various additives were tested under aerobic and anaerobic conditions. In Experiment I, the mcWDGS was stored under aerobic conditions for 5 days at 25 °C. Treatments included different organic acids applied at 0.3% or 0.6% of fresh matter (FM). In Experiment II, the mcWDGS was ensiled anaerobically for 8 weeks at 25 °C using organic acids, a commercial acid mixture, or a microbial inoculant at 0.2% FM. In aerobic conditions, the best preservability was achieved with propionic and formic acids at 0.6% FM, as indicated by the lowest temperature, pH, and microbial counts on days 3 and 5 (p ≤ 0.01). Under anaerobic storage, the highest lactic acid concentrations were recorded in the control, citric acid, and commercial acid mixture variants (p ≤ 0.01). Acetic acid levels were highest in the control (p ≤ 0.01). The highest NH3-N content was found in the formic acid variant and the lowest in the inoculant variant (p ≤ 0.01). Aerobic stability after ensiling was greatest in the control and propionic acid groups (p ≤ 0.01). Nutritional analysis showed that the citric acid group had the highest dry matter content (p ≤ 0.01), while the control group contained the most crude protein (p ≤ 0.01) and saturated fatty acids (p ≤ 0.05). The propionic acid and commercial acid mixture variants had the highest unsaturated fatty acids (p ≤ 0.05). Antioxidant capacity was also greatest in the control (p ≤ 0.01). In conclusion, mcWDGS can be effectively preserved aerobically with 0.6% FM of propionic or formic acid, and anaerobically via ensiling, even without additives. These findings support its potential as a stable and nutritious feed ingredient. Full article