Search Results (2,043)

The precise identification of plant diseases is essential for improving agricultural productivity and reducing reliance on human expertise. Deep learning frameworks, belonging to the YOLO series, have demonstrated significant potential in the real-time detection of plant diseases. There are various factors influencing model performance; activation functions play an important role in improving both accuracy and efficiency. This study proposes αSiLU, a modified activation function developed to optimize the performance of YOLOv11n for plant disease-detection tasks. By integrating a scaling factor α into the standard SiLU function, αSiLU improved the effectiveness of feature extraction. Experiments are conducted on two different plant disease datasets—tomato and cucumber—to demonstrate that YOLOv11n models equipped with αSiLU outperform their counterparts using the conventional SiLU function. Specifically, with α = 1.05, mAP@50 increased by 1.1% for tomato and 0.2% for cucumber, while mAP@50–95 improved by 0.7% and 0.2% each. Additional evaluations across various YOLO versions confirmed consistently superior performance. Furthermore, notable enhancements in precision, recall, and F1-score were observed across multiple configurations. Crucially, αSiLU achieves these performance improvements with minimal effect on inference speed, thereby enhancing its appropriateness for application in practical agricultural contexts, particularly as hardware advancements progress. This study highlights the efficiency of αSiLU in the plant disease-detection task, showing the potential in applying deep learning models in intelligent agriculture. Full article

This study aimed to assess the effect of acidic storage and simulated brushing on the translucency and color stability of 3D-printed resins for prosthodontic applications. Three 3D printed resin materials—Ceramic Crown (CC), OnX (ONX), and Tough 2 (T2)—were compared with a CAD/CAM milled nano-ceramic resin material (Lava Ultimate, LU). Twelve specimens were fabricated from each material and were allocated into two groups based on the storage medium (water or citric acid), followed by simulated tooth brushing for 3650 cycles. The specimens’ translucency (TP) and color stability (ΔE) were determined using a spectrophotometer. The data was compared using ANOVA, independent student t-tests, and a post hoc Tukey test (p < 0.05). Multiple comparisons of mean differences in TP revealed significant differences between the tested materials (p < 0.001), except for groups CC and ONX. Irrespective of the groups, all materials showed decreased TP values after simulated tooth brushing. Regarding color stability, CC (0.66 ± 0.42) and T2 (1.40 ± 0.34) in acid demonstrated the least and greatest color changes, respectively. The ΔE did not vary between the materials or between the storage media (p > 0.05). Except for T2 and LU in water, the other materials showed ΔE values below the perceptibility threshold of 1.2. The material type and storage media affected the translucency of the tested materials. However, regardless of the material type and storage media, there was no discernible impact on the color change of the tested materials. Full article

Elevated chloride (Cl⁻) concentrations in surface water and groundwater are an increasing concern in cold region urban environments, largely due to long-term road salt application. This study investigates the Cl⁻ distribution across southern Ontario, Canada, using geospatial methods to identify contamination hot spots and assess groundwater vulnerability at both regional and watershed scales. Chloride data from 2001 to 2010 and 2011 to 2020 were compiled from public sources and interpolated using inverse distance weighting. A regional-scale vulnerability index was developed using slope (SL), surficial geology (SG), and land use (LU) (SL-SG-LU), and compared it to a more detailed DRASTIC-LU index within the Credit River watershed. Results show that Cl⁻ hot spots are concentrated in urbanized areas, including the Greater Toronto Area and Golden Horseshoe, with some rural zones also exhibiting elevated concentrations. Vulnerability mapping corresponded well with the observed Cl⁻ patterns and highlighted areas at risk for groundwater discharge to surface waters. While the DRASTIC-LU method offered finer resolution, the simplified SL-SG-LU index effectively captured broad vulnerability trends and is suitable for data-limited regions. This work provides a transferable framework for identifying Cl⁻ risk areas and supports long-term monitoring and management strategies in cold climate watersheds. Full article

The present study applied a GIS-based methodology for assessing soil diversity in a protected mountain area of Italy. Using QGIS, morphological (i.e., altitude and slope), lithological, climatic, and land use layers were intersected to delineate 16 land units (LUs), each representing relatively homogeneous conditions for soil formation, according to Jenny’s equation. To obtain the soil map units, a total of 112 soil profiles were analyzed, including 79 from previous studies and 33 that were newly excavated during 2023–2024 to fill gaps in underrepresented LU types. Most soils were classified as Inceptisols/Cambisols, occurring in both Dystric and Eutric variants, mainly in relation to lithology (i.e., arenaceous or pelitic facies). Alfisols, Umbrisols, and hydromorphic soils were also identified. The physicochemical properties showed marked variability among LUs, with sand content ranging from 39 to 798 g kg⁻¹, pH from 4.4 to 7.9, and organic carbon content from 1.6 to 6.1%. This LU-based framework allowed efficient field sampling, if compared to grid-based surveys, while retaining information on fine-scale pedodiversity. No quantitative accuracy assessment (e.g., boundary precision, internal homogeneity metrics) was conducted, even if the spatial coherence of the delineated LUs was supported by the distribution of soil profiles, which provided empirical validation of the LU framework. Full article

Interstitial lung diseases (ILD) significantly impact health and mortality, affecting millions of individuals worldwide. During the COVID-19 pandemic, lung ultrasonography (LUS) became an indispensable diagnostic and management tool for lung disorders. However, utilising LUS to diagnose ILD requires significant expertise. This research aims to develop an automated and efficient approach for diagnosing ILD from LUS videos using AI to support clinicians in their diagnostic procedures. We developed a binary classifier based on a state-of-the-art CSwin Transformer to discriminate between LUS videos from healthy and non-healthy patients. We used a multi-centric dataset from the Royal Melbourne Hospital (Australia) and the ULTRa Lab at the University of Trento (Italy), comprising 60 LUS videos. Each video corresponds to a single patient, comprising 30 healthy individuals and 30 patients with ILD, with frame counts ranging from 96 to 300 per video. Each video is annotated using the corresponding medical report as ground truth. The datasets used for training the model underwent selective frame filtering, including reduction in frame numbers to eliminate potentially misleading frames in non-healthy videos. This step was crucial because some ILD videos included segments of normal frames, which could be mixed with the pathological features and mislead the model. To address this, we eliminated frames with a healthy appearance, such as frames without B-lines, thereby ensuring that training focused on diagnostically relevant features. The trained model was assessed on an unseen, separate dataset of 12 videos (3 healthy and 9 ILD) with frame counts ranging from 96 to 300 per video. The model achieved an average classification accuracy of 91%, calculated as the mean of three testing methods: Random Sampling (92%), Key Featuring (92%), and Chunk Averaging (89%). In RS, 32 frames were randomly selected from each of the 12 videos, resulting in a classification with 92% accuracy, with specificity, precision, recall, and F1-score of 100%, 100%, 90%, and 95%, respectively. Similarly, KF, which involved manually selecting 32 key frames based on representative frames from each of the 12 videos, achieved 92% accuracy with a specificity, precision, recall, and F1-score of 100%, 100%, 90%, and 95%, respectively. In contrast, the CA method, where the 12 videos were divided into video segments (chunks) of 32 consecutive frames, with 82 video segments, achieved an 89% classification accuracy (73 out of 82 video segments). Among the 9 misclassified segments in the CA method, 6 were false positives and 3 were false negatives, corresponding to an 11% misclassification rate. The accuracy differences observed between the three training scenarios were confirmed to be statistically significant via inferential analysis. A one-way ANOVA conducted on the 10-fold cross-validation accuracies yielded a large F-statistic of 2135.67 and a small p-value of 6.7 × 10⁻²⁶, indicating highly significant differences in model performance. The proposed approach is a valid solution for fully automating LUS disease detection, aligning with clinical diagnostic practices that integrate dynamic LUS videos. In conclusion, introducing the selective frame filtering technique to refine the dataset training reduced the effort required for labelling. Full article

Plugs of lignite coal from multiple formations were subjected to a series of tests to determine the amount of rare-earth elements (REEs) to be extracted from coal in an in situ mining operation. These tests were used to determine if extraction of REEs and other critical minerals in an in situ environment would be possible for future attempts as an alternative to extraction mining. The tests involved subjecting whole lignite coal plugs from the Twin Butte coal seams in North Dakota to flow-through tests of water, and concentrations of 1.0 M ammonium nitrate, 1.0 M and 1.5 M sulfuric acid, and 1.0 M and 1.5 M hydrochloric acid (HCl) solvents at different concentrations and combinations. The flow-through testing was conducted by alternating the solvent and water flow-through to simulate an in situ mining scenario. The samples were analyzed for their concentrations of REEs (lanthanum [La], cerium [Ce], praseodymium [Pr], neodymium [Nd], samarium [Sm], europium [Eu], gadolinium [Gd], terbium [Tb], dysprosium [Dy], holmium [Ho], erbium [Er], thulium [Tm], ytterbium [Yb], lutetium [Lu], yttrium [Y], and scandium [Sc], as well as germanium [Ge] and cobalt [Co], manganese [Mn], nickel [Ni], and barium [Ba]). Results from the testing showed that REEs were extracted in concentrations that were on average higher using sulfuric acid (8.9%) than with HCl (5.8%), which had a higher recovery than ammonium nitrate. Tests were performed over a standard time interval for comparison between solvents, while a second set of testing was done to determine recovery rates of REEs and critical minerals under certain static and constant flow-through times to determine extraction in relation to time. Critical minerals had a higher recovery rate than the REEs across all tests, with a slightly higher recovery of light REEs over heavy REEs. Full article

Background/Objectives: Parkinson’s disease (PD) is a progressive neurodegenerative disorder that requires early and accurate diagnosis. This study aimed to classify individuals with and without PD using volumetric brain MRI data and to improve model interpretability using explainable artificial intelligence (XAI) techniques. Methods: This retrospective study included 79 participants (39 PD patients, 40 controls) recruited at Inonu University Turgut Ozal Medical Center between 2013 and 2025. A deep neural network (DNN) was developed using a multilayer perceptron architecture with six hidden layers and ReLU activation functions. Seventeen volumetric brain features were used as the input. To ensure robust evaluation and prevent overfitting, a stratified five-fold cross-validation was applied, maintaining class balance in each fold. Model transparency was explored using two complementary XAI techniques: the Contrastive Explanation Method (CEM) and Local Interpretable Model-Agnostic Explanations (LIME). CEM highlights features that support or could alter the current classification, while LIME provides instance-based feature attributions. Results: The DNN model achieved high diagnostic performance with 94.1% accuracy, 98.3% specificity, 90.2% sensitivity, and an AUC of 0.97. The CEM analysis suggested that reduced hippocampal volume was a key contributor to PD classification (–0.156 PP), whereas higher volumes in the brainstem and hippocampus were associated with the control class (+0.035 and +0.150 PP, respectively). The LIME results aligned with these findings, revealing consistent feature importance (mean = 0.1945) and faithfulness (0.0269). Comparative analyses showed different volumetric patterns between groups and confirmed the DNN’s superiority over conventional machine learning models such as SVM, logistic regression, KNN, and AdaBoost. Conclusions: This study demonstrates that a deep learning model, enhanced with CEM and LIME, can provide both high diagnostic accuracy and interpretable insights for PD classification, supporting the integration of explainable AI in clinical neuroimaging. Full article

Targeted alpha therapy (TAT) has recently emerged as a highly promising approach for the management of metastatic castration-resistant prostate cancer (mCRPC), especially in patients with disease progression despite standard treatments. Among alpha-emitter radiopharmaceuticals, actinium-225-labelled prostate-specific membrane antigen ([225Ac]Ac-PSMA) has shown remarkable potential due to its high linear energy transfer (LET), short path length, and ability to induce potent, localised cytotoxic effects. This review summarises current clinical evidence regarding [225Ac]Ac-PSMA radioligand therapy (RLT), emphasising its efficacy, safety profile, and position relative to beta-emitter therapy with lutetium-177 ([177Lu]Lu-PSMA). Data from compassionate-use programs and small clinical trials demonstrate that [225Ac]Ac-PSMA produces significant biochemical and imaging responses, including > 50% declines in prostate-specific antigen (PSA) and lesion regression on [68Ga]Ga-PSMA PET/CT, even in heavily pre-treated mCRPC cohorts. Xerostomia, renal toxicity, and haematological adverse effects remain the main safety challenges, necessitating optimisation of patient selection, dosing strategies, and salivary gland protection protocols. Compared with [177Lu]Lu-PSMA, [225Ac]Ac-PSMA appears effective even in cases of beta-refractory disease, highlighting its complementary role rather than a competitive alternative. However, limited availability, high production costs, and the lack of large-scale, randomised trials hinder widespread clinical adoption. Future directions include combination protocols, improved radiopharmaceutical design, and trials evaluating its use in earlier disease stages. This review provides a comprehensive overview of the clinical aspects of [225Ac]Ac-PSMA RLT and its evolving role in advanced prostate cancer management. Full article

Due to the nearly identical chemical properties of Lu and Yb, the production of no-carrier-added (NCA) ¹⁷⁷Lu faces significant challenges in their separation. Achieving efficient and streamlined separation of Lu and Yb is crucial for the production of NCA ¹⁷⁷Lu. This study systematically investigated the separation performance of the commercial Rext-P350 extraction resin for Lu and Yb. Static adsorption experiments revealed that, at a solid–liquid ratio of 8 g/L, both Lu³⁺ and Yb³⁺ were nearly completely adsorbed, with saturation adsorption capacities of 25.8 mg/g and 21.5 mg/g, respectively. An increase in the nitric acid concentration in the aqueous phase significantly inhibited adsorption, but the separation factor for Lu³⁺/Yb³⁺ remained above 1.88. The adsorption kinetics followed a pseudo-second-order model (R² > 0.99), with equilibrium reached within 15 min, demonstrating fast adsorption kinetics. Characterization by SEM, FT-IR, and XPS confirmed the chemical coordination between the resin and Lu³⁺/Yb³⁺. Dynamic chromatographic separation experiments showed that the Rext-P350 resin exhibited significantly better separation performance for Lu³⁺/Yb³⁺ compared to 2-ethylhexylphosphoric acid mono-2-ethylhexyl ester (P507) extraction resin. Leveraging the excellent performance of Rext-P350 resin, a two-stage continuous extraction chromatography process was designed, achieving efficient separation of 0.045 mg of Lu³⁺ from 200 mg of Yb³⁺ with a Lu³⁺ purity of 90.9% and a yield of 98.4%. This study provides a feasible separation technique for the purification of NCA ¹⁷⁷Lu. Full article

Vehicle detection plays a pivotal role in traffic management as a key technology for intelligent traffic management and driverless driving. However, current deep learning-based vehicle detection models face several challenges in practical applications. These include slow detection speeds, large computational and parametric quantities, high leakage and misdetection rates in target-intensive environments, and difficulties in deploying them on edge devices with limited computing power and memory. To address these issues, this paper proposes an improved vehicle detection method called SMA-YOLO, based on the YOLOv7 model. Firstly, MobileNetV3 is adopted as the new backbone network to lighten the model. Secondly, the SimAM attention mechanism is incorporated to suppress background interference and enhance small-target detection capability. Additionally, the ACON activation function is substituted for the original SiLU activation function in the YOLOv7 model to improve detection accuracy. Lastly, SIoU is used to replace CIoU to optimize the loss of function and accelerate model convergence. Experiments on the UA-DETRAC dataset demonstrate that the proposed SMA-YOLO model achieves a lightweight effect, significantly reducing model size, computational requirements, and the number of parameters. It not only greatly improves detection speed but also maintains higher detection accuracy. This provides a feasible solution for deploying a vehicle detection model on embedded devices for real-time detection. Full article

During traditional aircraft landings, pilots face significant challenges in identifying runway numbers with the naked eye, particularly at decision height under adverse weather conditions. To address this issue, this study proposes a novel detection algorithm based on an optimized version of the YOLOv5s model (You Only Look Once, version 5) for recognizing runway markings during civil aircraft landings. By integrating a data augmentation strategy with external datasets, the method effectively reduces both false detections and missed targets through expanded feature representation. An Alpha Complete Intersection over Union (CIOU) Loss function is introduced in place of the original CIOU Loss function, offering improved gradient optimization. Additionally, the model incorporates several advanced modules and techniques, including a Convolutional Block Attention Module (CBAM), Soft Non-Maximum Suppression (Soft-NMS), cosine annealing learning rate scheduling, the FReLU activation function, and deformable convolutions into the backbone and neck of the YOLOv5 architecture. To further enhance detection, a specialized small-target detection layer is added to the head of the network and the resolution of feature maps is improved. These enhancements enable better feature extraction and more accurate identification of smaller targets. As a result, the optimized model shows significantly improved recall (R) and precision (P). Experimental results, visualized using custom-developed software, demonstrate that the proposed optimized YOLOv5s model achieved increases of 5.66% in P, 2.99% in R, and 2.74% in mean average precision (mAP) compared to the baseline model. This study provides valuable data and a theoretical foundation to support the accurate visual identification of runway numbers and other reference markings during aircraft landings. Full article

Deep-learning-based multiple label chest X-ray classification has achieved significant success, but existing models still have three main issues: fixed-scale convolutions fail to capture both large and small lesions, standard pooling is lacking in the lack of attention to important regions, and linear classification lacks the capacity to model complex dependency between features. To circumvent these obstacles, we propose CONVFCMAE, a lightweight yet powerful framework that is built on a backbone that is partially frozen (77.08 % of the initial layers are fixed) in order to preserve complex, multi-scale features while decreasing the number of trainable parameters. Our architecture adds (1) an intelligent global pooling module that is learnable, with $1\times 1$ convolutions that are dynamically weighted by their spatial location, and (2) a multi-head attention block that is dedicated to channel re-calibration, along with (3) a two-layer MLP that has been enhanced with ReLU, batch normalization, and dropout. This module is used to enhance the non-linearity of the feature space. To further reduce the noise associated with labels and the imbalance in class distribution inherent to the NIH ChestXray14 dataset, we utilize a combined loss that combines BCEWithLogits and Focal Loss as well as extensive data augmentation. On ChestXray14, the average ROC–AUC of CONVFCMAE is 0.852, which is 3.97 percent greater than the state of the art. Ablation experiments demonstrate the individual and collective effectiveness of each component. Grad-CAM visualizations have a superior capacity to localize the pathological regions, and this increases the interpretability of the model. Overall, CONVFCMAE provides a practical, generalizable solution to the problem of extracting features from medical images in a practical manner. Full article

Micro-motors are essential for power drive systems, and efficient fault diagnosis is crucial to reduce safety risks and economic losses caused by failures. However, the fault signals from micro-motors typically exhibit weak and unclear characteristics. To address this challenge, this paper proposes a novel fault diagnosis method that integrates jump plus AM-FM mode decomposition (JMD), symmetrized dot pattern (SDP) visualization, and an improved convolutional neural network (ICNN). Firstly, we employed the jump plus AM-FM mode decomposition technique to decompose the mixed fault signals, addressing the problem of mode mixing in traditional decomposition methods. Then, the intrinsic mode functions (IMFs) decomposed by JMD serve as the multi-channel inputs for symmetrized dot pattern, constructing a two-dimensional polar coordinate petal image. This process achieves both signal reconstruction and visual enhancement of fault features simultaneously. Finally, this paper designed an ICNN method with LeakyReLU activation function to address the vanishing gradient problem and enhance classification accuracy and training efficiency for fault diagnosis. Experimental results indicate that the proposed JMD-SDP-ICNN method outperforms traditional methods with a significantly superior fault classification accuracy of up to 99.2381%. It can offer a potential solution for the monitoring of electromechanical structures under complex conditions. Full article

The Sverdrup Basin, Arctic Canada, is ideally situated to contain an archive of tectono-magmatic and climatic events that occurred within the wider Arctic region, including the exhumation of the adjacent (northeastern) part of the Canadian-Greenlandic Shield. To test this, a multi-analytical provenance study of Middle Jurassic to Cretaceous sandstones from the eastern Sverdrup Basin was undertaken. Most of the samples analysed were recycled from sedimentary rocks of the Franklinian Basin, with possible additional contributions from the Mesoproterozoic Bylot basins and metasedimentary shield rocks. The amount of high-grade metamorphic detritus in samples from central Ellesmere Island increased from Middle Jurassic times. This is interpreted to reflect exhumation of the area to the southeast/east of the Sverdrup Basin. Exhumation may have its origins in Middle Jurassic extension and uplift along the northwest Sverdrup Basin margin. Rift-flank uplift along the Canadian–West Greenland conjugate margin and lithospheric doming linked with the proximity of the Iceland hotspot and/or the emplacement of the Cretaceous High Arctic Large Igneous Province may have contributed to exhumation subsequently. The southeast-to-northwest thickening of Jurassic to Early Cretaceous strata across the Sverdrup Basin may be a distal effect of exhumation rather than rifting in the Sverdrup or Amerasia basins. Full article

In order to improve mathematical model accuracy of electromechanical actuator (EMA) and solve the problems of low-frequency response and large overshoot for nonlinear systems by using traditional proportional integral derivative (PID) algorithm, a fuzzy single neuron (FSN)-PID algorithm is proposed. Firstly, a complete multi-nonlinear dynamic model of EMA is constructed, which introduces internal friction and current limiter of brushless direct current motors (BLDCMs), dead zone backlash of gear trains, and LuGre friction between output shaft and fin. Secondly, a FSN-PID controller is introduced into the automatic position regulator (APR) of EMA control system, where the gain coefficient K of SN algorithm is adjusted by fuzzy control, and the stability of the controller is proved. In addition, simulations are conducted on the response effect of different fin positions under different algorithms for the analysis of the 6° fin position response; it can be concluded that the rise time with FSN-PID algorithm can be reduced by about 4.561% compared to PID, about 1.954% compared to fuzzy (F)-PID, about 0.875% compared to single neuron (SN)-PID, and about 0.380% compared to back propagation (BP)-PID. For the 4°-2 Hz sine position tracking analysis, it can be concluded that the minimum phase error of FSN-PID algorithm is about 0.4705 ms, which is about 74.44% smaller than PID, about 73.43% smaller than F-PID, about 17.24% smaller than SN-PID, and about 10.81% smaller than BP-PID. Finally, wind tunnel experiments investigate the actual high dynamic flight environment and verify the excellent position tracking ability of FSN-PID algorithm. Full article