Training Accuracy Research Articles

Clustering of Urinary Biomarkers to Identify Interstitial Cystitis Subtypes and Different Clinical Characteristics and Treatment Outcomes

Purpose: Interstitial cystitis/bladder pain syndrome (IC/BPS) is mysterious and difficult to diagnose without cystoscopic hydrodistention. This study aimed to explore non-invasive and highly reliable urine biomarkers to identify Hunner’s IC (HIC) and different non-Hunner’s IC (NHIC) subtypes. Methods: In total, 422 women with and without clinically diagnosed IC/BPS (n = 376 and 46, respectively) were retrospectively enrolled. Patients were diagnosed with HIC or NHIC by cystoscopic hydrodistention under anesthesia. Then, the maximal bladder capacity (MBC) and glomerulation grade were determined. Thirteen urine inflammatory cytokines, chemokines, and oxidative stress biomarkers based on the previously reported predictors of IC/BPS were assayed using commercial microsphere kits. The dataset was randomly divided into training (70%) and test (30%) sets for model construction and validation using logistic regression and stepwise variable selection techniques. To construct the predictive models, univariate analysis was performed to evaluate the discriminative power of each urinary biomarker, measured by the area under the curve (AUC). Biomarkers with AUC values < 0.6 were excluded from further modeling. Multivariate logistic regression was then employed, with variables selected through stepwise forward selection based on log-likelihood criteria. For dichotomization, cutoff values were determined using quartile ranges from the control group. The final model’s performance was assessed using AUC, accuracy, sensitivity, and specificity in both training and test sets. Results: By setting the screening criterion to AUC ≥ 0.60, the potential urinary biomarkers for identifying IC/BPS cases were eotaxin, monocyte chemoattractant protein-1, tumor necrosis factor-alpha (TNF-α), 8-hydroxy-2′-deoxyguanosine (8-OHdG), and 8-isoprostane. Those for identifying HIC from the IC/BPS cohort were interleukin (IL)-6, IL-8, interferon γ-inducible protein 10 (IP-10), and regulated on activation, normal T-cell expressed and secreted (RANTES). A diagnostic algorithm using a cluster of urinary biomarkers included TNF-α ≥ 0.95 pg/mL or 8-OHDG ≥ 22.34 pg/mL and 8-isoprastane ≥ 22.34 pg/mL for identifying IC/BPS from the overall cohort; for identifying HIC from the IC/BPS cohort, the urinary IP-10 ≥ 3.74 pg/mL or IP-10 ≥ 19.94 pg/mL was added. Conclusions: Using a cluster of urinary biomarkers such as TNF-α or 8-OHdG and 8-isoprostane can identify IC/BPS from a study cohort, and adding the urinary IP-10 can distinguish HIC from IC/BPS cases.

Wireless 5G Network in Edge Computing Based On MIMO with Federated Learning and Clustering Integrated Reinforcement Learning

Edge Computing (EC) is a revolutionary architecture that brings Cloud Computing (CC) services closer to data sources than ever before. This research proposed novel technique in edge computing network based on wireless 5G technology using MIMO_federated learning integrated with Reinforcement neural network. Here the aim is to enhance the resource allocation by Decentralized Federated learning in multiple user based MIMO (De_Fed_L- MIMO) networks. Then the energy efficiency and channel optimization of the network is carried out using K-means clustering integrated with Reinforcement learning (K-means_RL). Here the experimental analysis is carried out in terms of number of users of network as well as number of edge server by DoF of 92%, Spectral efficiency of 92%, Energy efficiency of 96%, Signal to noise ratio (SNR) of 85%, Coverage area of 92%, RL training accuracy of 95%, FL training accuracy of 98%.

Optimized Random Forest Models for Rock Mass Classification in Tunnel Construction

The accurate prediction of rock mass quality ahead of the tunnel face is crucial for optimizing tunnel construction strategies, enhancing safety, and reducing geological risks. This study developed three hybrid models using random forest (RF) optimized by moth-flame optimization (MFO), gray wolf optimizer (GWO), and Bayesian optimization (BO) algorithms to classify the surrounding rock in real time during tunnel boring machine (TBM) operations. A dataset with 544 TBM tunneling samples included key parameters such as thrust force per cutter (TFC), revolutions per minute (RPM), penetration rate (PR), advance rate (AR), penetration per revolution (PRev), and field penetration index (FPI), with rock classification based on the Rock Mass Rating (RMR) method. To address the class imbalance, the Borderline Synthetic Minority Over-Sampling Technique was applied. Performance assessments revealed the MFO-RF model’s superior performance, with training and testing accuracies of 0.992 and 0.927, respectively, and key predictors identified as PR, AR, and RPM. Additional validation using 91 data sets confirmed the reliability of the MFO-RF model on unseen data, achieving an accuracy of 0.879. A graphical user interface was also developed, enabling field engineers and technicians to make instant and reliable rock classification predictions, greatly supporting safe tunnel construction and operational efficiency. These models contribute valuable tools for real-time, data-driven decision-making in tunneling projects.

Assessment of different U-Net backbones in segmenting colorectal adenocarcinoma from H&E histopathology.

Adenocarcinoma, the most prevalent type of colorectal cancer, makes up roughly 95 % of all cases and is associated with a notably high mortality rate. Owing to the various risk factors which might include personal choices and habits or genetic factors, the risk of developing the cancer for every individual might vary. However, given the statistics, the rate of acquiring the disease is pretty high. Therefore, based on the need for early detection and diagnosis of the disease, there is a pressing demand for an automated system to accurately identify adenocarcinoma in the colorectal region by utilizing the concept of binary segmentation wherein two classes are employed to indicate the presence as well as the absence of the condition. To address this, the project explored several deep learning-based segmentation methods-such as U-Net, Attention U-Net, U-Net with ResNet50 backbone, U-Net with MobileNet-v2 backbone, U-Net with EfficientNetB0 backbone, and U-Net with DenseNet121 backbone-to segment adenocarcinoma regions in histopathological images of the colon and rectum, which are essentially the various U-Net backbones. The performance of each method was then compared to identify the most effective approach, and subsequently, it was found that the U-Net with DenseNet121 backbone and U-Net with ResNet50 backbone performed better than the rest of the models in terms of accuracy with its respective training accuracy scores being 93.81 % and 93.39 % while the testing accuracy scores were 90.21 % and 89.81 %, respectively.

Automating container damage detection with the YOLO-NAS deep learning model.

Ensuring the integrity of shipping containers is crucial for maintaining product quality, logistics efficiency, and safety in the global supply chain. Damaged containers can lead to significant economic losses, delays, and safety hazards. Traditionally, container inspections have been manual, which are labor-intensive, time-consuming, and error-prone, especially in busy port environments. This study introduces an automated solution using the YOLO-NAS model, a cutting-edge deep learning architecture known for its adaptability, computational efficiency, and high accuracy in object detection tasks. Our research is among the first to apply YOLO-NAS to container damage detection, addressing the complex conditions of seaports and optimizing for high-speed, high-accuracy performance essential for port logistics. Our method showcases YOLO-NAS's superior efficacy in detecting container damage, achieving a mean average precision (mAP) of 91.2%, a precision rate of 92.4%, and a recall of 84.1%. Comparative analyses indicate that YOLO-NAS consistently outperforms other leading models like YOLOv8 and Roboflow 3.0, which showed lower mAP, precision, and recall values under similar conditions. Additionally, while models such as Fmask-RCNN and MobileNetV2 exhibit high training accuracy, they lack the real-time assessment capabilities critical for port applications, making YOLO-NAS a more suitable choice. The successful integration of YOLO-NAS for automated container damage detection has significant implications for the logistics industry, enhancing port operations with reliable, real-time inspection solutions that can seamlessly integrate into predictive maintenance and monitoring systems. This approach reduces operational costs, improves safety, and lessens the reliance on manual inspections, contributing to the development of "smart ports" with higher efficiency and sustainability in container management.

An improved approach for retrieval of tidal flat elevation based on inundation frequency

An improved approach for retrieval of tidal flat elevation based on inundation frequency

Predicting breast cancer recurrence using deep learning

Breast cancer and its recurrence are significant health concerns, emphasizing the critical importance of early detection and personalized treatment strategies for improved outcomes. This study introduces the BCR-HDL (Breast Cancer Recurrence using Hybrid Deep Learning) framework, a novel approach designed to predict breast cancer recurrence with high accuracy and interpretability. Utilizing the Wisconsin Diagnostic Breast Cancer and Wisconsin Prognostic Breast Cancer datasets, the framework integrates multiple deep learning architectures- Multi layer Perceptron (MLP), Visual Geometry Group (VGG), Residual Network (ResNet), and Extreme Inception (Xception)-with traditional machine learning models such as Support Vector Machine (SVM), Decision Trees (DT), Random Forest (RF), and Logistic Regression (LR). This hybridization leads to the creation of 16 robust models that enhance interpretability, facilitate generalization, and effectively manage challenges related to small datasets, class imbalance, and data preprocessing. The BCR-HDL framework’s unique contributions include its ability to predict not only diagnostic outcomes but also prognostic and recurrence timing, offering a comprehensive solution for breast cancer management. Specifically, the Hybrid MLP+RF and Xception+RF models achieved an exceptional diagnostic accuracy of 97% on the WDBC dataset, while the Hybrid MLP+RF model reached 78% prognostic accuracy on the WPBC dataset. Moreover, the Hybrid ResNet+SVM and ResNet+RF models demonstrated impressive performance in multi-classifying recurrence into different time intervals, achieving 92% accuracy in predicting recurrence within 2 years, between 2 to 4 years, and beyond 4 years. The study also provides a detailed analysis of model performance through training versus validation accuracy graphs and a comparison with existing approaches, demonstrating the superiority of the proposed framework in terms of diagnostic, prognostic, and recurrence time predictions. The BCR-HDL framework offers practical recommendations for clinicians, including its potential for personalized treatment strategies and improved patient monitoring, making it a valuable tool for advancing breast cancer management.

Eye Diseases Classification Using Transfer Learning Technique of Deep Neural Network

The human eye is a complex organ susceptible to various diseases and abnormalities that can impair vision. Identifying retinal diseases early and precisely categorizing them is crucial to ensure effective treatment is administered promptly. In this proposed study, we implemented Convolutional Neural Networks (CNNs) along with pretrained neural network architectures. We use existing methods as the foundation and develop models tailored for detecting eye diseases, such as Diabetic Retinopathy (DR), Cataracts, and Glaucoma, specifically among diabetic patients. These models are applied for screening of retinal fundus images for accurate diagnosis. Prompt identification and suitable treatment can aid in preventing the commencement and advancement of diabetic retinopathy, cataracts, and glaucoma among patients. Accurate identification of these conditions is paramount in the healthcare sector. Our research aims to categorize eye diseases through the utilization of retinal fundus images for multiclass classification. The multiclass classification outcomes, implemented on the Kaggle platform, indicated that CNN, MobileNet, DenseNet121, ResNet and EfficientNetB3 yielded training accuracies of 98.50%, 99.94%, 99.37%, 99.90% and 99.97 as well as the validation accuracies of 88.22%, 95.67%, 93.99%, 93.30% and 95.97% respectively.

A Surrogate-Assisted Intelligent Adaptive Generation Framework for Cost-Effective Coal Blending Strategy in Thermal Power Units

The coal cost of coal-fired units accounts for more than 70% of the total power generation cost. In addition to determining coal costs, coal blending strategies (CBS) significantly impact various types of costs, such as pollutant removal and emissions. To address these issues, we propose a framework for generating cost-effective CBS. The framework includes a unit output condition recognition module (UOCR) that enables the adaptive classification of output conditions based on historical operation datasets, performing intelligent condition recognition with Imitator and pre-trained image classification models using blending strategies and unit parameters as inputs. The cost-effective strategy generation module (CESG) employs a surrogate model to evaluate the economic viability of strategies in terms of coal and environmental costs, among other factors. It also employs UOCR as another surrogate model to validate strategy feasibility. Cost-effective strategies are generated via a population-based metaheuristic algorithm. In the case study, the UOCR achieved an average training accuracy of 96.64%, and the generated cost-effective strategies reduced costs by an average of 3.37% compared to currently implemented strategies.

Tuberculosis Detection Using Convolutional Neural Network

Tuberculosis (TB) is a chronic lung disease caused by bacterial infection and remains one of the leading causes of mortality worldwide. Early and accurate detection of TB is crucial for effective treatment and prevention of complications. In this study, we propose a deep learning-based system for TB detection using chest X-ray images. The system leverages the VGG16 architecture for classification, supported by image preprocessing and data augmentation to improve model performance. The dataset comprises X-ray images categorized into TB-infected and normal cases. The VGG16 model was trained, validated, and tested to classify chest X-rays, achieving a train accuracy of 98.55%, validation accuracy of 99.75% and test accuracy of 98.73%, for TB detection. Our results demonstrate the effectiveness of the VGG16-based model in reliably detecting TB from chest X-rays while providing practical healthcare accessibility through the user interface. This combined approach can serve as a valuable tool in computer-aided TB diagnosis, enabling timely and accurate detection to improve patient outcomes. Key Words: Chest X-ray Classification, Deep Learning, Medical Image Analysis, Computer-Aided Diagnosis (CAD), Transfer Learning, TB Diagnosis.

Artificial Intelligence for Enhancing Veterinary Healthcare: An Experiment

Background: The use of machine learning (ML) in veterinary medicine has gained significant attention, particularly for the early detection and classification of animal diseases. For example: lumpy Skin Disease (LSD) in cattle is one such condition that poses a substantial threat to livestock health. Traditional diagnostic methods can be labor-intensive and time-consuming. Therefore, leveraging ML techniques for automated disease detection could improve diagnostic efficiency and accuracy. Methods: In this study, a dataset comprising images of cattle from four breeds-Vechur, Swiss Holstein, Jersey and Ponwar-was collected to train a DenseNet-121 Convolutional Neural Network (CNN) model for identifying and classifying LSD in cattle. The dataset included both LSD-affected and non-affected cattle images, ensuring balanced representation across different breeds, ages and severity levels. Image preprocessing techniques such as resizing, normalization and data augmentation were applied to prepare the data for model training. The DenseNet-121 architecture was employed, utilizing a pretrained ImageNet model as the feature extractor, with additional layers for binary classification. Result: The model achieved excellent performance, with training accuracy reaching 99.76% and validation accuracy of 92.17%. The precision, recall and F1-score for the “Lumpy” class were 96.43%, 81.82% and 88.52%, respectively, while the “Normal” class had a precision of 92%, recall of 98.57% and F1-score of 95.17%. The overall accuracy of the model was 93.2%. Additionally, the model achieved an AUC score of 0.96 for both classes, indicating a high ability to distinguish between “Lumpy” and “Normal” cattle. These results highlight the potential of ML-based methods in enhancing the efficiency and accuracy of veterinary disease diagnostics.

A tool wear monitoring method driven by workpiece surface quality inversion and EfficientNet

Abstract Tool wear status monitoring plays an important role in the safety of machine and the efficiency of production. Due to the limitation of sensor installation and complex operation conditions, traditional machine vision methods are difficult to directly act on the tool itself to obtain corresponding wear status information. Meanwhile, the surface quality of the machined workpiece can provide indirect insights into the wear status of the tool to a significant degree. Therefore, an indirect tool wear status monitoring method is proposed in this paper by workpiece surface quality inversion and EfficientNet. Firstly, the surface texture data of workpiece are obtained in the form of image by an industrial camera, and the representative features that indirectly reflect tool wear status are further highlighted by using the weighted average image grayscale and the Laplacian operator. Secondly, to monitor tool wear effectively, we have optimized three key settings, width, model depth, and input image resolution, of the EfficientNet backbone network. This optimization boosts its ability to understand and generalize, ensuring it can handle various surface textures and extract crucial information accurately. This, in turn, enhances the accuracy of our tool wear monitoring system. Finally, the improved EfficientNet network is further utilized as a pre-trained network to achieve tool wear state monitoring. Utilizing the powerful representation ability of the pre-trained model and the surface texture features of workpiece, the model can accurately identify and judge the wear degree of the tool. The tool experiment results show that the proposed tool wear monitoring method has achieved remarkable results with recognition accuracy of 97.77%. While other popular methods such as VGG19, ResNet50, and ResNet152, whose training accuracies are 95.66%, 96.37%, and 91.27%, respectively. The comparative results highlight the excellent performance of the proposed method in recognition rate, training time and generalization performance.

EARLY WARNING PREDICTION WITH AUTOMATIC LABELLING IN EPILEPSY PATIENTS

Abstract Early warning for epilepsy patients is crucial for their safety and well being, in particular, to prevent or minimize the severity of seizures. Through the patients’ electroencephalography (EEG) data, we propose a meta learning framework to improve the prediction of early ictal signals. The proposed bilevel optimization framework can help automatically label noisy data at the early ictal stage, as well as optimize the training accuracy of the backbone model. To validate our approach, we conduct a series of experiments to predict seizure onset in various long-term windows, with long short-term memory (LSTM) and ResNet implemented as the baseline models. Our study demonstrates that not only is the ictal prediction accuracy obtained by meta learning significantly improved, but also the resulting model captures some intrinsic patterns of the noisy data that a single backbone model could not learn. As a result, the predicted probability generated by the meta network serves as a highly effective early warning indicator.

Colorectal cancer detection with enhanced precision using a hybrid supervised and unsupervised learning approach

The current work introduces the hybrid ensemble framework for the detection and segmentation of colorectal cancer. This framework will incorporate both supervised classification and unsupervised clustering methods to present more understandable and accurate diagnostic results. The method entails several steps with CNN models: ADa-22 and AD-22, transformer networks, and an SVM classifier, all inbuilt. The CVC ClinicDB dataset supports this process, containing 1650 colonoscopy images classified as polyps or non-polyps. The best performance in the ensembles was done by the AD-22 + Transformer + SVM model, with an AUC of 0.99, a training accuracy of 99.50%, and a testing accuracy of 99.00%. This group also saw a high accuracy of 97.50% for Polyps and 99.30% for Non-Polyps, together with a recall of 97.80% for Polyps and 98.90% for Non-Polyps, hence performing very well in identifying both cancerous and healthy regions. The framework proposed here uses K-means clustering in combination with the visualisation of bounding boxes, thereby improving segmentation and yielding a silhouette score of 0.73 with the best cluster configuration. It discusses how to combine feature interpretation challenges into medical imaging for accurate localization and precise segmentation of malignant regions. A good balance between performance and generalization shall be done by hyperparameter optimization-heavy learning rates; dropout rates and overfitting shall be suppressed effectively. The hybrid schema of this work treats the deficiencies of the previous approaches, such as incorporating CNN-based effective feature extraction, Transformer networks for developing attention mechanisms, and finally the fine decision boundary of the support vector machine. Further, we refine this process via unsupervised clustering for the purpose of enhancing the visualisation of such a procedure. Such a holistic framework, hence, further boosts classification and segmentation results by generating understandable outcomes for more rigorous benchmarking of detecting colorectal cancer and with higher reality towards clinical application feasibility.

Improvement of K-2 rifle’s live-fire accuracy using virtual reality shooting training system

Virtual reality (VR) is used in military training with simulators for marksmanship training. Despite evidence of the effectiveness of VR simulation-based training, less is known about the association between VR marksmanship training and record-fire accuracy. In addition, the VR simulations utilized in prior research were largely rooted in laser infrared technology, which does not consider the ballistic trajectory of bullets. Thus, this study was designed to examine the effect of VR-based shooting training systems, Multi-purpose Assault and sniper Rifles-shooting Simulator (MARS) training that applies a six-degree-of-freedom trajectory of a rigid body, on live-fire accuracy. The participants were divided into three groups (non-MARS training, two-MARS training, and four or more-MARS training), and their live-fire scores were compared, considering the effects of psychological variables. Additionally, the proportions of elite shooters were computed among the groups. Analysis of variance showed that individuals who were exposed to MARS training four times or more showed better K2C1 rifle live-fire accuracy than those in two-MARS training and non-MARS training. In addition, the rate of elite shooters was the highest in the four or more-MARS groups. These results demonstrate that the VR simulation training using MARS, which applies an algorithm for a six-degree-of-freedom trajectory of a rigid body, can improve the K2C1 rifle live-fire accuracy. Future studies are needed to provide stronger evidence for the optimal number of MARS training programs to improve live-fire shooting performance.

Attention-Based Interpretable Multiscale Graph Neural Network for MOFs.

Metal-organic frameworks (MOFs) hold great potential in gas separation and storage. Graph neural networks (GNNs) have proven effective in exploring structure-property relationships and discovering new MOF structures. Unlike molecular graphs, crystal graphs must consider the periodicity and patterns. MOFs' specific features at different scales, such as covalent bonds, functional groups, and global structures, influenced by interatomic interactions, exert varying degrees of impact on gas adsorption or selectivity. Moreover, redundant interatomic interactions hinder training accuracy, leading to overfitting. This research introduces a construction method for multiscale crystal graphs, which considers specific features at different scales by decomposing the crystal graph into multiple subgraphs based on interatomic interactions within varying distance ranges. Additionally, it takes into account the global structure of the crystal by encoding the periodic patterns of the unit cells. We propose MSAIGNN, a multiscale atomic interaction graph neural network with self-attention-based graph pooling mechanism, which incorporates three-body bond angle information, accounts for structural features at different scales, and minimizes interference from redundant interactions. Compared with traditional methods, MSAIGNN demonstrates higher prediction accuracy in assessing single-component adsorption, gas separation, and structural features. Visualization of attention scores confirms effective learning of structural features at different scales, highlighting MSAIGNN's interpretability. Overall, MSAIGNN offers a novel, efficient, multilayered, and interpretable approach for property prediction of complex porous crystal structures like MOFs using deep learning.

Designing Face Detection Systems with Gray Wolf Optimization

The main objective of this paper project was to create a state-of-the-art face identification technique that can handle the various difficulties caused by changes in illumination, occlusions, and facial emotions. Face detection is a cornerstone of computer vision, facilitating diverse applications ranging from surveillance systems to human-computer interaction. Throughout this paper, the comprehensive exploration of advancing face detection methodologies has been undertaken, culminating in developing and evaluating a novel approach. The challenges posed by variations in facial expressions, lighting conditions, and occlusions necessitated a multifaceted solution. Our proposed method, which consists of interconnected steps, works quite well to overcome these challenges. Using deep learning architectures to increase feature extraction and discrimination was beneficial in the initial stage of fine-tuning Residual Networks (ResNet-50) to serve as the Region-based Convolutional Neural Network (Faster R-CNN) framework classifier. The process of gradually optimizing thresholds, such as batch size, learning rate, and detection threshold, involved using the Gray Wolf optimization technique (GWO). The conversion process was accelerated and improved overall detection process efficiency and accuracy using a clever fusion of machine learning and metaheuristic optimization techniques. A key component of our methodology is the careful data processing, which was necessary to ensure. The suggested method was carefully examined on a particular dataset, and the 94% training accuracy that was attained together with an identical test dataset accuracy highlights the method's resilience. These findings support the effectiveness of our approach in reducing false positives and negatives, resulting in unmatched recall and precision in the detection system. The discovery has significant significance as it can potentially improve face detection systems' performance and reliability in various real-world applications, such as human-computer interaction and surveillance. Convolutional neural networks, deep learning architectures, and metaheuristic optimization approaches were synergized to produce a new and reliable solution.

Enhancing Coffee Leaf Rust Detection Using DenseNet201: A Comprehensive Analysis of the Mbozi and Public Datasets in Songwe, Tanzania

Coffee Leaf Rust (CLR) is a worldwide devastating fungal disease that threatens coffee production, upsetting economic and farmers' livelihoods. Traditional methods of detecting CLR heavily rely on using machine-learning (ML) models trained through weakly collected datasets and physical inspection which is tedious, time-consuming, and subject to human error. This study explores the performance of the DenseNet201 model using three datasets: Mbozi, Public, and Combined (a merger of Mbozi and Public datasets). Machine Learning Theory guided this research. The study objective is to assess the influence of dataset quality on CLR detection, analyze Mbozi and Public datasets using DenseNet201, and enhance robustness by merging the two datasets. A study on coffee leaf rot (CLR) severity was conducted using systematic sampling techniques. Leaves from multiple coffee farms were collected, representing different levels of infection. The Mbozi dataset, sourced from high-resolution images captured from Tanzania's Songwe coffee plantations, was analyzed for quality under controlled conditions, including environmental factors, image clarity, resolution, labeling consistency, and class balance, based on data completeness, image quality score, visual inspection, and model performance. DenseNet201 was trained and validated on each dataset achieving its highest accuracy with the Mbozi dataset at 98.72% and a validation accuracy of 97.65%, demonstrating the importance of consistent image quality and accurate annotations. In contrast, the public dataset suffered from inconsistencies in resolution and labeling, resulting in a lower training and validation accuracy of 96.86% and 96.42% respectively. The Combined dataset, which integrated the strengths of both datasets, exhibited a stronger generalization with an accuracy of 97.48% and validation accuracy of 97.49%, balancing the need for high-quality images with environmental variability. The study shows improved CLR detection speed and accuracy due to high-quality and consistently labeled images from the Mbozi dataset. It recommends future models integrate regionally relevant and high-resolution datasets for robust performance in real-world agricultural conditions, providing coffee farmers with timely disease intervention tools for better production management and economic stability in coffee-growing regions.

Changes in the Timing of Autumn Leaf Senescence of Maple and Ginkgo Trees in South Korea over the Past 30 Years: A Comparative Assessment of Process-Based, Linear Regression, and Machine-Learning Models

Changes in vegetation activities driven by climate change serve as both a sensitive indicator and a key driver of climate impacts, underscoring the need for accurate phenological predictions. Delays in leaf senescence due to rising air temperatures increase the risk of damage from early frost, potentially affecting growth and survival in subsequent years. This study aimed to quantify long-term changes in leaf senescence timing for palmate maple and ginkgo trees, explore their associations with environmental factors, and compare the performance of multiple modeling approaches to identify their strengths and limitations for phenological predictions. Using data from 48 sites across South Korea (1989–2020), this study analyzed trends in the timing of leaf senescence for maple and ginkgo trees and compared the performance of process-based models (CDD_T, CDD_P, TP_T, TP_P), a linear regression model, and machine-learning models (random forest, RF; gradient-boosting decision tree, GBTD). Leaf senescence timing for both species has progressively been delayed, with ginkgo trees showing a faster rate of change (0.20 vs. 0.17 days per year, p < 0.05). Delayed senescence was observed in most regions (81% for maple and 75% for ginkgo), with statistically significant delays (p < 0.05) at half of the sites. Machine-learning models demonstrated the highest training accuracy (RMSE < 4.0 days, r > 0.90). Evaluation with independent datasets revealed that the RF and process-based TP_P (including minimum temperature and photoperiod) using a site-specific approach performed best (RMSE < 5.5 days, r > 0.75). Key environmental factors identified by RF included autumn minimum or mean temperatures and a summer photoperiod. By conducting this comparative assessment, the study provides insights into the applicability of different modeling approaches for phenology research and highlights their implications for vegetation management and climate change adaptation.

FusionNet remote a hybrid deep learning ensemble model for remote image classification in multispectral images

Remote sensing plays a vital role in various industries, including smart cities, agriculture, and environmental monitoring, by capturing multispectral images of the Earth’s surface and analyzing its features. Accurate classification of these images is essential for extracting valuable information and making informed decisions. However, traditional image classification models, such as Convolutional Neural Networks (CNNs), often struggle to fully capture the complex spectral and spatial information inherent in multispectral images. Addressing this research gap, we propose FusionNet-Remote, a novel hybrid deep learning ensemble model that integrates CNNs with Random Forests (RF) to enhance the accuracy and robustness of remote image classification. The primary objective of this study is to develop a model that combines the strengths of CNNs for spatial feature extraction with the robust classification capabilities of RFs. Through comprehensive evaluations using LANDSAT data, FusionNet-Remote demonstrates superior performance compared to existing models, achieving an outstanding training accuracy of 99.8% and a testing accuracy of 98.7%. The model also exhibits the lowest training loss (0.0001) and testing loss (0.0043), significantly outperforming traditional models such as standalone CNNs, VGG16, and Inceptionv3. These results highlight the effectiveness of the hybrid ensemble approach in overcoming the limitations of conventional methods. This research contributes to the field by introducing an innovative ensemble technique that significantly improves remote image classification accuracy and reliability. The model’s high performance underscores its potential for critical applications such as land cover mapping, vegetation analysis, and disaster monitoring. Future work will explore the integration of more advanced CNN architectures and alternative fusion techniques to further enhance the model's performance and adaptability across different remote sensing scenarios.