Accurate measurement of range of motion (ROM) during physical rehabilitation is traditionally achieved using goniometers or multi-camera, marker-based motion capture systems. The latter, while highly accurate, require specialised laboratory infrastructure, are costly, and are unsuitable for home-based use. There is growing interest in markerless, camera-based alternatives that leverage artificial intelligence (AI) for joint angle estimation. This study presents a preliminary proof-of-concept evaluation of a hybrid AI vision system, combining the MediaPipe Pose framework (version 0.8.9.1, Full model variant) with a Mamdani-type fuzzy inference system, for joint angle estimation and exercise repetition counting using a single consumer-grade camera. Fourteen healthy adult rehabilitation staff members performed three exercises (elbow flexion, knee extension, and hip external rotation) simultaneously recorded by the proposed AI system and by a reference optical motion capture system (C-Motion Visual3D™, 14 calibrated infrared cameras, 53 retro-reflective markers). Reliability was assessed using two-way mixed-effects intraclass correlation coefficients (ICC[Formula: see text]) with absolute agreement. Individual-measure ICC values ranged from poor to good across joints and conditions (range: 0.005-0.68). The highest reliability was observed for the right hip external rotation ROM (individual ICC [Formula: see text]; average ICC [Formula: see text], 95% CI: 0.63-0.89). Several measures showed poor reliability, including the right elbow flexion ROM (individual ICC [Formula: see text]) and the minimum angle of the left hip external rotation (individual ICC [Formula: see text]). Confidence intervals were wide throughout. This proof of concept demonstrates that a single-camera AI system can capture general trends in joint angles during simple rehabilitation exercises in healthy adults under controlled conditions. However, reliability is inconsistent and frequently poor, and the system has not been evaluated in patient populations with pathological movement patterns. These findings do not support clinical deployment at this stage. Substantial further development and validation in patient cohorts are required before clinical or home-based use can be considered.

As solar photovoltaic (PV) systems become vital to sustainable energy, accurate monitoring of solar power plant capacity is essential for maximizing energy yield and maintaining operational reliability, especially under variable environmental conditions. This paper introduces a novel hybrid framework that uniquely integrates physics-informed feature engineering with boosted ensemble learning, AdaBoostM2 for classification and LSBoost for regression, to classify and predict the operational capacity of a 5 MW solar PV plant in Thailand. Leveraging meteorological data (global radiation, ambient/module temperatures) and electrical metrics (active power, generating capacity, and plant capacity) collected at hourly intervals for 365 days, the model accurately classifies plant performance into six operational modes: Optimal, Sub optimal, Severe sub-optimal, Critical, Emergency, and Offline, while predicting power output with robust precision. Unlike conventional physics-informed neural networks or data-driven ML, our approach embeds lightweight, interpretable physics constraints into tree-based ensembles, ensuring physical consistency, real-time scalability (<4 ms/sample inference), and causal interpretability via OOB feature importance. The AdaBoostM2 classifier achieves 100% test accuracy in mode classification, while the LSBoost regressor delivers a mean absolute error (MAE) of 0.0276 MW and an R-squared (R²) of 0.9975, surpassing standalone ML baselines in relative error reduction and outperforming state-of-the-art PIML baselines by 95% in RMSE reduction, and 88% in MAE reduction. By embedding validated physical constraints into a data-driven approach, this framework enables early detection of critical inefficiencies, supporting operational uptime improvements and maintenance prioritization in tropical climates. This work advances beyond existing PIML models by offering a deployable, interpretable, high-precision tool for sustainable solar asset management.

Central inverters are widely used in large photovoltaic systems, but struggle with inefficiency and large energy losses with direct current (DC). To solve these problems, Artificial Intelligence (AI) and the Internet of Things (IoT) offer alternative solutions. In this study, a low-cost open-source IoT system is proposed for the 5 MW Thap-Sakae photovoltaic plant in Thailand, where the DC power parameters are collected in an AI-based time-series fault classification mode. The fault diagnosis data includes five entries per set for no fault, open circuit fault and shading fault. The system achieves 94% accuracy in the diagnosis of non-linear faults and time series parameters. Compared to other solutions, the long-range private wide area network provides cost-effective communication and supports data transmission of up to 180 m. In this study, conducted in a photovoltaic plant with unregulated conditions, a low-cost AI-powered IoT solution has been shown to be effective in real-time fault classification on the DC side. The proposed solution is sustainable and easy to manage over time, as it can handle non-linear problems caused by the volatility of the system or the fluctuations of the sensors without significant changes.

Accurate prediction of stroke risk at an early stage is essential for timely intervention and prevention, especially given the serious health consequences and economic burden that strokes can cause. In this study, we proposed a class-balanced and data-augmented (CBDA-ResNet50) deep learning model to improve the prediction accuracy of the well-known ResNet50 architecture for stroke risk. Our approach uses advanced techniques such as class balancing and data augmentation to address common challenges in medical imaging datasets, such as class imbalance and limited training examples. In most cases, these problems lead to biased or less reliable predictions. To address these issues, the proposed model assures that the predictions are still accurate even when some stroke risk factors are absent in the data. The performance of CBDA-ResNet50 improves by using the Adam optimizer and the ReduceLROnPlateau scheduler to adjust the learning rate. The application of weighted cross entropy removes the imbalance between classes and significantly improves the results. It achieves an accuracy of 97.87% and a balanced accuracy of 98.27%, better than many of the previous best models. This shows that we can make more reliable predictions by combining modern deep-learning models with advanced data-processing techniques. CBDA-ResNet50 has the potential to be a model for early stroke prevention, aiming to improve patient outcomes and reduce healthcare costs.

A natural or man-made disaster must be predicted accurately and in time. Advanced modelling is essential to predict increasingly frequent and intense natural events, enabling better response and recovery strategies. Hybrid machine learning is used to predict floods, assess earthquake damage, and control wildfires. Convolutional Neural Networks (CNNs), Gradient Boosting Machines (GBMs) and Support Vector Machines (SVMs) are used in the proposed model. We evaluate the proposed machine learning model in comparison to conventional statistical analysis and single machine learning techniques. The proposed hybrid model provided 90% correct predictions for flood events in Bangladesh with precision and recall values of 88% and 85%, respectively. In the assessment of earthquake damage in Japan, an accuracy of 92% was achieved with a precision of 90% and a recall of 89% with an F1 score of 89%. In the management of wildfires in California, an accuracy of 88% with a precision of 85% were achieved. The proposed hybrid model outperforms conventional techniques due to its higher reliability in predicting floods.

Rehabilitation after a stroke is vital for regaining functional abilities. However, a shortage of rehabilitation professionals leads to many patients with severe disabilities. Traditional rehabilitation methods can be time-consuming and hard to measure for progress. This study introduces an innovative machine learning (ML) approach for lower limb rehabilitation in stroke patients. The proposed methodology integrates two models: a fuzzy logic rule-based system and a K-Nearest Neighbor(K-NN) machine learning model. The rule-based model utilizes the Fugl-Meyer Assessment to evaluate lower limb angles during exercises using a camera without human intervention. The hybrid fuzzy logic-based ML model continuously tracks the desired angle, counts exercise repetitions, and provides real-time feedback on patient progress. Furthermore, it measures the Range of Motion (ROM) for each repetition, presenting a graphical visualization of ROMs for ten repetitions simultaneously. The model facilitates real-time evaluation of rehabilitation progress by clinicians, with the lowest observed error rate of $$0.34^\circ$$ of angle measurement. The K-NN model assesses rehabilitation exercise accuracy levels, presenting results graphically, with machine learning accuracy rates of $$97\%$$ , $$92\%$$ , and $$91\%$$ for hip flexion, hip external rotation, and knee extension rehabilitation exercises. Model training utilized data from 30 experienced physical therapists at King Chulalongkorn Memorial Hospital, Bangkok, Thailand, garnering positive evaluations from rehabilitation doctors. The proposed ML-based models offer real-time and prerecorded video capabilities, enabling telerehabilitation applications. This research highlights the potential of ML-based methodologies in stroke rehabilitation to enhance accuracy, efficiency, and patient outcomes.

Photovoltaic systems have been under development for many years. However, their energy production is low compared to other systems. From 2016 to 2022, the research goal was to develop a fault diagnosis that evolves from inverter functional analysis to comprehensive monitoring systems. Currently, it is a major challenge to propose a commercially viable product. This review article focuses on fault diagnosis algorithms on the PV DC side for real-world applications. We analyzed 100 studies published between 2017 and 2024, considering nine factors PV array size on the DC side, data collection method, number of data sets, the fault itself, fault location, diagnosis accuracy, diagnosis time, input data for the algorithm and the diagnosis algorithm used. We summarize the strengths and weaknesses of each study in terms of practical implementation and highlight new technological trends. In addition, we discuss how new technologies for PV fault diagnosis on the DC side are tested and evaluated. Our contribution is intended to guide research in the field of PV fault diagnostics and help ensure that it can be used commercially in the future.

Effective disaster prediction is essential for disaster management and mitigation. This study addresses a multi-classification problem and proposes the Neural-XGBoost disaster prediction model (N-XGB), a hybrid model that combines neural networks (NN) for feature extraction with XGBoost for classification. The NN component extracts high-level features, while XGBoost uses gradient-boosted decision trees for accurate predictions, combining the strengths of deep learning and boosting techniques for improved accuracy. The N-XGB model achieves an accuracy of 94.8% and an average F1 score of 0.95 on a real-world dataset that includes wildfires, floods and earthquakes, significantly outperforming baseline models such as random forest, Support vector machine and logistic regression 85% accuracy. The balanced F1 scores for wildfires 0.96, floods 0.93, and earthquakes 0.96 demonstrate the model’s robustness in multi-class classification. The Synthetic Minority Oversampling Technique (SMOTE) balances datasets and improves model efficiency and capability. The proposed N-XGB model provides a reliable and accurate solution for predicting disasters and contributes to improving preparedness, resource allocation and risk management strategies.

The natural abundance of solar energy and its favorable environmental impacts have led to growing acceptance and utilization as the world transitions from a fossil fuel-based economy to a cleaner, more sustainable, and ecologically friendly alternative. An increasing number of solar photovoltaic projects have emerged to capitalize on these benefits. However, factors such as weather conditions, exposure to natural elements, device deterioration, improper installations, and maintenance concerns can eventually compromise these systems' power-generating performance and dependability over time. This study demonstrates how the random forest machine learning classifier can be used to monitor and categorize the performance of solar PV plants to ensure their reliability and efficiency. The study collected and preprocessed a field dataset containing various environmental and electrical parameters. These parameters include plant capacity, active power, generating capacity, ambient temperature, global radiation, and module temperature. This data was used to train a random forest machine learning algorithm to categorize the operational condition of the solar power plant. The results demonstrate that the model efficiently classified the operational condition of the solar power plant by effectively identifying conditions such as optimal mode, moderate sub-optimal mode, severe sub-optimal mode, critical mode, emergency mode, and offline mode, achieving 99.92% accuracy on the test set. The proposed method provides an effective technique for estimating and monitoring the generating capacity and operational condition of solar power plants in near real time, thereby serving as a distinctive instrument for boosting the efficiency and reliability of solar power generation.

Dry gas pipelines can encounter various operational, technical, and environmental issues, such as corrosion, leaks, spills, restrictions, and cyber threats. To address these difficulties, proactive maintenance and management and a new technological strategy are needed to increase safety, reliability, and efficiency. A novel neural network model for forecasting the life of a dry gas pipeline system and detecting the metal loss dimension class that is exposed to a harsh environment is presented in this study to handle the missing data. The proposed strategy blends the strength of deep learning techniques with industry-specific expertise. The main advantage of this study is to predict the pipeline life with a significant advantage of predicting the dimension classification of metal loss simultaneously employing a Bayesian regularization-based neural network framework when there are missing inputs in the datasets. The proposed intelligent model, trained on four pipeline datasets of a dry gas pipeline system, can predict the health condition of pipelines with high accuracy, even if there are missing parameters in the dataset. The proposed model using neural network technology generated satisfactory results in terms of numerical performance, with MSE and R2 values closer to 0 and 1, respectively. A few cases with missing input data are carried out, and the missing data is forecasted for each case. Then, a model is developed to predict the life condition of pipelines with the predicted missing input variables. The findings reveal that the model has the potential for real-world applications in the oil and gas sector for estimating the health condition of pipelines, even if there are missing input parameters. Additionally, multi-model comparative analysis and sensitivity analysis are incorporated, offering an extensive comprehension of multi-model prediction abilities and beneficial insights into the impact of various input variables on model outputs, thereby improving the interpretability and reliability of our results. The proposed framework could help business plans by lowering the chance of severe accidents and environmental harm with better safety and reliability.

The area of computer numerical control (CNC) machines has grown fast, and their use has risen significantly in recent years. This article presents the design and development of a CNC writing machine that uses an Arduino, a motor driver, a stepper motor, and a servo motor. The machine is meant to create 2D designs and write in numerous input languages using 3-axis simultaneous interpolated operations. The suggested machine is low-cost, simple to build, and can be operated with merely G codes. The performance of the CNC writing machine was assessed by testing it on a range of solid surfaces, including paper, cardboard, and wood. The results reveal that the machine can generate high-quality text and images with great accuracy and consistency. The proposed machine's ability to write in several input languages makes it appropriate for various applications, including art, design, and engineering.

Cerebrovascular diseases such as stroke are among the most common causes of death and disability worldwide and are preventable and treatable. Early detection of strokes and their rapid intervention play an important role in reducing the burden of disease and improving clinical outcomes. In recent years, machine learning methods have attracted a lot of attention as they can be used to detect strokes. The aim of this study is to identify reliable methods, algorithms, and features that help medical professionals make informed decisions about stroke treatment and prevention. To achieve this goal, we have developed an early stroke detection system based on CT images of the brain coupled with a genetic algorithm and a bidirectional long short-term Memory (BiLSTM) to detect strokes at a very early stage. For image classification, a genetic approach based on neural networks is used to select the most relevant features for classification. The BiLSTM model is then fed with these features. Cross-validation was used to evaluate the accuracy of the diagnostic system, precision, recall, F1 score, ROC (Receiver Operating Characteristic Curve), and AUC (Area Under The Curve). All of these metrics were used to determine the system’s overall effectiveness. The proposed diagnostic system achieved an accuracy of 96.5%. We also compared the performance of the proposed model with Logistic Regression, Decision Trees, Random Forests, Naive Bayes, and Support Vector Machines. With the proposed diagnosis system, physicians can make an informed decision about stroke.

Age-related macular degeneration (AMD) diagnosis using fundus images is one of the critical missions of the eye-care screening program in many countries. Various proposed deep learning models have been studied for this research interest, which aim to achieve the mission and outperform human-based approaches. However, research efforts are still required for the improvement of model classification accuracy, sensitivity, and specificity values. In this study, we proposed the model named as ViT-AMD, which is based on the latest Vision Transformer (ViT) structure, to diagnosis a fundus image as normal, dry AMD, or wet AMD types. Unlike convolution neural network models, ViT consists of the attention map layers, which show more effective performance for image classification task. Our training process is based on the 5-fold cross-validation and transfer learning techniques using Chula-AMD dataset at the Department of Ophthalmology, the King Chulalongkorn Memorial Hospital, Bangkok. Furthermore, we also test the performance of trained model using an independent image datasets. The results showed that for the 3-classes AMD classification (normal vs. dry AMD vs. wet AMD) on the Chula-AMD dataset, the averaged accuracy, precision, sensitivity, and specificity of our trained model are about 93.40%, 92.15%, 91.27%, and 96.57%, respectively. For result testing on independent datasets, the averaged accuracy, precision, sensitivity, and specificity of trained model are about 74,20%, 75.35%, 74.13%, and 87.07%, respectively. Compared with the results from the baseline CNN-based model (DenseNet201), the trained ViT-AMD model has outperformed significantly. In conclusion, the ViT-AMD model have proved their usefulness to assist the ophthalmologist to diagnosis the AMD disease.

There has been a notable increase in the adoption of solar energy as a sustainable energy choice in recent years due to its positive environmental impact, and widespread availability. However, factors such as weather, component deterioration, improper installation, and maintenance concerns gradually create conditions in the system that threaten its performance and reliability. This study provides a performance evaluation of different supervised machine learning methods for diagnosing photovoltaic cells. Solar power plant datasets containing environmental and electrical parameters were collected, pre-processed, and used to train various machine learning models, including Decision Tree, Long-Short Term Memory, Multilayer Perceptron, Temporal Convolutional Network, Convolutional Neural Network, Naive Bayes, Support Vector Machine, K-Nearest Neighbor, Random Forest, and Linear Discriminant Analysis algorithms. The models are evaluated using accuracy, precision, recall, f1-score, and confusion matrix metrics. Results demonstrated that the models diagnosed short circuit, abnormal voltage, partial shading, and hotspot faults with varying degrees of accuracy.

In order to support the development of more efficient spectrum management by using Big Data and Artificial Intelligence (AI), the authors study and propose a methodological framework that allows the application of Big Data and AI into spectrum management. The authors benchmark how spectrum regulators across the world are currently applying Big Data and AI technologies into their spectrum management, together with advantage(s) and disadvantage(s) of each of the approaches. The authors analyze the current status of the spectrum management under Thailand’s Office of the National Broadcasting and Telecommunications Commission (NBTC). Moreover, the authors identify gaps that might exist between the current status and the aimed future in which Big Data and AI technologies could be applied, and how to close the gaps so that the more efficient spectrum management could be achieved. Based on these studies and analyses, the authors propose a framework and a prototype of a web application applying Big Data and AI Platform to support the mission on spectrum management of the Office of the NBTC, Thailand.

Stroke has a negative impact on people’s lives and is one of the leading causes of death and disability worldwide. Early detection of symptoms can significantly help predict stroke and promote a healthy lifestyle. Researchers have developed several methods to predict strokes using machine learning (ML) techniques. However, the proposed systems have suffered from the following two main problems. The first problem is that the machine learning models are biased due to the uneven distribution of classes in the dataset. Recent research has not adequately addressed this problem, and no preventive measures have been taken. Synthetic Minority Oversampling (SMOTE) has been used to remove bias and balance the training of the proposed ML model. The second problem is to solve the problem of lower classification accuracy of machine learning models. We proposed a learning system that combines an autoencoder with a linear discriminant analysis (LDA) model to increase the accuracy of the proposed ML model for stroke prediction. Relevant features are extracted from the feature space using the autoencoder, and the extracted subset is then fed into the LDA model for stroke classification. The hyperparameters of the LDA model are found using a grid search strategy. However, the conventional accuracy metric does not truly reflect the performance of ML models. Therefore, we employed several evaluation metrics to validate the efficiency of the proposed model. Consequently, we evaluated the proposed model’s accuracy, sensitivity, specificity, area under the curve (AUC), and receiver operator characteristic (ROC). The experimental results show that the proposed model achieves a sensitivity and specificity of 98.51% and 97.56%, respectively, with an accuracy of 99.24% and a balanced accuracy of 98.00%.

Age-related macular degeneration (AMD) is a typical fundus disease that affects the central vision of elderly people. It causes difficulties in everyday activities such as reading and recognizing faces. AMD can progress slowly or rapidly, and it leads to severe vision loss if left untreated. Therefore, early detection and diagnosis of AMD are crucial to prevent or delay vision impairment in the elderly. To handle this requirement, researchers are exploring deep learning-based models as an AI tool to assist ophthalmologist in AMD diagnosis. However, conducting an appropriate deep learning model for the AMD classification is challenging and cost-intensive. This research aims to evaluate the efficacy of various deep learning models for obtaining the best performance results when identifying AMD disease using retinal images. To meet this objective, the retinal images from the Department of Ophthalmology, the King Chulalongkorn Memorial Hospital, Thailand were collected for transfer learning and other publicly available datasets for testing. Then, seven deep learning models VGG19, Xception, DenseNet201, EfficientNetB7, InceptionV3, NASNetLarge, and ResNet152V2 were chosen to training for the 2-labels (Normal vs. AMD) and the 3-labels (Normal vs. Dry AMD vs. Wet AMD) classifications. From the experimental results, the DenseNet201 model with Dense block in its structure showed the best efficacy in both 2-labels and 3-labels AMD classifications since its performance always include in the Top-3 models accuracy and generalization performance measured by total accuracy and total F1-Score, respectively. Furthermore, the accuracy performance of deep learning models in Top-3 are comparable with the performance of retinal specialist. These results contribute consolidated knowledge to the process of implementation effective deep learning as production that detects AMD automatically in the clinical and enhance the quality of healthcare service.

Stroke is a significant cause of mortality and disability globally, with its occurrence in the human brain and motor function being linked to various parts of the human body. Stroke victims often experience disabilities or mobility problems in affected body parts, either on one or both sides of the body. Physiotherapy exercises are the primary treatment and medication for stroke patients, necessitating daily monitoring by a physiotherapist. However, this approach is expensive, and rehabilitation centers and physicians are scarce. Numerous research studies have been conducted to address this issue from various perspectives. This study proposes a Convolutional Neural Network (CNN) based pose net machine learning (ML) model for stroke home rehabilitation using pose detection and classification with a skeleton base model and human pose estimation drawing. We trained our ML model using different human pose images for this model. The tested accuracy of our CNN model is 100% for our exercise pose image test case. Later we build and mobile application for remote rehabilitation. Tested the application with ten different subjects, achieving a 98% for elbow extension, and in the case of elbow flexion and Normal position of balancing both sides, all the persons achieved a 100 % accuracy rate under a laboratory environment.

Solar energy is one of the most promising renewable energy sources, however, anomalies like short circuit faults, abnormal voltage, partial shading, and hotspot faults may reduce its dependability and efficiency. This research offers a machine learning-based solar cell anomaly detection system that employs discrete wavelet transform, convolutional neural network, and multilayer perceptron algorithm for anomaly detection and classification. The authors collected and processed solar farm datasets of various operating parameters, including daily generating capacity, total active power, solar irradiance, ambient temperature, and solar module temperature. A CNN and an MLP machine learning models were trained for anomaly detection and classification using pre-processed data obtained from discrete wavelet transform decomposition. The proposed method accurately identified short-circuit faults, abnormal voltage, partial shading, and hotspot faults, which are difficult to detect using traditional approaches. Machine learning-based solar cell anomaly detection can improve solar power plant efficiency and reliability by detecting irregularities in real-time.