KMITL

Permanent URI for this communityhttps://dspace.kmitl.ac.th/handle/123456789/1

Browse

Search Results

Now showing 1 - 10 of 20

Some of the metrics are blocked by your
consent settings
DDoS Detection Framework Using Machine Learning Optimized by Bayesian and PSO Techniques
(2026-07-01)
Sathaporn, Posathip
;
Krungseanmuang, Woranidtha
;
Chaowalittawin, Vasutorn
;
Benjangkaprasert, Chawalit
;
Archevapanich, Tuanjai
This paper presents a distributed denial of service (DDoS) detection framework using machine learning techniques enhanced with hyperparameter optimization for network traffic classification and evaluated on the BCCC-cPacket-Cloud-DDoS-2024 dataset. The framework includes data preprocessing with normalization and class imbalance handling via the synthetic minority over-sampling technique. A critical contribution of this study is the rigorous analysis of the trade-off between detection accuracy and model complexity. Unlike arbitrary feature selection methods, we empirically determined the optimal feature set using information gain, identifying that the top 100 features represent the saturation point that balances high accuracy with minimal overhead. Model performance was further improved through hyperparameter optimization using particle swarm optimization and Bayesian algorithms. The extreme gradient boosting (XGBoost) model optimized using Bayesian optimization and the top 100 features achieved the highest performance, with an accuracy of 99.29% and an F1-score of 98.91%. As a result, the proposed framework improves detection performance while reducing model complexity by selecting an optimal feature set to improve model stability and efficiency.
Some of the metrics are blocked by your
consent settings
INTERNET OF THINGS BASED PRACTICAL SMART ENVIRONMENTAL MONITORING SYSTEM FOR POULTRY FARM
(2026-01-01)
Anuwongpinit, Thanavit
;
Manthawornsiri, Chananont
;
Archevapanich, Tuanjai
;
Purahong, Boonchana
In this article, we propose developing digital innovation of smart monitoring systems in the poultry farm using the Internet of Things (IoT) technology. This work aimed to design and develop a monitoring system based on an IoT system that transforms a traditional farm that uses a manual management system to apply an IoT system for environmental monitoring in a poultry farm. The main components include a hardware component that was designed and implemented to gather data of the poultry houses. Temperature and humidity sensor nodes are applied to monitoring the environment of poultry houses. The LoRa communication module in the sensor node will forward data to the gateway. The second principal component is the cloud server for data acquisition from the gateway. The cloud will be responsible for back-end processing and a web-based dashboard displaying mechanism. This system can work as an alarm notification system using LINE notify API for the LINE application that is the most popular communication application in Thailand. The system was implemented practically in one of the poultry farms in Prachinburi province, Thailand. The results indicate that the proposed system provides significant advantages, including enhanced monitoring accuracy, reduced energy consumption, and improved real-time environmental tracking for poultry farms.
Some of the metrics are blocked by your
consent settings
Banana quality classification using lightweight CNN model with microservice integration system
(2025-06-10)
Chaowalittawin, Vasutorn
;
Krungseanmuang, Woranidtha
;
Sathaporn, Posathip
;
Morita, Fuka
;
Archevapanich, Tuanjai
Banana sorting has been performed manually, which often leads to human error due to the high volume and diverse characteristics involved. This paper presents a banana quality classification system using ConsolutechMobileNetV2 (CST-MobileNetV2) to classify banana ripeness into four categories unripe, ripe, overripe, and rotten. A lightweight deep learning model is proposed and integrated with a uniquely designed microservice system to optimize performance while minimizing computational demands. A publicly available dataset containing 13,478 images was used, and the data split into 56% for training, 14% for validation, and 30% for testing. Image normalization and augmentation techniques were applied to enhance the model's robustness. The model's performance was evaluated using a confusion matrix, achieving 98% precision, recall, and F1-score. The proposed model was compared with other deep learning models to benchmark its performance and deployed in different operating systems to evaluate its flexibility and capabilities. The LINE platform was employed as the user interface, enabling practical interaction with users. The system also demonstrated an average response time of 9.25 seconds per image, ensuring efficient processing, delivers high accuracy and scalability making it a practical and efficient solution for automated banana quality classification.
Some of the metrics are blocked by your
consent settings
2D and 3D LiDAR with CNN Models for Detecting Sediment Accumulation Underground after Disasters
(2025-01-01)
Krungseanmuang, Woranidtha
;
Morita, Fuka
;
Chaowalittawin, Vasutorn
;
Sathaporn, Posathip
;
Kanamori, Chisato
Global climate change impacts all regions and leads to natural disasters such as typhoons, which cause destruction, debris, and flooding. Postdisaster restoration is a very important activity that is mostly done manually and can be time-consuming and challenging, especially in subterranean environments owing to accumulated objects such as pipes, pillars, and mud distributed in confined underground areas. Therefore, in this study, we aim to utilize emerging AI technologies by comparing deep learning algorithms and evaluating four models for 2D object detection and four for 3D point cloud segmentation for detecting sediment accumulation and navigating around obstacles in underground areas after a disaster. Additionally, a custom dataset was developed to simulate underground disaster scenarios. As a result, the You Only Look Once version 11 (YOLOv11) model achieved the highest mean average precision 50 (mAP50: 91.1%) for general detection within the pillar-pipe dataset, whereas the YOLOv12 model performed the best in detecting pipes (mAP50: 87.7%). In the mud dataset, the YOLOv8 segmentation (YOLOv8-seg) model demonstrated superior performance with mAP50 scores of 93.0% (detection) and 86.4% (segmentation). For 3D point cloud segmentation, PointNet achieved the highest accuracy (98.61%), whereas RandLA-Net was optimal for pipe segmentation, achieving an intersection over union score of 37.1%. These findings highlight AI’s potential to accelerate disaster recovery, reduce manual labor, and ensure faster cleanup. Integrating deep learning models into post-typhoon restoration efforts can enable communities to recover more quickly and efficiently after climate change impacts or disaster events.
Some of the metrics are blocked by your
consent settings
Comparative Study of Machine Learning Models for Soil Fertilizer Classification in Precision Agriculture
(2025-01-01)
Archevapanich, Tuanjai
;
Sirikham, Thanapat
;
Chaowalittawin, Vasutorn
;
Krungseanmuang, Woranidtha
;
Sathaporn, Posathip
This study explores the machine learning techniques compare for fertilizer classification based on soil nutrient dataset aligning with the goals of precision agriculture. Five models include Random Forest, Logistic Regression, SVM, XGBoost and Neural Network(ANN) were tested using precision, accuracy, F1-score, recall and confusion matrices. The highest F1-score is XGBoost model, while the best precision performance delivered by Random Forest model. Results emphasize the significance of model selection in handling imbalanced agricultural data. The approach supports data-driven decision-making for sustainable farming aligned with Thailand's 20-Year Agricultural Strategic Plan.
Some of the metrics are blocked by your
consent settings
Implementation of Cloud Computing and Internet of Things (IoT) by Performance Evaluation
(2024-01-01)
Sithiyopasakul, Jiran
;
Archevapanich, Tuanjai
;
Sithiyopasakul, Saran
;
Lasakul, Attasit
;
Purahong, Boonchana
The integration of cloud computing and the Internet of Things (IoT) holds transformative potential across diverse industries. Performance assessment is essential to gauge the quality and efficiency of cloud computing and IoT systems. This paper presents a comprehensive performance evaluation of cloud computing and IoT systems, focusing on three major platforms: Amazon Web Services (AWS), Google Cloud Platform (GCP), and Microsoft Azure. Experimental results encompass various scenarios, including normal operation, heavy load conditions, IoT applications, and scalability testing. The outcomes reveal distinct performance metrics such as response time, throughput, latency, and reliability for each cloud platform.
Some of the metrics are blocked by your
consent settings
Multi-sector Collaboration for Students of different levels to Learn Digital Technology Skills using Competency Assessment from Learning to Create pieces with a 3D printers for Young Innovators Case Study: MogroWittayakom School
(2024-01-01)
Terdyothin, Alongkot
;
Archevapanich, Tuanjai
;
Chansuthirangkool, Manit
;
Inyoo, Pondthip
;
Anuwongpinit, Thanavit
This article is the result of collaboration across various associations to create learning spaces for students in compulsory and optional education as well as living skills. This article provides an example of learning digital technology skills through collaboration, starting with the establishment of an innovation room, followed by instructional activities to explore assessment methods for competency in creating works using a 3D printer tailored for young innovators. By employing rubric- based technology tailored to real-world conditions, it aims to fulfill teachers' expectations for student outcomes. This approach helps students understand and apply a rubric to evaluate and improve their own creations. The demographic involves 20 high school students from MogroWittayakhom, selected through purposive sampling. The students participated in a practical performance test, receiving step-by-step training over two days, totaling 12 hours. Upon completing the instruction, students designed and fabricated their projects using Tinkercad software, which were then printed with a 3D printer. The instructor evaluated the students using a rubric-based competency assessment with three scoring levels. The results showed that all 20 participants achieved a high competency level as per the rubric, validating the hypothesis with 100% of the participants meeting the performance criteria.
Some of the metrics are blocked by your
consent settings
Ensemble of CNN classifiers using Choquet Fuzzy Integral Technique for PCB Defect Classification
(2024-01-01)
Purahong, Boonchana
;
Krungseanmuang, Woranidtha
;
Tenghongsakul, Kasi
;
Archevapanich, Tuanjai
;
Khunthawiwone, Parkpoom
This paper presents a novel method for detecting defects in printed circuit boards (PCBs) using an ensemble of classifiers based on the Choquet fuzzy integral. Our approach employs convolutional neural network (CNN) models, specifically ResNet152, VGG19, and InceptionV3 as base classifiers to identify six types of PCB defects: spurs, mouse bites, short circuits, open circuits, spurious copper, and pinholes. Given the critical role of PCBs in ensuring electronic equipment reliability, effective defect detection methods like ours are essential. We employ pre-trained CNN models for feature extraction and classification of PCB defects. Following this, we combine the prediction scores using the Choquet fuzzy integral to derive more accurate final labels, exceeding the accuracy of standalone models. Our approach is tested on PCB images obtained from public repositories, captured using a linear scan CCD. The evaluation results demonstrate average precision, recall, F-score, and accuracy of 93.0%, 95.2%, 95.1%, and 95.1%, respectively.
Some of the metrics are blocked by your
consent settings
Performance Evaluation of Infrastructure as a Service across Cloud Service Providers
(2023-01-01)
Sithiyopasakul, Saran
;
Archevapanich, Tuanjai
;
Purahong, Boonchana
;
Sithiyopasakul, Paisan
;
Lasakul, Attasit
The purpose of this research aims to monitor, analyze, and compare the performance of infrastructure as a service (IaaS) between the selective cloud providers. To assure which cloud provider has more stability, reliability, and scalability. This paper focuses on performance testing based on a deployed web server in the cloud environment. The main feature of cloud computing is scalability thus most common IaaS cloud service providers (CSPs) have Auto Scaling features for instances or virtual machines. Not only does this paper gives the experimental results of the scaling scalability testing, but it also provides the results of recovery testing to inspect how long a web server is able to recover from failures and load testing which simulated traffic requests. Testing was conducted in the major public clouds of Google Cloud Platform (GCP), Microsoft Azure, and Amazon Web Services (AWS). Azure performs the most efficiently of almost all testing but hardest to configure.
Some of the metrics are blocked by your
consent settings
Deep transfer learning for brain tumor detection based on MRI images
(2023-01-01)
Tenghongsakul, Kasi
;
Kanjanasurat, Isoon
;
Archevapanich, Tuanjai
;
Purahong, Boonchana
;
Lasakul, Attasit
Brain tumors are among the main causes of cancer-related mortality in humans. Early detection of brain tumors is a vital job in the medical task of diagnosis and cure planning for patients. The automatic detection greatly facilitates medical personnel. Magnetic resonance imaging (MRI) is an accepted imaging strategy for diagnosing brain tumors. Presently, deep learning approaches have proven effective in handling various computer vision problems, such as image classification, because of their high performance and also determine models that can learn and decide based on sample data. In this study, the deep transfer learning method, namely InceptionResNet-V2, ResNet50, MobileNet-V2, and VGG16, was used to compare and find the most suitable model for brain tumor detection from the public MRI dataset. Also, CLAHE was employed as an image enhancement technique to improve the quality of the image data set before being used as the model input. As a result, the suggested method performed a prediction accuracy of up to 100%.

KMITL

Browse

Filters

Settings

Sort By

Results per page

Search Results