KMITL

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

Browse

Search Results

Now showing 1 - 10 of 34

Some of the metrics are blocked by your
consent settings
Enhancing risk communication and environmental crisis management through satellite imagery and AI for air quality index estimation
(2024-06-01)
Jitkajornwanich, Kulsawasd
;
Vijaranakul, Nattadet
;
Jaiyen, Saichon
;
Srestasathiern, Panu
;
Lawawirojwong, Siam
Due to climate change, the air pollution problem has become more and more prominent [23]. Air pollution has impacts on people globally, and is considered one of the leading risk factors for premature death worldwide; it was ranked as number 4 according to the website [24]. A study, ‘The Global Burden of Disease,’ reported 4,506,193 deaths were caused by outdoor air pollution in 2019 [22,25]. The air pollution problem is become even more apparent when it comes to developing countries [22], including Thailand, which is considered one of the developing countries [26]. In this research, we focus and analyze the air pollution in Thailand, which has the annual average PM2.5 (particulate matter 2.5) concentration falls in between 15 and 25, classified as the interim target 2 by 2021′s WHO AQG (World Health Organization's Air Quality Guidelines) [27]. (The interim targets refer to areas where the air pollutants concentration is high, with 1 being the highest concentration and decreasing down to 4 [27,28]). However, the methodology proposed here can also be adopted in other areas as well. During the winter in Thailand, Bangkok and its surrounding metroplex have been facing the issue of air pollution (e.g., PM2.5) every year. Currently, air quality measurement is done by simply implementing physical air quality measurement devices at designated—but limited number of locations. In this work, we propose a method that allows us to estimate the Air Quality Index (AQI) on a larger scale by utilizing Landsat 8 images with machine learning techniques. We propose and compare hybrid models with pure regression models to enhance AQI prediction based on satellite images. Our hybrid model consists of two parts as follows: • The classification part and the estimation part, whereas the pure regressor model consists of only one part, which is a pure regression model for AQI estimation. • The two parts of the hybrid model work hand in hand such that the classification part classifies data points into each class of air quality standard, which is then passed to the estimation part to estimate the final AQI. From our experiments, after considering all factors and comparing their performances, we conclude that the hybrid model has a slightly better performance than the pure regressor model, although both models can achieve a generally minimum R2 (R2 > 0.7). We also introduced and tested an additional factor, DOY (day of year), and incorporated it into our model. Additional experiments with similar approaches are also performed and compared. And, the results also show that our hybrid model outperform them. Keywords: climate change, air pollution, air quality assessment, air quality index, AQI, machine learning, AI, Landsat 8, satellite imagery analysis, environmental data analysis, natural disaster monitoring and management, crisis and disaster management and communication.
Some of the metrics are blocked by your
consent settings
FAST DEFECT DETECTION FOR GLASS BOTTLE USING AUTOENCODER AND ERROR THRESHOLD
(2024-06-01)
Hanskunatai, Anantaporn
;
Jaiyen, Saichon
;
Claypo, Niphat
Glass bottle defect detection is an important part of quality control process in any glass manufacturing industry. The bottles must be inspected before packaging. Machine vision for glass bottle defect detection is the technology and method to inspect and analyze the defects for images automatically. Machine vision requires high-ability method to detect the defect and reject the bottle with the defect quickly. In this paper, defect detection framework for glass bottle defect detection tasks using autoencoders and error threshold is proposed. The fast detection method, a small autoencoder neural network architecture was designed with only good bottle images to train an autoencoder neural network. The decoded images are representations of normal bottle images and calculate threshold errors value. Defect detection is done by comparing the error between the normal background image and the encoded images to a threshold error from the training set. The performance of our method was compared to several other methods: VGG16, MobileNetV3, ADA, edge detection and image threshold. The experimental results show that our method yields 80% of accuracy on the body dataset and 92% of accuracy on the neck dataset. The average training time of our method is faster than that of all other neural network-based methods. From the experimental results, we can conclude that our defect detection framework outperforms other approaches both in accuracy and training time for defect detection on the side wall of a glass bottle.
Some of the metrics are blocked by your
consent settings
Improving OpenAI's Whisper Model for Transcribing Homophones in Legal News
(2024-01-01)
Siriket, Lattapon
;
Jitkajornwanich, Kulsawasd
;
Jaiyen, Saichon
;
Intakosum, Sarun
The 'Whisper' model provides a tool for those who require transcription of human voice. It equips with opensource features and diverse functionalities. The model is capable of effectively deciphering messages in multiple languages, including support for the Thai language. This paper focuses on improving the transcription process of Thai homophones using the Whisper model in reducing the word error rate (WER). We focus on words in the legal news category and identify factors that lead to Whisper's incorrect sound predictions. We examined homophones using snippets of legal news video clips and compiled them into a homophone dictionary. We compare words extracted from the Whisper model by determining the word error rate and spelling of words. Based on the initial results obtained from the original Whisper model and the created homophone dictionary, 48 % of the words were incorrectly transcribed out of a total of 94 words. Then, we propose a methodology by which the performance of the Whisper is improved. That way, the automatic speech recognition of Thai language using the Whisper model can fully be utilized and used in other applications.
Some of the metrics are blocked by your
consent settings
FACE SPOOFING DETECTION BASED ON DEEP FEATURE EXTRACTION AND INSTANCE-BASED CLASSIFICATION
(2023-02-01)
Claypo, Niphat
;
Jaiyen, Saichon
;
Hanskunatai, Anantaporn
Face recognition is an important task in smart home security for detecting a face or monitoring a person in a live video and verifying the identity of an authentic user. However, there have been spoofing face methods that can trick a face recognition algorithm into wrongly verifying the identity of the person. In this paper, we propose a new hybrid framework for spoofing face detection based on Convolutional Neural Network and Long Short-Term Memory (CNNLSTM) and instance-based learning algorithm. In addition, a new dataset called FSA-CCTV is proposed, which contains face images from CCTV video clips with many types of spoofing attacks. The performance of our method was compared to several other anti-spoofing methods: CNN and RI-LBP, SLRNN, HSV+YCbCr, ResNet50, YCbCr+SVM and YCbCr+KNN. The experimental results show that our method yielded 93.2% of Accuracy, 96.8% of Recall, 94% of Precision, 94.8% of F<inf>1</inf>-score and 0.93 of AUC on the FSA-CCTV dataset. From the experimental results we can conclude that the proposed algorithm outperforms other approaches and yielded the most stable classification accuracy on the proposed dataset.
Some of the metrics are blocked by your
consent settings
Question Classification for Thai Conversational Chatbots Using Artificial Neural Networks and Multilingual BERT Models
(2023-01-01)
Thananukhun, Kit
;
Jaiyen, Saichon
;
Jitkajornwanich, Kulsawasd
;
Hanskunatai, Anantaporn
Question-Answering (QA) models are part of Natural Language Processing (NLP) field used for ensuring questions match the answers appropriately. QA consists of several steps, one of which is called Question Classification, which is to classify the context of communication. In this step, it categorizes group of questions based on what users need to know in order to combine answers within the same category and respond accurately. It helps saving us time to search for answers as well. In this paper, we present a question classification model for Thai Conversational Chatbot using Artificial Neural Network and Multilingual Bidirectional Encoder Representations from Transformer (BERT) models using BERT-base multilingual cased combined with Multilayer Perceptron (MLP). The method yields the highest accuracy of 92.57%, compared to the BERT-base multilingual cased combined with other classification models, including Support Vector Machine (SVM), Naive Bayes (NB), K-Nearest Neighbors (KNN) and Decision Trees (DTs) with the accuracy scores of 88.57%, 80.00%, 78.57% and 60.29%, respectively. In addition, we also compare the performance of our proposed BERT model with another well-known Thai word embedding model, called Thai2Vec, which also combines with other classification models including MLP, SVM, NB, KNN and DTs, and their results of accuracies are: 85.71%, 85.71%, 75.71%, 75.71% and 58.86%, respectively. From the experiments, the BERT model combined with MLP can achieve the highest performance in term of accuracy among other methods.
Some of the metrics are blocked by your
consent settings
Inspection System for Glass Bottle Defect Classification based on Deep Neural Network
(2023-01-01)
Claypo, Niphat
;
Jaiyen, Saichon
;
Hanskunatai, Anantaporn
The problem of defects in glass bottles is a significant issue in glass bottle manufacturing. There are various types of defects that can occur, including cracks, scratches, and blisters. Detecting these defects is crucial for ensuring the quality of glass bottle production. The inspection system must be able to accurately detect and automatically determine that the defects in a bottle affect its appearance and functionality. Defective bottles must be identified and removed from the production line to maintain product quality. This paper proposed glass bottle defect classification using Convolutional Neural Network with Long Short-Term Memory (CNNLSTM) and instant base classification. CNNLSTM is used for feature extraction to create a representation of the class data. The instant base classification predicts anomalies based on the similarity of representations of class data. The convolutional layer of the CNNLSTM method incorporates a transfer learning algorithm, using pre-trained models such as ResNet50, AlexNet, MobileNetV3, and VGG16. In this experiment, the results were compared with ResNet50, AlexNet, MobileNetV3, VGG16, ADA, Image threshold, and Edge detection methods. The experimental results demonstrate the effectiveness of the proposed method, achieving high classification accuracies of 77% on the body dataset, 95% on the neck dataset, and an impressive 98% on the rotating dataset.
Some of the metrics are blocked by your
consent settings
A Deep Single-Pass Learning for Recognition of Handwritten Digits
(2022-03-01)
Thongsuwan, Setthanun
;
Jaiyen, Saichon
We describe a deep learning model-Deep Single-Pass Learning (DSPL)-that can learn a data set, with a single pass for recognition, and predict with high accuracy, when evaluated for visual recognition of handwritten digits. DSPL consists of several stacked convolutional layers to learn features automatically and Extreme gradient boosting (XGBoost) was set as the last layer for predicting class labels. The learning time complexity is O(Lc2 mnpq), or less than the learning time of deep learning-Convolutional Neural Networks (CNNs). The network does not need iteration to re-adjust weights during feature learning. Tests showed that our model provided better accuracy than other models i.e. CNNs, XGBoost, LR, ETC, GBC, RFC, GNB, and DTC, including MLP and SVC families: in the worst case, DSPL provided 99.95% accuracy.
Some of the metrics are blocked by your
consent settings
A Performance Comparison between GIS-based and Neuron Network Methods for Flood Susceptibility Assessment in Ayutthaya Province
(2022-01-15)
Vajeethaveesin, Thanat
;
Panboonyuen, Teerapong
;
Lawawironjwong, Siam
;
Srestasathiern, Panu
;
Jaiyen, Saichon
Flooding has been a long withstanding issue in Thailand. Due to its geographical setup, mitigation and management of floods are challenging and hard to execute. One of the tools used in managing the events is “flood susceptibility mapping,” in which an incident probability as well as a rescue path is estimated and planned. To create one, the traditional GIS method called FRAM (flood risk assessment model), combined with AHP (analytical hierarchy process), is used and implemented on ArcGIS software. In this method, we first created a comparison table to compute weights for each of the selected factors. Then the computed weights were used in the FRAM model in ArcGIS to create a flood susceptibility map for each region. Each region was then classified as very high, high, medium, low, and very low risk. On the other hand, in computer science, machine learning and AI are prevalent and being adopted to various domains, promising the effectiveness of the method, potentially beat the forementioned traditional method. Therefore, ANN (artificial neural network) is adopted in this work to create the flood susceptibility map. The ANN technique is developed by using causal factors. The ANN classifies areas as either flood areas or flood-free areas. The 2 methods from different disciplines (GIS and Computer Science) are applied and described in this paper with the intention to prove whether the machine learning is really efficient and can outperform the traditional GIS approach. Data on Thailand’ s Ayutthaya Province is used in this work as a case study-in order to assess flood prone areas and compared for performance evaluation. Both of which use the 6 selected factors according to the literature: (i) flow accumulation, (ii) elevation, (iii) land use, (iv) rainfall intensity, (v) slope and (vi) soil types. The results from the 2 methods were verified with historical flood data and compared. The results showed that ANN (obtained via sensitivity analysis) outperformed the FRAM with precision of 79.90 %, recall of 79.04 %, F1-score of 79.08 % and accuracy of 79.31 %. In addition, we found that (according to our ANN experiments) the main causal factors related to flood susceptibility map only included 3 factors: flow accumulation, elevation, and soil types. Therefore, the proposed methodology for assessment of flood susceptibility areas using these 3 factors could be considered sufficient and applied to other regions in related applications, when needed.
Some of the metrics are blocked by your
consent settings
ConvXGB: A new deep learning model for classification problems based on CNN and XGBoost
(2021-02-01)
Thongsuwan, Setthanun
;
Jaiyen, Saichon
;
Padcharoen, Anantachai
;
Agarwal, Praveen
We describe a new deep learning model - Convolutional eXtreme Gradient Boosting (ConvXGB) for classification problems based on convolutional neural nets and Chen et al.’s XGBoost. As well as image data, ConvXGB also supports the general classification problems, with a data preprocessing module. ConvXGB consists of several stacked convolutional layers to learn the features of the input and is able to learn features automatically, followed by XGBoost in the last layer for predicting the class labels. The ConvXGB model is simplified by reducing the number of parameters under appropriate conditions, since it is not necessary re-adjust the weight values in a back propagation cycle. Experiments on several data sets from UCL Repository, including images and general data sets, showed that our model handled the classification problems, for all the tested data sets, slightly better than CNN and XGBoost alone and was sometimes significantly better.
Some of the metrics are blocked by your
consent settings
Air Quality Assessment Based on Landsat 8 Images Using Supervised Machine Learning Techniques
(2020-06-01)
Vijaranakul, Nattadet
;
Jaiyen, Saichon
;
Srestasathiern, Panu
;
Lawawirojwong, Siam
;
Jitkajornwanich, Kulsawasd
Since 2018 during the winter of every year (December - January), Thailand has been suffering from air pollution problems known as PM 2.5 toxic dust, affecting people's daily lives especially in Bangkok and its metroplex. To cope with this problem, one of the traditional methods used is to implement physical air quality measurement devices at specific locations. Currently there are 21 stations across Bangkok and surrounding areas. Each station can assess air quality at the station point along with the given radius, meaning that areas far away from the station will not be assessed properly. In this paper, we propose a methodology that incorporates satellite images for air quality assessment with supervised machine learning techniques. Several classification models tested in this paper are Decision Tree, Naïve Bayes, k-Nearest Neighbors (kNN), Random Forest, and Gradient Boosting. From our experiments, the highest performance model is Random Forest that has averaged accuracy of 0.914, averaged precision of 0.89, averaged recall of 0.814 and averaged F-1 score of 0.84825.

KMITL

Browse

Filters

Settings

Sort By

Results per page

Search Results