Pub Date : 2024-02-10DOI: 10.37936/ecti-cit.2024181.254621
Paisit Khanarsa, Satanat Kitsiranuwat
Cervical cancer screening allows the early signs of precancerous abnormalities in the cervix before they develop into invasive cancer. The Pap Smear is a widely used screening for early detection and prevention of cervical cancer. In many remote areas, the number of cytologists available to interpret pap smear screening tests is insufficient. This lack of personnel makes the test interpretation very time-consuming. To address this, deep learning techniques have been employed to detect cervical cancer cells and support cytologists. Therefore, an integrative approach with deep learning models and the ensemble techniques such as the maximum occurrence and the maximum probability score of cervical cells was proposed. The multi-cell assessment of the Pap smear slide allowed aggregate predictions of single cervical cell images using the proposed method. The classification results between pre-trained deep learning models and the proposed method were compared. In the experimental results, the proposed method can achieve an accuracy score of more than 97%, while the best pre-trained deep learning model can attain an accuracy score of more than 85%. Hence, the proposed method may have the potential to assist physicians or cytologists in the classification of cervical cell types for Pap Smear images.
{"title":"Deep Learning-based Ensemble Approach for Conventional Pap Smear Image Classification","authors":"Paisit Khanarsa, Satanat Kitsiranuwat","doi":"10.37936/ecti-cit.2024181.254621","DOIUrl":"https://doi.org/10.37936/ecti-cit.2024181.254621","url":null,"abstract":"Cervical cancer screening allows the early signs of precancerous abnormalities in the cervix before they develop into invasive cancer. The Pap Smear is a widely used screening for early detection and prevention of cervical cancer. In many remote areas, the number of cytologists available to interpret pap smear screening tests is insufficient. This lack of personnel makes the test interpretation very time-consuming. To address this, deep learning techniques have been employed to detect cervical cancer cells and support cytologists. Therefore, an integrative approach with deep learning models and the ensemble techniques such as the maximum occurrence and the maximum probability score of cervical cells was proposed. The multi-cell assessment of the Pap smear slide allowed aggregate predictions of single cervical cell images using the proposed method. The classification results between pre-trained deep learning models and the proposed method were compared. In the experimental results, the proposed method can achieve an accuracy score of more than 97%, while the best pre-trained deep learning model can attain an accuracy score of more than 85%. Hence, the proposed method may have the potential to assist physicians or cytologists in the classification of cervical cell types for Pap Smear images.","PeriodicalId":507234,"journal":{"name":"ECTI Transactions on Computer and Information Technology (ECTI-CIT)","volume":" 12","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-02-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139786807","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-01-20DOI: 10.37936/ecti-cit.2024181.252738
Supakan Janthong, Rakkrit Duangsoithong, K. Chalermyanont
Non-technical loss (NTL) is one of the problems that has been a major issue in lost revenue for many years. Electricity distributors have attempted to reduce NTL by detecting electricity theft using various methods. Some events are difficult to detect that conventional meters inspection is inadequate. Moreover, many anomaly patterns found are very complex, confusing in identifying or distinguishing what types of electricity customers are at abnormal risk or energy theft that affects NTL. This paper proposes five key feature extraction methods and six classifying electricity customers using supervised learning. The main problem was studied and collected information, including kilowatt meters, electronic meters, TOU meters, and AMR meters, which cover four customer types that were recorded in the Provincial Electricity Authority (PEA) of Thailand. An electrical profile to be extracted for in-depth analysis of the behavior of each type of electricity customer, combined with the information of physical data to help enhance and increase efficiency. All features examined the relationships in each feature using Pearson correlation and handled unbalanced data using random oversampling (ROS). Then, the extracted data has been trained, validated, and tested to classify three classes: normal, risk, and theft, where we evaluate the results with performance metrics. The results show that random forest (RF) outperforms the rest of the classifiers by achieving a precision-recall area under the curve of 90% and a receiver operating characteristic curve of 78%. Significantly, the results were compared to previous studies and benchmark datasets, which revealed that the proposed method gave better results than other techniques.
非技术性损失（NTL）是多年来收入损失的主要问题之一。配电公司试图通过各种方法检测窃电行为来减少 NTL。有些事件很难被发现，传统的电表检查是不够的。此外，发现的许多异常模式非常复杂，在识别或区分哪些类型的电力客户存在异常风险或影响 NTL 的窃电行为时容易混淆。本文提出了五种关键特征提取方法和六种利用监督学习对电力客户进行分类的方法。主要问题是研究和收集信息，包括千瓦表、电子表、TOU 表和 AMR 表，涵盖泰国省电力局（PEA）记录的四种客户类型。将提取的电力概况与物理数据信息相结合，对各类电力客户的行为进行深入分析，以帮助增强和提高效率。所有特征均使用皮尔逊相关性检验每个特征中的关系，并使用随机超采样（ROS）处理不平衡数据。然后，对提取的数据进行训练、验证和测试，将其分为三个等级：正常、风险和盗窃，并用性能指标对结果进行评估。结果表明，随机森林（RF）的精度-召回曲线下面积达到 90%，接收者操作特征曲线达到 78%，优于其他分类器。值得注意的是，将结果与以前的研究和基准数据集进行比较后发现，所提出的方法比其他技术取得了更好的结果。
{"title":"Feature Extraction of Risk Group and Electricity Theft by using Electrical Profiles and Physical Data for Classification in the Power Utilities","authors":"Supakan Janthong, Rakkrit Duangsoithong, K. Chalermyanont","doi":"10.37936/ecti-cit.2024181.252738","DOIUrl":"https://doi.org/10.37936/ecti-cit.2024181.252738","url":null,"abstract":"Non-technical loss (NTL) is one of the problems that has been a major issue in lost revenue for many years. Electricity distributors have attempted to reduce NTL by detecting electricity theft using various methods. Some events are difficult to detect that conventional meters inspection is inadequate. Moreover, many anomaly patterns found are very complex, confusing in identifying or distinguishing what types of electricity customers are at abnormal risk or energy theft that affects NTL. This paper proposes five key feature extraction methods and six classifying electricity customers using supervised learning. The main problem was studied and collected information, including kilowatt meters, electronic meters, TOU meters, and AMR meters, which cover four customer types that were recorded in the Provincial Electricity Authority (PEA) of Thailand. An electrical profile to be extracted for in-depth analysis of the behavior of each type of electricity customer, combined with the information of physical data to help enhance and increase efficiency. All features examined the relationships in each feature using Pearson correlation and handled unbalanced data using random oversampling (ROS). Then, the extracted data has been trained, validated, and tested to classify three classes: normal, risk, and theft, where we evaluate the results with performance metrics. The results show that random forest (RF) outperforms the rest of the classifiers by achieving a precision-recall area under the curve of 90% and a receiver operating characteristic curve of 78%. Significantly, the results were compared to previous studies and benchmark datasets, which revealed that the proposed method gave better results than other techniques.","PeriodicalId":507234,"journal":{"name":"ECTI Transactions on Computer and Information Technology (ECTI-CIT)","volume":"32 4","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140501961","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-01-20DOI: 10.37936/ecti-cit.2024181.253800
J. Panyavaraporn, Paramate Horkaew
One of the essential processes of construction quality control is tile bonding inspection. Hollows beneath tile tessellation can lead to unbounded or completely broken tiles. An interior inspector typically used a hollowsounding technique. However, it relies on skill and judgment that greatly vary among individuals. Moreover, equipment and interpretation are difficult to calibrate and standardize. This paper addresses these issues by employing machine-learning strategies for tile-tapping sound classification. Provided that a tapping signal was digitally acquired, the proposed method was fully computerized. Firstly, the signal was analyzed and its wavelets and MFCC were extracted. The corresponding spectral features were then classified by SVM, k-NN, Naïve Bayes, and Logistic Regression algorithm, in turn. The results were subsequently compared against those from a previous works that employed a deep learning strategy. It was revealed that when the proposed method was properly configured, it required much less computing resources than the deep learning based one, while being able to distinguish dull from hollow sounding tiles with 93.67% accuracy.
{"title":"Non-Destructive Inspection of Tile Debonding by DWT and MFCC of Tile-Tapping Sound with Machine versus Deep Learning Models","authors":"J. Panyavaraporn, Paramate Horkaew","doi":"10.37936/ecti-cit.2024181.253800","DOIUrl":"https://doi.org/10.37936/ecti-cit.2024181.253800","url":null,"abstract":"\u0000\u0000\u0000One of the essential processes of construction quality control is tile bonding inspection. Hollows beneath tile tessellation can lead to unbounded or completely broken tiles. An interior inspector typically used a hollowsounding technique. However, it relies on skill and judgment that greatly vary among individuals. Moreover, equipment and interpretation are difficult to calibrate and standardize. This paper addresses these issues by employing machine-learning strategies for tile-tapping sound classification. Provided that a tapping signal was digitally acquired, the proposed method was fully computerized. Firstly, the signal was analyzed and its wavelets and MFCC were extracted. The corresponding spectral features were then classified by SVM, k-NN, Naïve Bayes, and Logistic Regression algorithm, in turn. The results were subsequently compared against those from a previous works that employed a deep learning strategy. It was revealed that when the proposed method was properly configured, it required much less computing resources than the deep learning based one, while being able to distinguish dull from hollow sounding tiles with 93.67% accuracy.\u0000\u0000\u0000","PeriodicalId":507234,"journal":{"name":"ECTI Transactions on Computer and Information Technology (ECTI-CIT)","volume":"68 20","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139611546","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-01-20DOI: 10.37936/ecti-cit.2024181.253577
Allyza Ruth Cadeliña, Mori Assanty Cuevas, Marselean Kallos, Morris John Montemayor, Ronald Jay Uy, Joshua Benjamin Rodriguez, Orland D. Tubola
Quantum computing shows a positive approach for addressing optimization challenges in NP-hard problems such as the vehicle routing problem (VRP). This study focuses on improving the efficiency of disaster response operations by localizing the application of D-wave quantum annealing in Marikina City. This study uses the Solution Partitioning Solver (SPS) and the Quadratic Unconstrained Binary Optimization (QUBO) formulation to convert the VRP into an equation that can be solved using quantum annealing. The study demonstrates that quantum computing effectively distributes resources during emergency response operations and improves overall operational efficiency. In determining the most effective route for each vehicle, the D-wave Leap API and QUBO representation compute the distances traveled by each vehicle. These findings contribute to the practical applications of quantum computing to revolutionize various fields, including disaster management. Implementing D-wave quantum annealing in Marikina City shows relevance for future advancements in optimizing resource allocation and improving disaster response operations.
量子计算为解决车辆路由问题（VRP）等 NP 难问题的优化挑战提供了一种积极的方法。本研究的重点是通过在马里基纳市本地化应用 D 波量子退火来提高救灾行动的效率。本研究使用解决方案分区求解器（SPS）和二次无约束二元优化（QUBO）公式将 VRP 转换为可使用量子退火求解的方程。研究表明，量子计算能在应急响应行动中有效分配资源，提高整体运行效率。在确定每辆车的最有效路线时，D-wave Leap API 和 QUBO 表示法计算了每辆车的行驶距离。这些研究成果有助于量子计算的实际应用，为包括灾难管理在内的各个领域带来变革。在马里基纳市实施 D 波量子退火表明，未来在优化资源分配和改善灾难响应行动方面的进步具有现实意义。
{"title":"D-Wave Implementation of Quantum Annealing for Optimal Resource Allocation in Disaster Response Operation of Marikina City","authors":"Allyza Ruth Cadeliña, Mori Assanty Cuevas, Marselean Kallos, Morris John Montemayor, Ronald Jay Uy, Joshua Benjamin Rodriguez, Orland D. Tubola","doi":"10.37936/ecti-cit.2024181.253577","DOIUrl":"https://doi.org/10.37936/ecti-cit.2024181.253577","url":null,"abstract":"Quantum computing shows a positive approach for addressing optimization challenges in NP-hard problems such as the vehicle routing problem (VRP). This study focuses on improving the efficiency of disaster response operations by localizing the application of D-wave quantum annealing in Marikina City. This study uses the Solution Partitioning Solver (SPS) and the Quadratic Unconstrained Binary Optimization (QUBO) formulation to convert the VRP into an equation that can be solved using quantum annealing. The study demonstrates that quantum computing effectively distributes resources during emergency response operations and improves overall operational efficiency. In determining the most effective route for each vehicle, the D-wave Leap API and QUBO representation compute the distances traveled by each vehicle. These findings contribute to the practical applications of quantum computing to revolutionize various fields, including disaster management. Implementing D-wave quantum annealing in Marikina City shows relevance for future advancements in optimizing resource allocation and improving disaster response operations.","PeriodicalId":507234,"journal":{"name":"ECTI Transactions on Computer and Information Technology (ECTI-CIT)","volume":"62 17","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139611563","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-01-20DOI: 10.37936/ecti-cit.2024181.253483
Surapon Riyana, Kittikorn Sasujit, N. Homdoung
A well-known privacy preservation model is k-anonymity. It is simple and widely applied in several real-life systems. To achieve k-anonymity constraints in datasets, all explicit identifiers of users are removed. Furthermore, the unique quasi-identifiers of users are distorted by their less specific values to be at least k indistinguishable tuples. For this reason, after datasets are satisfied by k-anonymity constraints, they can guarantee that all possible query conditions to them always have at least k tuples that are satisfied. Aside from achieving privacy preservation constraints, the data utility and the complexity of data transformation are serious issues that must also be considered when datasets are released. Therefore, both privacy preservation models are proposed in this work. They are based on k-anonymity constraints in conjunction with the weighted graph of correlated distortion tuples and the adjacency matrix of tuple distances. The proposed models aim to preserve data privacy in datasets. Moreover, the data utility and data transform complexities are also considered in the privacy preservation constraint of the proposed models. Furthermore, we show that the proposed data transformation technique is more efficient and effective by using extensive experiments.
众所周知的隐私保护模式是 k 匿名。它操作简单，在现实生活中的多个系统中得到广泛应用。为了在数据集中实现 k-anonymity 约束，用户的所有显式标识符都会被移除。此外，用户的唯一准标识符会被其不太具体的值扭曲，使其成为至少 k 个无法区分的图元。因此，数据集在满足 k 个匿名约束后，可以保证所有可能的查询条件总是至少有 k 个图元得到满足。除了实现隐私保护约束外，数据效用和数据转换的复杂性也是数据集发布时必须考虑的严重问题。因此，本文提出了两种隐私保护模式。它们都基于 k-anonymity 约束，并结合了相关失真元组的加权图和元组距离的邻接矩阵。提出的模型旨在保护数据集中的数据隐私。此外，拟议模型的隐私保护约束还考虑了数据效用和数据转换复杂性。此外，我们还通过大量实验证明，所提出的数据转换技术更加高效和有效。
{"title":"Achieving Privacy Preservation Constraints based on K-Anonymity in conjunction with Adjacency Matrix and Weighted Graphs","authors":"Surapon Riyana, Kittikorn Sasujit, N. Homdoung","doi":"10.37936/ecti-cit.2024181.253483","DOIUrl":"https://doi.org/10.37936/ecti-cit.2024181.253483","url":null,"abstract":"A well-known privacy preservation model is k-anonymity. It is simple and widely applied in several real-life systems. To achieve k-anonymity constraints in datasets, all explicit identifiers of users are removed. Furthermore, the unique quasi-identifiers of users are distorted by their less specific values to be at least k indistinguishable tuples. For this reason, after datasets are satisfied by k-anonymity constraints, they can guarantee that all possible query conditions to them always have at least k tuples that are satisfied. Aside from achieving privacy preservation constraints, the data utility and the complexity of data transformation are serious issues that must also be considered when datasets are released. Therefore, both privacy preservation models are proposed in this work. They are based on k-anonymity constraints in conjunction with the weighted graph of correlated distortion tuples and the adjacency matrix of tuple distances. The proposed models aim to preserve data privacy in datasets. Moreover, the data utility and data transform complexities are also considered in the privacy preservation constraint of the proposed models. Furthermore, we show that the proposed data transformation technique is more efficient and effective by using extensive experiments.","PeriodicalId":507234,"journal":{"name":"ECTI Transactions on Computer and Information Technology (ECTI-CIT)","volume":"136 3","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140501931","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This research aims to develop an IoT-driven smart farming system for beef cattle management in Chiang Rai Province, Thailand. The system empowers small-scale farmers by enabling precise criteria for cattle care, optimized feeding, growth monitoring, breeding analysis, and cost estimation through WSN and cloud-based platforms. The sensors gather raw data on consumption from the feeding troughs and then transmit it to the cloud-based platform. Consumption data is then analyzed using Linear Regression Analysis. Key findings indicate a substantial correlation (0.995) between feed quantity and cattle weight gain, with a predictive capability of 99%. This system enhances precision and decision-making in cattle farming, offering significant benefits to small-scale farmers in the region.
{"title":"Enhancing Smart Farming Capabilities for Small-Scale Cattle Farmers in Chiang Rai, Thailand","authors":"Bunyarat Umsura, Kamonlak Chaidee, Kingkan Puansurin, Dueanpen Manoruang, Pornthipat Wimooktayone, Kanjana Boontasri, Wisoot Kaenmueng","doi":"10.37936/ecti-cit.2024181.253823","DOIUrl":"https://doi.org/10.37936/ecti-cit.2024181.253823","url":null,"abstract":"This research aims to develop an IoT-driven smart farming system for beef cattle management in Chiang Rai Province, Thailand. The system empowers small-scale farmers by enabling precise criteria for cattle care, optimized feeding, growth monitoring, breeding analysis, and cost estimation through WSN and cloud-based platforms. The sensors gather raw data on consumption from the feeding troughs and then transmit it to the cloud-based platform. Consumption data is then analyzed using Linear Regression Analysis. Key findings indicate a substantial correlation (0.995) between feed quantity and cattle weight gain, with a predictive capability of 99%. This system enhances precision and decision-making in cattle farming, offering significant benefits to small-scale farmers in the region.","PeriodicalId":507234,"journal":{"name":"ECTI Transactions on Computer and Information Technology (ECTI-CIT)","volume":" 29","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139623644","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：