2023 IEEE International Conference on Industry 4.0, Artificial Intelligence, and Communications Technology (IAICT)最新文献

{"title":"Distance Metrics Comparison on K-Nearest Neighbor for Landslide Susceptibility Mapping","authors":"Martinus Fajarbudi Kurnia Jati, O. Wahyunggoro, Agus Bejo, T. B. Adimedha","doi":"10.1109/IAICT59002.2023.10205840","DOIUrl":"https://doi.org/10.1109/IAICT59002.2023.10205840","url":null,"abstract":"Landslides are natural disasters that often occur in certain areas in Indonesia, which cause great damage to aspects of human life. To reduce or prevent major losses due to landslide disasters, previous research used the kind of study that is called Landslide Susceptibility Map (LSM). Meanwhile the purpose of this study is to compare 6 distance metrics on K-Nearest Neighbor (KNN) with distance weight. It is to find the optimal distance metric to create a landslide susceptibility map from the data which is located in the Kejajar sub-district, Wonosobo regency, Central Java, Indonesia, with 25 factors that cause landslides. Data processing is carried out using information gain as a feature selection which aims to reduce the factors that can cause landslide. The distance metrics that used in this study are: Euclidean, City Block, Chebyshev, Cosine, Hamming, and Jaccard. Every distance metrics were trained with the dataset and compare each performance model. The result is Chebyshev distance has a better performance than other distance metrics with precision 0.8289, recall 0.9763, f1-score 0.8966, and accuracy 0.9874.","PeriodicalId":339796,"journal":{"name":"2023 IEEE International Conference on Industry 4.0, Artificial Intelligence, and Communications Technology (IAICT)","volume":"117 1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126699729","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}

{"title":"Coordinate-Based Geometric Features and Nearest Neighbor Performance in 2D Facial Classification","authors":"E. Saputra, Risanuri Hidayat, Agus Bejo","doi":"10.1109/IAICT59002.2023.10205618","DOIUrl":"https://doi.org/10.1109/IAICT59002.2023.10205618","url":null,"abstract":"Geometric feature is one of many kinds of feature that is used in facial recognition system. This feature represents the relative position between facial objects. In recent days, those relative positions are represented by facial landmarks. Many extracted features from facial landmark had been used in facial recognition, except the coordinate-based feature. This paper transforms raw facial landmarks coordinate into coordinate-based geometric features. The raw two dimensions facial landmarks coming from 68-points facial landmark dataset. The feature extraction process utilizes two points of outer eye (OE), right OE is placed at (0,0) and the left OE is placed at (100, 0) using translation, rotation, and scaling matrices. There are 208 class in classification process. This process involves 208 subjects, where each subject has one file in probe image and nine files in gallery images. The k-NN (nearest neighbor) with k=1 is chosen as classification method. On test session, the use of 11 to 17 features is resulting high recognition accuracy (more than 90%) with 94.23% at the peak. Compared to other studies which use triangular-based feature, angular-based feature, and distance-based feature, coordinate-based feature gives better accuracy.","PeriodicalId":339796,"journal":{"name":"2023 IEEE International Conference on Industry 4.0, Artificial Intelligence, and Communications Technology (IAICT)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123439010","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}

{"title":"A Deep Learning-based General Defect Detection Framework for Automated Optical Inspection","authors":"Chia-Yu Lin, Yan-Hung Chou, Yun-Chiao Cheng","doi":"10.1109/IAICT59002.2023.10205799","DOIUrl":"https://doi.org/10.1109/IAICT59002.2023.10205799","url":null,"abstract":"Artificial intelligence (AI) is applied in automated optical inspection (AOI) to help inspect defects and reduce the false discovery rate of AOI in manufacturing industries. In current studies, the training data of AI models are sufficient, and the source data are from the specific production line. However, defect samples are insufficient, and the data source is variant. The current models need more generalization to all machines and take a long training time to build a new model for other appliances. This paper proposes a Deep Learning-based General Defect Detection Framework (DL-GDD) solves the insufficient data issue and the generalization issue of models. We implement a color preprocessing module, a data augmentation module, a data generation module, and four classification models to detect defects and generalize the utilization of DLG-DD. In experiments, we evaluate DLG-DD based on the NEU-CLS and AIdea datasets. The accuracy of DLG-DD is 90%, and the false omission rate and false discovery rate are less than 1%. DLG-DD is a general framework that tackles insufficient data and decreases the false discovery rate of AOI.","PeriodicalId":339796,"journal":{"name":"2023 IEEE International Conference on Industry 4.0, Artificial Intelligence, and Communications Technology (IAICT)","volume":"59 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123758856","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}

{"title":"Detection of Physical Adversarial Attacks on Traffic Signs for Autonomous Vehicles","authors":"B. Villarini, Panagiotis I. Radoglou-Grammatikis, T. Lagkas, P. Sarigiannidis, V. Argyriou","doi":"10.1109/IAICT59002.2023.10205591","DOIUrl":"https://doi.org/10.1109/IAICT59002.2023.10205591","url":null,"abstract":"Current vision-based detection models within Autonomous Vehicles, can be susceptible to changes within the physical environment, which cause unexpected issues. Physical attacks on traffic signs could be malicious or naturally occurring, causing incorrect identification of the traffic sign which can drastically alter the behaviour of the autonomous vehicle. We propose two novel deep learning architectures which can be used as detection and mitigation strategy for environmental attacks. The first is an autoencoder which detects anomalies within a given traffic sign, and the second is a reconstruction model which generates a clean traffic sign without any anomalies. As the anomaly detection model has been trained on normal images, any abnormalities will provide a high reconstruction error value, indicating an abnormal traffic sign. The reconstruction model is a Generative Adversarial Network (GAN) and consists of two networks; a generator and a discriminator. These map the input traffic sign image into a meta representation as the output. By using anomaly detection and reconstruction models as mitigation strategies, we show that the performance of the other models in pipelines such as traffic sign recognition models can be significantly improved. In order to evaluate our models, several types of attack circumstances were designed and on average, the anomaly detection model achieved 0.84 accuracy with a 0.82 F1-score in real datasets whereas the reconstruction model improved performance of traffic sign recognition model from average F1-score 0.41 to 0.641.","PeriodicalId":339796,"journal":{"name":"2023 IEEE International Conference on Industry 4.0, Artificial Intelligence, and Communications Technology (IAICT)","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133214051","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}

{"title":"Tree Boosting Methods Comparison for Landslide Susceptibility Maps, Case Study: Kejajar, Wonosobo","authors":"Rokhmat Arifianto, O. Wahyunggoro, I. Mustika, T. B. Adimedha","doi":"10.1109/IAICT59002.2023.10205882","DOIUrl":"https://doi.org/10.1109/IAICT59002.2023.10205882","url":null,"abstract":"Landslides in Indonesia occur almost every year and cause large material losses. Early prevention by creating a landslide susceptibility map is one way to anticipate losses due to landslides. The search for the best method for predicting landslides using machine learning with several tree boosting methods has been carried out, but the comparison between the tree boosting methods is unknown. This study aims to compare the tree boosting methods in their use for creating landslide susceptibility maps. The case study used in this research is Kejajar District, Wonosobo. There are 25 data features used to determine landslide. The landslide data in this study is 84 polygons. The tree boosting methods used include XGBoost, LGBM, Adaboost and Catboost. Hyperparameter tuning and k-fold cross validation were used to get the best model. The results of the comparison show that LGBM is the best method with accuracy, recall, f1 score, and ROC AUC values of 0.9903, 0.9360, 0.9154, and 0.9648 respectively. It indicates that the boosting method using LGBM can provide good results for creating a landslide susceptibility map.","PeriodicalId":339796,"journal":{"name":"2023 IEEE International Conference on Industry 4.0, Artificial Intelligence, and Communications Technology (IAICT)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133505177","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}

{"title":"MetaForecast: Harnessing Model-Agnostic Meta-Learning Approach to Predict Key Metrics of Interconnected Network Topologies","authors":"Shruti Jadon, Aryan Jadon","doi":"10.1109/IAICT59002.2023.10205730","DOIUrl":"https://doi.org/10.1109/IAICT59002.2023.10205730","url":null,"abstract":"Meta-learning, an approach in machine learning that focuses on “learning how to learn prioritizes generalization over specialization, mirroring the human’s ability to derive generalizations from experiences and specialize when tasks are repeated. Training a meta-model requires the procurement of similar tasks or similar data distribution. In our study, we explored a model-agnostic meta-learning approach to predict telemetry data collected from a network of devices. We also proposed a custom architecture “MetaForecast” wherein a meta-learner learns the generalized intricacies of each site/device’s data, allowing us to fine-tune the base learner and create site/device-specific models. Based on our experiments, we have observed that by using MetaForecast in such complex telemetry system we can:1)Significantly reducing the training time of a forecasting model for newly added devices/sites: Our proposed approach enables fine-tuning to a site-specific model within a small number (less than 10) of epochs.2)Minimizing data gathering requirements: By requiring fewer epochs for model tuning, our approach greatly reduces the data gathering needs. Hence, a new site doesn’t necessitate extensive historical data beyond a few recent entries based on granularity.3)Enabling day 1 prediction: We assert that if a new site/device is added, the new model can be trained within a few epochs and doesn’t rely on a large amount of past data for training. To establish a baseline, we compared the performance of our MAML-inspired architecture to individual models per site and transfer learning. Our findings revealed approximately a ~ 10% reduction in mean squared error, a ~ 50% reduction in computing resources, and a ~ 65% reduction in data gathering requirements. Based on our comprehensive research, we assert that the integration of meta-learning techniques and our proposed architecture yields notable improvements in forecasting accuracy, accompanied by substantial reductions in training time, data requirements, and computing resources.","PeriodicalId":339796,"journal":{"name":"2023 IEEE International Conference on Industry 4.0, Artificial Intelligence, and Communications Technology (IAICT)","volume":"173 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133607903","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}

{"title":"Joint Angular-Frequency Power Spectrum Sensing Based on K-Means Clustering","authors":"A. N. Jati, S. Wibowo, D. D. Ariananda","doi":"10.1109/IAICT59002.2023.10205879","DOIUrl":"https://doi.org/10.1109/IAICT59002.2023.10205879","url":null,"abstract":"The critical component in the cognitive radio (CR) network is spectrum sensing (SS) used to locate the accessible spectrum and to minimize interference with the licensed users (LUs). The CR SS could take into account the direction-of-arrival (DoA) of signals sent by LUs from the perspective of a secondary user (SU) to further improve the spectrum utility. This prevents the SU to interfere the LU signals by occupying a frequency band that the LU is currently using but aiming its transmission at a specific angle away from the LU. We use a uniform linear array, sample the signal received by each antenna, compute the correlation between samples, apply classical beamforming, and compute the discrete Fourier transform to reconstruct the complete two-dimensional angular-frequency power spectrum. In order to determine whether LUs or other SUs are currently occupying a certain frequency band or not, we use K-means clustering to compute the detection threshold. Numerical studies are carried out to demonstrate that the threshold based on the K-means clustering is applicable to detect signals from LUs based on the result of frequency and DoA power SS.","PeriodicalId":339796,"journal":{"name":"2023 IEEE International Conference on Industry 4.0, Artificial Intelligence, and Communications Technology (IAICT)","volume":"103 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130185300","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}

{"title":"ResUnet-GAN with Dynamic Memory for Mura Defect Detection","authors":"Chia-Yu Lin, Pin-Fan Lin, Wei-Kuang Chung, Yu-Hsien Lee","doi":"10.1109/IAICT59002.2023.10205662","DOIUrl":"https://doi.org/10.1109/IAICT59002.2023.10205662","url":null,"abstract":"“Mura” is a phenomenon in which panels have uneven display defects, irregular shapes, and different sizes. It is impossible to produce perfect panels on production lines, so panel inspection is necessary to differentiate between “light Mura” and “serious Mura” manually. The performance of conventional defect detection models for Mura detection is worse since they only differentiate between “normal” and “abnormal” samples. To reduce human cost and increase the accuracy of Mura detection, we propose a “ResUnet-GAN with Dynamic Memory Model an unsupervised anomaly detection method based on a Generative Adversarial Network (GAN) with a memory module to distinguish panel defects. In the dynamic memory, we designed a dynamic feature filtering (DFF) method to choose important features of images, enhancing the ability to recognize light Mura features of the ResUnet-GAN. The proposed model can achieve an Area Under Curve (AUC) of approximately 0.8 for accurate Mura detection. The mechanism of this paper is novel, and the result contributes to practical application.","PeriodicalId":339796,"journal":{"name":"2023 IEEE International Conference on Industry 4.0, Artificial Intelligence, and Communications Technology (IAICT)","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131801451","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}

{"title":"Increasing Automatic Proctoring System Performance using Distributed Round-Robin Load Balancer","authors":"M. A. Nugroho, M. Abdurohman, B. Erfianto","doi":"10.1109/IAICT59002.2023.10205750","DOIUrl":"https://doi.org/10.1109/IAICT59002.2023.10205750","url":null,"abstract":"In the context of online examinations, ensuring a reliable and efficient proctoring system is crucial for maintaining academic integrity. The increasing demand for automatic proctoring services raises challenges related to handling growing traffic, server failures, and ensuring fair resource utilization. This study investigates the performance of an automatic proctoring load balancing system to address these challenges, focusing on key metrics such as response times, request distribution, and failover times. The system utilizes a Round Robin load balancing algorithm to effectively manage incoming requests and distribute them among multiple proctoring engines. The evaluation reveals a 22.22% decrease in response times, ensuring low latency and improved efficiency. Additionally, the algorithm evenly distributes requests across the proctoring engines, enabling fair resource utilization and adaptability to varying workloads. Despite an increase in failover times with the number of requests, the system maintains an average failover time of 700 ms, demonstrating resilience in handling server failures and ensuring uninterrupted service. Overall, the automatic proctoring load balancing system offers a more efficient, scalable, and resilient proctoring infrastructure capable of addressing the challenges of increasing traffic and server failures.","PeriodicalId":339796,"journal":{"name":"2023 IEEE International Conference on Industry 4.0, Artificial Intelligence, and Communications Technology (IAICT)","volume":"58 29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128671955","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}

{"title":"Performance Evaluation of SDN-Enabled Switching System For IoT Infrastructure","authors":"M. Elham, S. Sam, A. Azizan, Y. Yusof, Norliza Mohamed, N. Ahmad","doi":"10.1109/IAICT59002.2023.10205953","DOIUrl":"https://doi.org/10.1109/IAICT59002.2023.10205953","url":null,"abstract":"The revolution of the Internet of Things (IoT) has had a major impact on the infrastructure of networks and information technology. Software Defined Networking (SDN) is a way to enhance infrastructure agility and thus facilitate dynamic and scalable design, delivery and operation of network services. It is therefore believed that the agility brought about by network programming or SDN is essential to address the IoT revolution which is pushing the infrastructure to its limit with numerous and diverse requirements to meet. IoT nodes usually depend on the network layers to communicate with each other. However, it is almost impossible to construct a single end-to-end infrastructure that addresses the whole set of IoT constraints. The objective of this project is thus to share components in a converged layer where Raspberry acts as an IoT gateway as well as SDN enabled device that can satisfy various IoT requirements. Raspberry Pi has been a common IoT commodity for quite a few years now and is targeted to be redesigned for additional networking functionality. The most recent version # 4 of Raspberry Pi has been installed with a very well-known virtual Open vSwitch. The Raspberry Pi serves as the data path of network communication throughout the project with SDN controller Ryu operating on the Raspberry Pi to provide programmability. A system of wired, wireless and SDN networks combined is then designed to validate the operations of the Raspberry Pi. After successful functionality, the Raspberry Pi is then evaluated in terms of QoS parameters (bandwidth, packet loss, delay and jitter) to ensure the suitability of Raspberry Pi utilized as a common device for both IoT and SDN operations.","PeriodicalId":339796,"journal":{"name":"2023 IEEE International Conference on Industry 4.0, Artificial Intelligence, and Communications Technology (IAICT)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133104413","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}