• Proceedings of the Korean Society of Computer Information Conference
• /
• 2022.01a
• /
• pp.3-5
• /
• 2022

본 연구는 VGGNet을 사용하여 광섬유 복합가공 지선 설비의 양/불량 판별을 수행한다. 광섬유 복합가공 지선이란, 전력선의 보호 및 전력 시설 간 통신을 담당하는 중요 설비로 고장 발생 전, 결함의 조기 발견 및 유지 관리가 중요하다. 현재 한국전력공사에서는 드론에서 촬영된 영상을 점검원이 이상 여부를 점검하는 방식이 주로 사용되고 있으나 이는 점검원의 숙련도, 경험에 따른 정확성 및 비용과 시간 측면에서 한계를 지니고 있다. 본 연구는 드론에서 촬영된 영상으로 VGGNet 기반의 양/불량 판정을 수행했다. 그 결과, 정확도 약 95.15%, 정밀도 약 96%, 재현율 약 95%, f1 score 약 95%의 성능을 확인하였다. 결과 확인 방법으로는 설명 가능한 인공지능(XAI) 알고리즘 중 하나인 Grad-CAM을 적용하였다. 이러한 광섬유 복합가공 지선 설비의 양/불량 판별은 점검원의 단순 작업에 대한 비용 및 점검 시간을 줄이며, 부가가치가 높은 업무에 집중할 수 있게 해준다. 또한, 고장 결함 발견에 있어서 객관적인 점검을 수행하기 때문에 일정한 점검 품질을 유지한다는 점에서 적용 가치가 있다.

• Annual Conference on Human and Language Technology
• /
• 2023.10a
• /
• pp.563-567
• /
• 2023

본 연구는 한국 통계청이 수행하는 가계동향조사와 생활시간조사에서 자료처리 과정 및 방법을 혁신하려는 시도로, 기존의 통계 생산 방법론의 한계를 극복하고, 대규모 데이터의 효과적인 관리와 분석을 가능하게 하는 인공지능 기반의 통계 생산을 목표로 한다. 본 연구는 데이터 과학과 통계학의 교차점에서 진행되며, 인공지능 기술, 특히 자연어 처리와 딥러닝을 활용하여 비정형 텍스트 분류 방법의 성능을 검증하며, 인공지능 기반 통계분류 방법론의 확장성과 추가적인 조사 확대 적용의 가능성을 탐구한다. 이 연구의 결과는 통계 데이터의 품질 향상과 신뢰성 증가에 기여하며, 국민의 생활 패턴과 행동에 대한 더 깊고 정확한 이해를 제공한다.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.14 no.2
• /
• pp.115-121
• /
• 2021

This paper used big data and artificial intelligence technology to predict the rapidly increasing internet traffic. There have been various studies on traffic prediction in the past, but they have not been able to reflect the increasing factors that induce huge Internet traffic such as smartphones and streaming in recent years. In addition, event-like factors such as the release of large-capacity popular games or the provision of new contents by OTT (Over the Top) operators are more difficult to predict in advance. Due to these characteristics, it was impossible for an ISP (Internet Service Provider) to reflect real-time service quality management or traffic forecasts in the network business environment with the existing method. Therefore, in this study, in order to solve this problem, an Internet traffic collection system was constructed that searches, discriminates and collects traffic data in real time, separate from the existing NMS. Through this, the flexibility and elasticity to automatically register the data of the collection target are secured, and real-time network quality monitoring is possible. In addition, a large amount of traffic data collected from the system was analyzed by machine learning (AI) to predict future traffic of OTT operators. Through this, more scientific and systematic prediction was possible, and in addition, it was possible to optimize the interworking between ISP operators and to secure the quality of large-scale OTT services.

• Journal of Digital Contents Society
• /
• v.19 no.4
• /
• pp.789-799
• /
• 2018

This article presents machine learning based approach on Big data to analyzing time series data for anomaly detection in such industrial complex system. Long Short-Term Memory (LSTM) network have been demonstrated to be improved version of RNN and have become a useful aid for many tasks. This LSTM based model learn the higher level temporal features as well as temporal pattern, then such predictor is used to prediction stage to estimate future data. The prediction error is the difference between predicted output made by predictor and actual in-coming values. An error-distribution estimation model is built using a Gaussian distribution to calculate the anomaly in the score of the observation. In this manner, we move from the concept of a single anomaly to the idea of the collective anomaly. This work can assist the monitoring and management of Smart Factory in minimizing failure and improving manufacturing quality.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.10
• /
• pp.391-398
• /
• 2021

Since the leased line is a structure that exclusively uses two connected areas for data transmission, a stable quality level and security are ensured, and despite the rapid increase in the number of switched lines, it is a line method that is continuously used a lot in companies. However, because the cost is relatively high, one of the important roles of the network operator in the enterprise is to maintain the optimal state by properly arranging and utilizing the resources of the network leased line. In other words, in order to properly support business service requirements, it is essential to properly manage bandwidth resources of leased lines from the viewpoint of data transmission, and properly predicting and managing leased line usage becomes a key factor. Therefore, in this study, various prediction models were applied and performance was evaluated based on the actual usage rate data of leased lines used in corporate networks. In general, the performance of each prediction was measured and compared by applying the smoothing model and ARIMA model, which are widely used as statistical methods, and the representative models of deep learning based on artificial neural networks, which are being studied a lot these days. In addition, based on the experimental results, we proposed the items to be considered in order for each model to achieve good performance for prediction from the viewpoint of effective operation of leased line resources.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.17 no.4
• /
• pp.577-586
• /
• 2022

Detection and classification of steel surface defects are critical for product quality control in the steel industry. However, due to its low accuracy and slow speed, the traditional approach cannot be effectively used in a production line. The current, widely used algorithm (based on deep learning) has an accuracy problem, and there are still rooms for development. This paper proposes a method of steel surface defect detection combining EfficientNetV2 for image classification and YOLOv5 as an object detector. Shorter training time and high accuracy are advantages of this model. Firstly, the image input into EfficientNetV2 model classifies defect classes and predicts probability of having defects. If the probability of having a defect is less than 0.25, the algorithm directly recognizes that the sample has no defects. Otherwise, the samples are further input into YOLOv5 to accomplish the defect detection process on the metal surface. Experiments show that proposed model has good performance on the NEU dataset with an accuracy of 98.3%. Simultaneously, the average training speed is shorter than other models.

• Journal of Internet Computing and Services
• /
• v.21 no.6
• /
• pp.23-31
• /
• 2020

This study was carried out to generate various images of railroad surfaces with random defects as training data to be better at the detection of defects. Defects on the surface of railroads are caused by various factors such as friction between track binding devices and adjacent tracks and can cause accidents such as broken rails, so railroad maintenance for defects is necessary. Therefore, various researches on defect detection and inspection using image processing or machine learning on railway surface images have been conducted to automate railroad inspection and to reduce railroad maintenance costs. In general, the performance of the image processing analysis method and machine learning technology is affected by the quantity and quality of data. For this reason, some researches require specific devices or vehicles to acquire images of the track surface at regular intervals to obtain a database of various railway surface images. On the contrary, in this study, in order to reduce and improve the operating cost of image acquisition, we constructed the 'Defective Railroad Surface Regeneration Model' by applying the methods presented in the related studies of the Generative Adversarial Network (GAN). Thus, we aimed to detect defects on railroad surface even without a dedicated database. This constructed model is designed to learn to generate the railroad surface combining the different railroad surface textures and the original surface, considering the ground truth of the railroad defects. The generated images of the railroad surface were used as training data in defect detection network, which is based on Fully Convolutional Network (FCN). To validate its performance, we clustered and divided the railroad data into three subsets, one subset as original railroad texture images and the remaining two subsets as another railroad surface texture images. In the first experiment, we used only original texture images for training sets in the defect detection model. And in the second experiment, we trained the generated images that were generated by combining the original images with a few railroad textures of the other images. Each defect detection model was evaluated in terms of 'intersection of union(IoU)' and F1-score measures with ground truths. As a result, the scores increased by about 10~15% when the generated images were used, compared to the case that only the original images were used. This proves that it is possible to detect defects by using the existing data and a few different texture images, even for the railroad surface images in which dedicated training database is not constructed.

• Journal of the Korean earth science society
• /
• v.31 no.4
• /
• pp.392-402
• /
• 2010

The purpose of this study is to analyze and evaluate the self-directed learning contents of Earth science area in the basic course of the 7th grade. For this purpose, we applied the 'Cyber Home Study Content Quality Control Tool' presented in 'Elementary Secondary Education e-Learning Quality Management Guidelines (Ver.2.0)' of Korea Education & Research Information Service (2008). The results of contents analysis are as follow: First, it was presented that the study guide introduced the contents which should be studied for one class, properly. And it was not analyzed that the diagnosis assesment was not completed in the initiative study; Second, it was possible to study choosing the contents fitting the learner's level of learning in the main study, it was comprised of about 15 minutes. Third, it was performed without feedback for incorrect answers in the learning assessment, just the number of wrong questions. And the learning arrangement present the important contents learned in that class, summarizing and arranging again. The results of content evaluation are as follows: First, a big difference was not showed against the needs analysis, instructional design, interaction in each class. And the evaluation of the ethics was not included a word or sentence not suitable. The evaluation of copyright, it was analyzed that Work within the content display in compliance with international copyright Second, the evaluation of instructional design presented mainly the description of a simple picture based, the visible resources like flash card were poor. And in the evaluation of Supporting System, it was presented that the contents were installed so that it was freely available for learners. But it was analyzed that there was no memo-function learners were able to jot down something during the studying contents. And in the evaluation for evaluation, the clear valuation basis about the described content was not presented. So there were slightly differences for each class. Third, in the evaluation and analysis for learning content, it was presented that there were some big differences for each class because it was not composed of the latest information, not corrected and complementary.

• Journal of Science Education
• /
• v.34 no.2
• /
• pp.225-236
• /
• 2010

The purpose of this study is to analyze and evaluate the Cyber Home Learning contents of Earth science area in the basic course of the $10^{th}$ grade. For this purpose, we applied the 'Cyber Home Study Content Quality Control Tool' presented in Elementary Secondary Education e-Learning Quality Management Guidelines (Ver.2.0)' of Korea Education & Research Information Service(2008). The results of Cyber home learning contents analysis are as follow: First, it was presented that the study guide introduced the contents which should be studied for one class, properly. And it was not analyzed that the diagnosis assesment was not completed in the initiative study; Second, it was possible to study choosing the contents fitting the learner's level of learning in the main study, it was comprised of about 10 minutes. Third, it was performed without feedback for incorrect answers in the learning assessment, just the number of wrong questions. And the learning arrangement present the important contents learned in that class, summarizing and arranging again. The results of evaluating the contents in Cyber Home Learning are as follows: First, in evaluation section of instructional design, many text materials which were so difficult for learners to read were explained, being provided. Besides, the systematic structures leaves much to be desired, in view of learners' learning experience, contents, and environment. And in evaluation section of learning contents, the error of contents caused the learning contents not to appear, the amount of learning in each section was found too much. Second, in evaluation section of the strategy for Teaching and Learning, when we mention the strategy of Self Directed Learning, the environment to make learners search for information free and self-study possible was not possessed well. And in evaluation section of interaction, it was found that a simple click caused the learning to go on. Third, in evaluation section of evaluating, it was evaluated that there was wanting in consistency in learning aims, contents, evaluation contents.

• Research in Plant Disease
• /
• v.28 no.3
• /
• pp.113-121
• /
• 2022

Rice blast is a major plant disease that occurs worldwide and significantly reduces rice yields. Rice blast disease occurs periodically in Korea, causing significant socio-economic damage due to the unique status of rice as a major staple crop. A disease outbreak prediction system is required for preventing rice blast disease. Epidemiological investigations of disease outbreaks can aid in decision-making for plant disease management. Currently, plant disease prediction and epidemiological investigations are mainly based on quantitatively measurable, structured data such as crop growth and damage, weather, and other environmental factors. On the other hand, text data related to the occurrence of plant diseases are accumulated along with the structured data. However, epidemiological investigations using these unstructured data have not been conducted. The useful information extracted using unstructured data can be used for more effective plant disease management. This study analyzed news articles related to the rice blast disease through text mining to investigate the years and provinces where rice blast disease occurred most in Korea. Moreover, the average temperature, total precipitation, sunshine hours, and supplied rice varieties in the regions were also analyzed. Through these data, it was estimated that the primary causes of the nationwide outbreak in 2020 and the major outbreak in Jeonbuk region in 2021 were meteorological factors. These results obtained through text mining can be combined with deep learning technology to be used as a tool to investigate the epidemiology of rice blast disease in the future.