• KIPS Transactions on Software and Data Engineering
• /
• v.6 no.5
• /
• pp.267-278
• /
• 2017

As the same letter can be pronounced differently depending on word contexts, one should refer to a lexicon in order to pronounce a word correctly. Phonetic alphabets that lexicons adopt as well as pronunciations that lexicons describe for the same word can be different from lexicon to lexicon. In this paper, we use a sequence-to-sequence model that is widely used in deep learning research area in order to convert automatically from one pronunciation to another. The 12 seq2seq models are implemented based on pronunciation training data collected from 4 different lexicons. The exact accuracy of the models ranges from 74.5% to 89.6%. The aim of this study is the following two things. One is to comprehend a property of phonetic alphabets and pronunciations used in various lexicons. The other is to understand characteristics of seq2seq models by analyzing an error.

• Journal of KIISE
• /
• v.43 no.7
• /
• pp.795-800
• /
• 2016

In image classification, recent CNNs compete with human performance. However, there are limitations in more general recognition. Herein we deal with indoor images that contain too much information to be directly processed and require information reduction before recognition. To reduce the amount of data processing, typically variational inference or variational Bayesian methods are suggested for object detection. However, these methods suffer from the difficulty of marginalizing over the given space. In this study, we propose an image-text integrated recognition system using active vision based on Spatial Transformer Networks. The system attempts to efficiently sample a partial region of a given image for a given language information. Our experimental results demonstrate a significant improvement over traditional approaches. We also discuss the results of qualitative analysis of sampled images, model characteristics, and its limitations.

• Journal of Advanced Navigation Technology
• /
• v.23 no.5
• /
• pp.473-480
• /
• 2019

The smart home is a technology that can monitor and control by connecting everything to a communication network in various fields such as home appliances, energy consumers, and security devices. The Smart home is developing not only automatic control but also learning situation and user's taste and providing the result accordingly. This paper proposes a model that can provide a comfortable indoor environment control service for the user's characteristics by detecting the user's behavior as well as the automatic remote control service. The whole system consists of ESP 8266 with sensor and Wi-Fi, Firebase as a real-time database, and a smartphone application. This model is divided into functions such as learning mode when the home appliance is operated, learning control through learning results, and automatic ventilation using indoor and outdoor sensor values. The study used moving averages for temperature and humidity in the control of home appliances such as air conditioners, humidifiers and air purifiers. This system can provide higher quality service by analyzing and predicting user's characteristics through various machine learning and deep learning.

• Journal of Korea Society of Industrial Information Systems
• /
• v.24 no.4
• /
• pp.1-8
• /
• 2019

Plant phenomics is a technique for observing and analyzing morphological features in order to select plant varieties of excellent traits. The conventional methods is difficult to apply to the phenomics system. because the color threshold value must be manually changed according to the detection target. In this paper, we propose the convolution neural network (CNN) structure that can automatically segment plants from the background for the phenomics system. The LeafNet consists of nine convolution layers and a sigmoid activation function for determining the presence of plants. As a result of the learning using the LeafNet, we obtained a precision of 98.0% and a recall rate of 90.3% for the plant seedlings images. This confirms the applicability of the phenomics system.

• Journal of IKEEE
• /
• v.23 no.1
• /
• pp.120-126
• /
• 2019

Recently, accurate prediction of power consumption based on machine learning techniques in Internet of Energy (IoE) has been actively studied using the large amount of electricity data acquired from advanced metering infrastructure (AMI). In this paper, we propose a deep learning model based on Gated Recurrent Unit (GRU) as an artificial intelligence (AI) network that can effectively perform pattern recognition of time series data such as the power consumption, and analyze performance of the prediction based on real household power usage data. In the performance analysis, performance comparison between the proposed GRU-based learning model and the conventional learning model of Long Short Term Memory (LSTM) is described. In the simulation results, mean squared error (MSE), mean absolute error (MAE), forecast skill score, normalized root mean square error (RMSE), and normalized mean bias error (NMBE) are used as performance evaluation indexes, and we confirm that the performance of the prediction of the proposed GRU-based learning model is greatly improved.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.8
• /
• pp.31-37
• /
• 2021

In this paper, we propose a method to effectively search by automatically classifying scenes in which specific images such as pitching or swing appear in baseball game images using a CNN(Convolution Neural Network) model. In addition, we propose a video scene search system that links the classification results of specific motions and game records. In order to test the efficiency of the proposed system, an experiment was conducted to classify the Korean professional baseball game videos from 2018 to 2019 by specific scenes. In an experiment to classify pitching scenes in baseball game images, the accuracy was about 90% for each game. And in the video scene search experiment linking the game record by extracting the scoreboard included in the game video, the accuracy was about 80% for each game. It is expected that the results of this study can be used effectively to establish strategies for improving performance by systematically analyzing past game images in Korean professional baseball games.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.8
• /
• pp.1248-1254
• /
• 2022

Efforts are being made to prevent traffic accidents in the school zone in advance. However, traffic accidents in school zones continue to occur. If the driver can know the situation information in the child protection area in advance, accidents can be reduced. In this paper, we design a camera that eliminates blind spots in school zones and a number recognition camera system that can collect pre-traffic information. It is designed by improving the LIDAR system that recognizes vehicle speed and pedestrians. It collects and processes pedestrian and vehicle image information recognized by cameras and LIDAR, and applies artificial intelligence time series analysis and artificial intelligence algorithms. The artificial intelligence traffic accident prevention system learned by deep learning proposed in this paper provides a forced push service that delivers school zone information to the driver to the mobile device in the vehicle before entering the school zone. In addition, school zone traffic information is provided as an alarm on the LED signboard.

• The Journal of Society for e-Business Studies
• /
• v.27 no.1
• /
• pp.1-13
• /
• 2022

With the development of network technologies, the security to protect organizational resources from internal and external intrusions and threats becomes more important. Therefore in recent years, the anomaly detection algorithm that detects and prevents security threats with respect to various security log events has been actively studied. Security anomaly detection algorithms that have been developed based on rule-based or statistical learning in the past are gradually evolving into modeling based on machine learning and deep learning. In this study, we propose a deep-autoencoder model that transforms LSTM-autoencoder as an optimal algorithm to detect insider threats in advance using various machine learning analysis methodologies. This study has academic significance in that it improved the possibility of adaptive security through the development of an anomaly detection algorithm based on unsupervised learning, and reduced the false positive rate compared to the existing algorithm through supervised true positive labeling.

• Journal of the Korea Society of Computer and Information
• /
• v.27 no.9
• /
• pp.33-40
• /
• 2022

In this paper, we propose a system that analyzes drone photographic images of panel-type factory roofs and conducts abnormal detection of bolts. Currently, inspectors directly climb onto the roof to carry out the inspection. However, safety accidents caused by working conditions at high places are continuously occurring, and new alternatives are needed. In response, the results of drone photography, which has recently emerged as an alternative to the dangerous environment inspection plan, will be easily inspected by finding the location of abnormal bolts using deep learning. The system proposed in this study proceeds with scanning the captured drone image using a sample image for the situation where the bolt cap is released. Furthermore, the scanned position is discriminated by using AI, and the presence/absence of the bolt abnormality is accurately discriminated. The AI used in this study showed 99% accuracy in test results based on VGGNet.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.4
• /
• pp.13-20
• /
• 2023

In this paper, the actual data of VOC reduction facilities was analyzed through a model that detects and predicts data anomalies. Using the USAD model, which shows stable performance in the field of anomaly detection, anomalies in real-time data are detected and sensors that cause anomalies are searched. In addition, we propose a method of predicting and warning, when abnormalities that time will occur by predicting future outliers with an auto-regressive model. The experiment was conducted with the actual data of the VOC reduction facility, and the anomaly detection test results showed high detection rates with precision, recall, and F1-score of 98.54%, 89.08%, and 93.57%, respectively. As a result, averaging of the precision, recall, and F1-score for 8 sensors of detection rates were 99.64%, 99.37%, and 99.63%. In addition, the Hamming loss obtained to confirm the validity of the detection experiment for each sensor was 0.0058, showing stable performance. And the abnormal prediction test result showed stable performance with an average absolute error of 0.0902.