• 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.

• KIISE Transactions on Computing Practices
• /
• v.23 no.7
• /
• pp.440-445
• /
• 2017

Due to the advance of e-commerce systems, the number of people using online shopping and products has significantly increased. Therefore, the need for an accurate recommendation system is becoming increasingly more important. Recurrent neural network is a deep-learning algorithm that utilizes sequential information in training. In this paper, an evaluation is performed on the application of recurrent neural networks to recommendation systems. We evaluated three recurrent algorithms (RNN, LSTM and GRU) and three optimal algorithms(Adagrad, RMSProp and Adam) which are commonly used. In the experiments, we used the TensorFlow open source library produced by Google and e-commerce session data from RecSys Challenge 2015. The results using the optimal hyperparameters found in this study are compared with those of RecSys Challenge 2015 participants.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.17 no.5
• /
• pp.173-187
• /
• 2018

This paper proposes a computer vision and deep learning-based technique for surveillance camera system for vehicle counting as one part of parking lot management system. We applied the You Only Look Once version 2 (YOLOv2) detector and come up with a deep convolutional neural network (CNN) based on YOLOv2 with a different architecture and two models. The effectiveness of the proposed architecture is illustrated using a publicly available Udacity's self-driving-car datasets. After training and testing, our proposed architecture with new models is able to obtain 64.30% mean average precision which is a better performance compare to the original architecture (YOLOv2) that achieved only 47.89% mean average precision on the detection of car, truck, and pedestrian.

• Smart Structures and Systems
• /
• v.25 no.5
• /
• pp.605-617
• /
• 2020

The existence of damages in structures causes changes in the physical properties by reducing the modal parameters. In this paper, we develop a two-stages approach based on normalized Modal Strain Energy Damage Indicator (nMSEDI) for quick applications to predict the location of damage. A two-dimensional IsoGeometric Analysis (2D-IGA), Machine Learning Algorithm (MLA) and optimization techniques are combined to create a new tool. In the first stage, we introduce a modified damage identification technique based on frequencies using nMSEDI to locate the potential of damaged elements. In the second stage, after eliminating the healthy elements, the damage index values from nMSEDI are considered as input in the damage quantification algorithm. The hybrid of Teaching-Learning-Based Optimization (TLBO) with Artificial Neural Network (ANN) and Particle Swarm Optimization (PSO) are used along with nMSEDI. The objective of TLBO is to estimate the parameters of PSO-ANN to find a good training based on actual damage and estimated damage. The IGA model is updated using experimental results based on stiffness and mass matrix using the difference between calculated and measured frequencies as objective function. The feasibility and efficiency of nMSEDI-PSO-ANN after finding the best parameters by TLBO are demonstrated through the comparison with nMSEDI-IGA for different scenarios. The result of the analyses indicates that the proposed approach can be used to determine correctly the severity of damage in beam structures.

• KIPS Transactions on Software and Data Engineering
• /
• v.6 no.3
• /
• pp.155-160
• /
• 2017

Recently, Deep learning has been used successfully to solve many recognition problems. It has many advantages over existing machine learning methods that extract feature points through hand-crafting. Deep neural networks for human activity recognition split video data into frame images, and then classify activities by analysing the connectivity of frame images according to the time. But it is difficult to apply to actual problems which has small-scale activity classes. Because this situations has a problem of overfitting and insufficient training data. In this paper, we defined 5 type of small-scale human activities, and classified them. We construct video database using 700 video clips, and obtained a classifying accuracy of 74.00%.

• The Journal of Korean Institute of Information Technology
• /
• v.17 no.12
• /
• pp.81-92
• /
• 2019

Weeds are a major object which is very harmful to crops. To remove the weeds effectively, we have to classify them accurately and use herbicides. As computing technology has developed, image-based machine learning methods have been studied in this field, specially convolutional neural network(CNN) based models have shown good performance in public image dataset. However, CNN with numerous training parameters and high computational amount. Thus, it works under high hardware condition of expensive GPUs in real application. To solve these problems, in this paper, a hierarchical architecture based deep-learning model is proposed. The experimental results show that the proposed model successfully classify 21 species of the exotic weeds. That is, the model achieve 97.2612% accuracy with a small number of parameters. Our proposed model with a few parameters is expected to be applicable to actual application of network based classification services.

• Journal of the Korea Convergence Society
• /
• v.12 no.2
• /
• pp.21-28
• /
• 2021

In this study, we propose a deep learning-based NILM technique using actual measured power data for 5 kinds of home appliances and verify its effectiveness. For about 3 weeks, the active power of the central power measuring device and five kinds of home appliances (refrigerator, induction, TV, washing machine, air cleaner) was individually measured. The preprocessing method of the measured data was introduced, and characteristics of each household appliance were analyzed through spectogram analysis. The characteristics of each household appliance are organized into a learning data set. All the power data measured by the central power measuring device and 5 kinds of home appliances were time-series mapping, and training was performed using a LSTM neural network, which is excellent for time series data prediction. An algorithm that can disaggregate five types of energies using only the power data of the main central power measuring device is proposed.

• Smart Structures and Systems
• /
• v.29 no.1
• /
• pp.117-127
• /
• 2022

Data anomalies seriously threaten the reliability of the bridge structural health monitoring system and may trigger system misjudgment. To overcome the above problem, an efficient and accurate data anomaly detection method is desiderated. Traditional anomaly detection methods extract various abnormal features as the key indicators to identify data anomalies. Then set thresholds artificially for various features to identify specific anomalies, which is the artificial experience method. However, limited by the poor generalization ability among sensors, this method often leads to high labor costs. Another approach to anomaly detection is a data-driven approach based on machine learning methods. Among these, the bidirectional long-short memory neural network (BiLSTM), as an effective classification method, excels at finding complex relationships in multivariate time series data. However, training unprocessed original signals often leads to low computation efficiency and poor convergence, for lacking appropriate feature selection. Therefore, this article combines the advantages of the two methods by proposing a deep learning method with manual experience statistical features fed into it. Experimental comparative studies illustrate that the BiLSTM model with appropriate feature input has an accuracy rate of over 87-94%. Meanwhile, this paper provides basic principles of data cleaning and discusses the typical features of various anomalies. Furthermore, the optimization strategies of the feature space selection based on artificial experience are also highlighted.

• The Journal of Korean Institute of Next Generation Computing
• /
• v.13 no.2
• /
• pp.102-110
• /
• 2017

Current word sense disambiguation techniques employ various machine learning-based methods. Various approaches have been proposed to address this problem, including the knowledge base approach. This approach defines the sense of an ambiguous word in accordance with knowledge base information with no training corpus. In unsupervised learning techniques that use a knowledge base approach, graph-based and similarity-based methods have been the main research areas. The graph-based method has the advantage of constructing a semantic graph that delineates all paths between different senses that an ambiguous word may have. However, unnecessary semantic paths may be introduced, thereby increasing the risk of errors. To solve this problem and construct a fine-grained graph, in this paper, we propose a model that iteratively constructs the graph while eliminating unnecessary nodes and edges, i.e., senses and semantic paths. The hybrid similarity estimation model was applied to estimate a more accurate sense in the constructed semantic graph. Because the proposed model uses BabelNet, a multilingual lexical knowledge base, the model is not limited to a specific language.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.11a
• /
• pp.377-379
• /
• 2022

POI refers to the point of Interest in Location-Based Social Networks (LBSNs). With the rapid development of mobile devices, GPS, and the Web (web2.0 and 3.0), LBSNs have attracted many users to share their information, physical location (real-time location), and interesting places. The tremendous demand of the user in LBSNs leads the recommendation systems (RSs) to become more widespread attention. Recommendation systems assist users in discovering interesting local attractions or facilities and help social network service (SNS) providers based on user locations. Therefore, it plays a vital role in LBSNs, namely POI recommendation system. In the machine learning model, most of the training data are stored in the centralized data storage, so information that belongs to the user will store in the centralized storage, and users may face privacy issues. Moreover, sharing the information may have safety concerns because of uploading or sharing their real-time location with others through social network media. According to the privacy concern issue, the paper proposes a recommendation model to prevent user privacy and eliminate traditional RS problems such as cold-start and data sparsity.