• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.32 no.6
• /
• pp.384-395
• /
• 2020

Breaking waves generated by wave shoaling in coastal areas have a close relationship with various physical phenomena in coastal regions, such as sediment transport, longshore currents, and shock wave pressure. Therefore, it is crucial to accurately predict breaker index such as breaking wave height and breaking depth, when designing coastal structures. Numerous scientific efforts have been made in the past by many researchers to identify and predict the breaking phenomenon. Representative studies on wave breaking provide many empirical formulas for the prediction of breaking index, mainly through hydraulic model experiments. However, the existing empirical formulas for breaking index determine the coefficients of the assumed equation through statistical analysis of data under the assumption of a specific equation. In this paper, we applied a representative linear-based supervised machine learning algorithms that show high predictive performance in various research fields related to regression or classification problems. Based on the used machine learning methods, a model for prediction of the breaking index is developed from previously published experimental data on the breaking wave, and a new linear equation for prediction of breaker index is presented from the trained model. The newly proposed breaker index formula showed similar predictive performance compared to the existing empirical formula, although it was a simple linear equation.

• Convergence Security Journal
• /
• v.22 no.3
• /
• pp.25-32
• /
• 2022

Recently, the image processing industry has been activated as deep learning-based technology is introduced in the image recognition and detection field. With the development of deep learning technology, learning model vulnerabilities for adversarial attacks continue to be reported. However, studies on countermeasures against poisoning attacks that inject malicious data during learning are insufficient. The conventional countermeasure against poisoning attacks has a limitation in that it is necessary to perform a separate detection and removal operation by examining the training data each time. Therefore, in this paper, we propose a technique for reducing the attack success rate by applying modifications to the training data and inference data without a separate detection and removal process for the poison data. The One-shot kill poison attack, a clean label poison attack proposed in previous studies, was used as an attack model. The attack performance was confirmed by dividing it into a general attacker and an intelligent attacker according to the attacker's attack strategy. According to the experimental results, when the proposed defense mechanism is applied, the attack success rate can be reduced by up to 65% compared to the conventional method.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.3
• /
• pp.91-98
• /
• 2021

Today's information and communication technology is rapidly developing, the security of IT infrastructure is becoming more important, and at the same time, cyber attacks of various forms are becoming more advanced and sophisticated like intelligent persistent attacks (Advanced Persistent Threat). Early defense or prediction of increasingly sophisticated cyber attacks is extremely important, and in many cases, the analysis of network-based intrusion detection systems (NIDS) related data alone cannot prevent rapidly changing cyber attacks. Therefore, we are currently using data generated by intrusion detection systems to protect against cyber attacks described above through Host-based Intrusion Detection System (HIDS) data analysis. In this paper, we conducted a comparative study on machine learning algorithms using LID-DS (Leipzig Intrusion Detection-Data Set) host-based intrusion detection data including thread information, metadata, and buffer data missing from previously used data sets. The algorithms used were Decision Tree, Naive Bayes, MLP (Multi-Layer Perceptron), Logistic Regression, LSTM (Long Short-Term Memory model), and RNN (Recurrent Neural Network). Accuracy, accuracy, recall, F1-Score indicators and error rates were measured for evaluation. As a result, the LSTM algorithm had the highest accuracy.

• Journal of KIISE:Software and Applications
• /
• v.34 no.1
• /
• pp.30-39
• /
• 2007

The existing content-based spam mail filtering algorithms have difficulties in filtering spam mails when e-mails contain images but little text. In this thesis we propose an efficient spam mail classification algorithm that utilizes the link structure of e-mails. We compute the number of hyperlinks in an e-mail and the in-link frequencies of the web pages hyperlinked in the e-mail. Using these two features we classify spam mails and legitimate mails based on the decision tree trained for spam mail classification. We also suggest a hybrid system combining three different algorithms by majority voting: the link structure analysis algorithm, a modified link structure analysis algorithm, in which only the host part of the hyperlinked pages of an e-mail is used for link structure analysis, and the content-based method using SVM (support vector machines). The experimental results show that the link structure analysis algorithm slightly outperforms the existing content-based method with the accuracy of 94.8%. Moreover, the hybrid system achieves the accuracy of 97.0%, which is a significant performance improvement over the existing method.

• The Journal of Korean Institute of Information Technology
• /
• v.17 no.5
• /
• pp.13-20
• /
• 2019

Recently, object detection technology based on CNN has been actively studied. Object detection technology is used as an important technology in autonomous vehicles, intelligent image analysis, and so on. In this paper, we propose a rotation change robust object detection system by applying MI-FL (Moment Invariant-Feature Layer) to SSD (Single Shot Multibox Detector) which is one of CNN-based object detectors. First, the features of the input image are extracted based on the VGG network. Then, a total of six feature layers are applied to generate bounding boxes by predicting the location and type of object. We then use the NMS algorithm to get the bounding box that is the most likely object. Once an object bounding box has been determined, the invariant moment feature of the corresponding region is extracted using MI-FL, and stored and learned in advance. In the detection process, it is possible to detect the rotated image more robust than the conventional method by using the previously stored moment invariant feature information. The performance improvement of about 4 ~ 5% was confirmed by comparing SSD with existing SSD and MI-FL.

• Journal of the Korea Computer Graphics Society
• /
• v.25 no.3
• /
• pp.123-131
• /
• 2019

In this study, we modified CT images of femoral head in consideration of anatomically meaningful structure, proposing the method to augment the training data of convolution Neural network for segmentation of femur mesh model. First, the femur mesh model is obtained from the CT image. Then divide the mesh model into meaningful parts by using cluster analysis on geometric characteristic of mesh surface. Finally, transform the segments by using an appropriate mesh deformation algorithm, then create new CT images by warping CT images accordingly. Deep learning models using the data enhancement methods of this study show better image division performance compared to data augmentation methods which have been commonly used, such as geometric conversion or color conversion.

• The Korean Journal of Applied Statistics
• /
• v.34 no.3
• /
• pp.389-399
• /
• 2021

In recent years, word embedding has been a popular field of natural language processing research and a skip-gram has become one successful word embedding method. It assigns a word embedding vector to each word using contexts, which provides an effective way to analyze text data. However, due to the limitation of vector space model, primary word embedding methods assume that every word only have a single meaning. As one faces multi-sense words, that is, words with more than one meaning, in reality, Neelakantan (2014) proposed a multi-sense skip-gram (MSSG) to find embedding vectors corresponding to the each senses of a multi-sense word using a clustering method. In this paper, we propose a modified method of the MSSG to improve statistical accuracy. Moreover, we propose a data-adaptive choice of the number of clusters, that is, the number of meanings for a multi-sense word. Some numerical evidence is given by conducting real data-based simulations.

• Journal of Industrial Convergence
• /
• v.19 no.4
• /
• pp.1-8
• /
• 2021

This study was conducted with the aim of developing a mobile app so that parents and children can provide questions and conversation-based Home Humanity Education. App development was developed in four stages: analysis, design, development, and evaluation by modifying the ADDIE model. In the analysis stage, education goals were set and seven core competencies were extracted for the app. In the design stage, 15 menus, including character to implement seven core competencies, were selected for learning contents. In the development stage, questions were asked based on 1,000 photos, and parents and children naturally talked to strengthen their core competencies. In the evaluation stage, the user's evaluation interview was followed by a reflux process that produced and reflected improvements. Based on research results, it is going to suggest follow-up research on developing apps for home humanity education. First, it is necessary to verify the learning experiences and changes that occur when parents and children use apps for a long time through longitudinal research. Second, a follow-up study is needed on how to use the home humanity education app by age and how to improve it accordingly. Third, as the number of cases in which grandparents raise their grandchildren increases, it is necessary to study the process and learning effects of grandparents and grandchildren using the app.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.10
• /
• pp.35-42
• /
• 2020

In this study, we propose a crack detection method using limited data with a U-Net based image inpainting technique that is a modified unsupervised anomaly detection method. Concrete cracking occurs due to a variety of causes and is a factor that can cause serious damage to the structure in the long term. In general, crack investigation uses an inspector's visual inspection on the concrete surfaces, which is less objective in judgment and has a high possibility of human error. Therefore, a method with objective and accurate image analysis processing is required. In recent years, the methods using deep learning have been studied to detect cracks quickly and accurately. However, when the amount of crack data on the building or infrastructure to be inspected is small, existing crack detection models using it often show a limited performance. Therefore, in this study, an unsupervised anomaly detection method was used to augment the data on the object to be inspected, and as a result of learning using the data, we confirmed the performance of 98.78% of accuracy and 82.67% of harmonic average (F1_Score).

• Journal of KIISE:Software and Applications
• /
• v.30 no.7_8
• /
• pp.649-658
• /
• 2003

We propose a hierarchical architecture of ART2 Network for performance improvement and fast pattern classification model using fitness selection. This hierarchical network creates coarse clusters as first ART2 network layer by unsupervised learning, then creates fine clusters of the each first layer as second network layer by supervised learning. First, it compares input pattern with each clusters of first layer and select candidate clusters by fitness measure. We design a optimized fitness function for pruning clusters by measuring relative distance ratio between a input pattern and clusters. This makes it possible to improve speed and accuracy. Next, it compares input pattern with each clusters connected with selected clusters and finds winner cluster. Finally it classifies the pattern by a label of the winner cluster. Results of our experiments show that the proposed method is more accurate and fast than other approaches.