• Proceedings of the KSR Conference
• /
• 2003.10c
• /
• pp.132-137
• /
• 2003

This paper concerns the prediction of railway noise propagation using scale model experiment in acoustics. In order to make acoustical experiment the digital signal processing technique are applied and spark discharge sound sources have been developed in which impulse response measured in 1/20 scale model railway. In the case of scale model experiment, it is difficult to realize sufficiently small size and directivity and to get sufficient sound energy and to get repeatability. Several type of Spark discharge sound source is made in laboratory. Experiment results are compared with the calculated results by the prediction model. As the results, it was found that railway noise could be predicted in acoustical scale model experiment using spark discharge sound source.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2013.10a
• /
• pp.564-569
• /
• 2013

An accurate railway environmental noise prediction model is required to make the proper solution of the railway noise problems. In this paper, an engineering model for predicting the noise of conventional passenger cars is presented considering the acoustic source strength in octave-band frequencies and the propagation over grounds with varying surface properties. Since the formation of a train can be variable, the source strength of each locomotive and passenger car was estimated by measuring the pass-by noise and analysing the wheel-rail rolling noise. Some validation cases show on the average small differences between the predictions of the present model and the measurement results.

• International Journal of Internet, Broadcasting and Communication
• /
• v.11 no.1
• /
• pp.47-53
• /
• 2019

The idea of ensemble learning is to train multiple models, each with the objective to predict or classify a set of results. Most of the errors from a model's learning are from three main factors: variance, noise, and bias. By using ensemble methods, we're able to increase the stability of the final model and reduce the errors mentioned previously. By combining many models, we're able to reduce the variance, even when they are individually not great. In this paper we propose an ensemble model and applied it to classification problem. In iris, Pima indian diabeit and semiconductor fault detection problem, proposed model classifies well compared to traditional single classifier that is logistic regression, SVM and random forest.

• Journal of Multimedia Information System
• /
• v.9 no.2
• /
• pp.93-102
• /
• 2022

Dependency parsing is a decision problem of the syntactic relation between words in a sentence. Recently, deep learning models are used for dependency parsing based on the word representations in a continuous vector space. However, it causes a mislabeled tagging problem for the proper nouns that rarely appear in the training corpus because it is difficult to express out-of-vocabulary (OOV) words in a continuous vector space. To solve the OOV problem in dependency parsing, we explored the proper noun embedding method according to the embedding unit. Before representing words in a continuous vector space, we replace the proper nouns with a special token and train them for the contextual features by using the multi-layer bidirectional LSTM. Two models of the syllable-based and morpheme-based unit are proposed for proper noun embedding and the performance of the dependency parsing is more improved in the ensemble model than each syllable and morpheme embedding model. The experimental results showed that our ensemble model improved 1.69%p in UAS and 2.17%p in LAS than the same arc-eager approach-based Malt parser.

• Journal of information and communication convergence engineering
• /
• v.21 no.1
• /
• pp.17-23
• /
• 2023

Recently, several high-performance instance segmentation models have used the Mask R-CNN model as a baseline, which reached a historical peak in instance segmentation in 2017. There are numerous derived models using the Mask R-CNN model, and if the performance of Mask R-CNN is improved, the performance of the derived models is also anticipated to improve. The Mask R-CNN uses interpolation to adjust the image size, and the input differs depending on the interpolation method. Therefore, in this study, the performance change of Mask R-CNN was compared when various interpolation methods were applied to the transform layer to improve the performance of Mask R-CNN. To train and evaluate the models, this study utilized the PennFudan and Balloon datasets and the AP metric was used to evaluate model performance. As a result of the experiment, the derived Mask R-CNN model showed the best performance when bicubic interpolation was used in the transform layer.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• fall
• /
• pp.91-94
• /
• 2021

As deep learning technologies becoming developed, realistic fake videos synthesized by deep learning models called "Deepfake" videos became even more difficult to distinguish from original videos. As fake news or Deepfake blackmailing are causing confusion and serious problems, this paper suggests a novel model detecting Deepfake videos. We chose Residual Convolutional Neural Network (Resnet50) as an extraction model and Long Short-Term Memory (LSTM) which is a form of Recurrent Neural Network (RNN) as a classification model. We adopted cosine similarity with hinge loss to train our extraction model in embedding the features of Deepfake and original video. The result in this paper demonstrates that temporal features in the videos are essential for detecting Deepfake videos.

• Smart Media Journal
• /
• v.12 no.11
• /
• pp.125-133
• /
• 2023

Traditional attire has assumed a pivotal role within the contemporary fashion industry. The objective of this study is to construct a computer vision model tailored to the recognition of traditional costumes originating from five distinct countries, namely India, Korea, Japan, Tanzania, and Vietnam. Leveraging a dataset comprising 1,608 images, we proceeded to train the cutting-edge computer vision model YOLOv8. The model yielded an impressive overall mean average precision (MAP) of 96%. Notably, the Indian sari exhibited a remarkable MAP of 99%, the Tanzanian kitenge 98%, the Japanese kimono 92%, the Korean hanbok 89%, and the Vietnamese ao dai 83%. Furthermore, the model demonstrated a commendable overall box precision score of 94.7% and a recall rate of 84.3%. Within the realm of the fashion industry, this model possesses considerable utility for trend projection and the facilitation of personalized recommendation systems.

• Journal of information and communication convergence engineering
• /
• v.22 no.3
• /
• pp.256-266
• /
• 2024

Sasang constitutional medicine (SCM) is one of the best traditional therapeutic approaches used in Korea. SCM prioritizes personalized treatment that considers the unique constitution of an individual and encompasses their physical characteristics, personality traits, and susceptibility to specific diseases. Facial features are essential for diagnosing Sasang constitutional types (SCTs). This study aimed to develop a real-time artificial intelligence-based model for diagnosing SCTs using facial images, building an SCTs prediction model based on a machine learning method. Facial features from all images were extracted to develop this model using feature engineering and machine learning techniques. The fusion of these features was used to train the AI model. We used four machine learning algorithms, namely, random forest (RF), multilayer perceptron (MLP), gradient boosting machine (GBM), and extreme gradient boosting (XGB), to investigate SCTs. The GBM outperformed all the other models. The highest accuracy achieved in the experiment was 81%, indicating the robustness of the proposed model and suitability for real-time applications.

• Journal of the Korean Society for Railway
• /
• v.16 no.1
• /
• pp.7-13
• /
• 2013

A scaled version of a roller rig is developed to demonstrate the dynamic characteristics of a railway vehicle for academic purposes. This rig is designed based on Jaschinski's similarity law. It is scaled to 1/10 of actual size and allows 9-DOF motion to examine the up and down vibration of a train set. The test rig consists of three sub-hardware components: (i) a driving roller mechanism with a three-phase AC motor and an inverter, (ii) a bogie structure with first and second suspensions, and (iii) the vehicle body. The motor of the rig is capable of 3,600rpm, allowing the test to simulate a vehicle up to a maximum speed of 400Km/hr. Because bearings and joints are properly connected to the sub-structures, various motion analyses, such as a lateral, pitching, and yawing motion, are allowed. The slip motion between the rail and the wheel set is also monitored by several sensors mounted in the rig. After the construction of the hardware, an experiment is conducted to obtain the natural frequencies of the dynamic behavior of the specimen. First, the test rig is run and data are collected from six sets of accelerometers. Then, a numerical analysis of the model based on the ADAMS program is derived. Finally, the measurement data of the first three fundamental frequencies are compared to the analytical result and the validation of the test rig is conducted. The results show that the developed roller rig provides good accuracy in simulating the dynamic behavior of the vehicle motion. Although the roller rig designed in this paper is intended for academia, it can easily be implemented as part of a dynamic experiment of a bogie and a vehicle body for a high-speed train as part of the research efforts in this area.

• Journal of the Korean Society of Safety
• /
• v.31 no.6
• /
• pp.45-51
• /
• 2016

The purpose of this study is to investigate the influence of train-induced end rotation of simple supported track girder on the performance of a direct fixation track system (DFTS) in Yeongjong grand bridge. In this study, the influences of deflection of a DFTS and track girder on dynamic rail-track girder interaction forces for the track girder ends currently employed in airport express lines were assessed by performing field tests using actual vehicles running along the service lines. Therefore, the dynamic displacement of rail and track girder and the fastener stress on the center and ends sections of DFTS were measured for two different trains (AREX and KTX) running in Yeongjong grand bridge. A three-dimensional finite element analysis (FEA) model using the time-history function based on the design wheel load was used to predict the train-induced track and track girder displacement, and the FEA and field test results were compared. The analytical results reproduced the experimental results well within about 3-7% difference in the values. Therefore, the FEA model of DFTS on track girder is considered to provide sufficiently reliable FEA results in the investigation of the behavior of DFTS. Using the analytical and experimental results, the influence of train-induced end rotation of simple supported track girder on the interaction behavior of rail and track girder installed on a simple supported track girder ends, i.e., upward displacement of rail-track girder and the fastener stress, was investigated. It was found that the train-induced end rotation effect of track girder was not significantly affected by the upward displacement of rails and the fastener stresses of track girder ends. Further, the interaction behavior of rail and track girder were similar to or less than that of the general railway bridge deck ends, nevertheless the vertical displacement of track was higher than that of conventional DFTS on the general railway bridge. From the results, the dynamic responses of the DFTS on track girder ends were not significantly affected by the safety and stability of DFTS ends.