• Journal of the Korean Society for Precision Engineering
• /
• v.27 no.11
• /
• pp.30-36
• /
• 2010

Previous researches show that linearly integrated Hall sensor arrays (LIHaS) can detect cracks in the steel structure fast and effectively This paper proposes an algorithm that estimates the size and shape of cracks for the developed LIHaS. In most nondestructive testing (NDT), just crack existence and location are obtained by processing 1-dimensional data from the sensor that scans the object with relative speed in single direction. The proposed method is composed with two steps. The first step is constructing 2-dimensionally mapped data space by combining the converted position data from the time-based scan data with the position information of sensor arrays those are placed in the vertical direction to the scan direction. The second step is applying designed Laplacian filter and smoothing filter to estimate the size and shape of cracks. The experimental results of express train wheels show that the proposed algorithm is not only more reliable and accurate to detecting cracks but also effective to estimate the size and shape of cracks.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.11 no.1
• /
• pp.67-72
• /
• 2010

In this paper, we verify the driverless operation possibility of MBS, which could overcome the defects of conventional track-circuit-based FBS, such as additional needs of maintenance and others problems according to short-circuit sensibility and, and which could allow the minimal interval between trains. With MBS, we can expect the reduction of headway, then the increase of transportation demand, and the protection of unnecessary speed variation because it allows the real time detection of train position from central office, and direct transmission of data between preceding trains and the following ones. In addition, it is possible to reduce the number of wayside-equipment substantially, to improve the passenger service, and to the achieve the positive economic effects by comfortable ride.

• Korean Journal of Radiology
• /
• v.24 no.6
• /
• pp.541-552
• /
• 2023

Objective: Detection of pneumoperitoneum using abdominal radiography, particularly in the supine position, is often challenging. This study aimed to develop and externally validate a deep learning model for the detection of pneumoperitoneum using supine and erect abdominal radiography. Materials and Methods: A model that can utilize "pneumoperitoneum" and "non-pneumoperitoneum" classes was developed through knowledge distillation. To train the proposed model with limited training data and weak labels, it was trained using a recently proposed semi-supervised learning method called distillation for self-supervised and self-train learning (DISTL), which leverages the Vision Transformer. The proposed model was first pre-trained with chest radiographs to utilize common knowledge between modalities, fine-tuned, and self-trained on labeled and unlabeled abdominal radiographs. The proposed model was trained using data from supine and erect abdominal radiographs. In total, 191212 chest radiographs (CheXpert data) were used for pre-training, and 5518 labeled and 16671 unlabeled abdominal radiographs were used for fine-tuning and self-supervised learning, respectively. The proposed model was internally validated on 389 abdominal radiographs and externally validated on 475 and 798 abdominal radiographs from the two institutions. We evaluated the performance in diagnosing pneumoperitoneum using the area under the receiver operating characteristic curve (AUC) and compared it with that of radiologists. Results: In the internal validation, the proposed model had an AUC, sensitivity, and specificity of 0.881, 85.4%, and 73.3% and 0.968, 91.1, and 95.0 for supine and erect positions, respectively. In the external validation at the two institutions, the AUCs were 0.835 and 0.852 for the supine position and 0.909 and 0.944 for the erect position. In the reader study, the readers' performances improved with the assistance of the proposed model. Conclusion: The proposed model trained with the DISTL method can accurately detect pneumoperitoneum on abdominal radiography in both the supine and erect positions.

• Journal of the Korean Society for Railway
• /
• v.20 no.5
• /
• pp.602-611
• /
• 2017

In railway systems, the exchange of information between running trains and wayside equipment is a very important role in various applications such as position detection and train control. Track circuits have been used as the medium for information transmission between trains and wayside. However, track circuits must be installed continuously along the track on the ground, resulting in an inevitable increase in installation and maintenance costs. One of the most promising solutions to reduce these costs is to mix continuous information transmission (via wireless communication) and discontinuous information transmission (via transponder). In this study, we designed antennas of railway high-speed transponder readers and tags for wireless power transmission, which can be used to transmit information from ground to high-speed trains with a maximum speed of 400km/h. We also verified system performance through computational simulation and prototyping.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.12
• /
• pp.3868-3888
• /
• 2022

A widely used social networking service like Twitter has the ability to disseminate information to large groups of people even during a pandemic. At the same time, it is a convenient medium to share irrelevant and unverified information online and poses a potential threat to society. In this research, conventional machine learning algorithms are analyzed to classify the data as either non-rumor data or rumor data. Machine learning techniques have limited tuning capability and make decisions based on their learning. To tackle this problem the authors propose a deep learning-based Rumor Detection Neural Network model to predict the rumor tweet in real-world events. This model comprises three layers, AttCNN layer is used to extract local and position invariant features from the data, AttBi-LSTM layer to extract important semantic or contextual information and HPOOL to combine the down sampling patches of the input feature maps from the average and maximum pooling layers. A dataset from Kaggle and ground dataset #gaja are used to train the proposed Rumor Detection Neural Network to determine the veracity of the rumor. The experimental results of the RDNN Classifier demonstrate an accuracy of 93.24% and 95.41% in identifying rumor tweets in real-time events.

• Journal of the Korean Society for Railway
• /
• v.14 no.6
• /
• pp.515-525
• /
• 2011

In this study, Urban maglev is applied to inductive-loop speed and position detection system for the top-level classification system for the entire system, and performed functional analysis On-board signal equipment, Wayside-signal equipment divided by the reliability, availability, maintainability, and safety through analysis of the proposed formula. RDB and by applying a system service for each device was calculated to availability, safety analysis. The PHA, FMEA, HAZOP over the Top Event of the FTA is performed by presenting the results. This also shows approach methods and relative activities for project to accomplish and ensure the system requirements.

• Journal of the Korean Society for Nondestructive Testing
• /
• v.31 no.2
• /
• pp.111-117
• /
• 2011

Vibration monitoring is required for reliable thermosonic testing to decide whether sufficient vibration is achieved in each test for the detection of cracks. From a practical point of view, a cheaper and convenient monitoring method is better for the application to real tests. Therefore, the performance of different sensors for vibration monitoring was investigated and compared in this study to find a convenient and acceptable measurement method for thermosonics. Velocity measured by a laser vibrometer and strain provide an equivalent HI when measured at the same position. The microphone can provide a cheaper vibration monitoring device than the laser and the heating index calculated by a microphone signal shows similar characteristics to that calculated from velocity measured by the laser vibrometer. The microphone frequency response shows that it underestimates high frequency components but it is applicable to practical tests because it gives a conservative value of HI.

• Proceedings of the KIEE Conference
• /
• 2003.04a
• /
• pp.415-417
• /
• 2003

TOA(Time of Arrival) 및 TDOA(Time Difference of Arrival)경우 무선국의 시간동기화를 위해서 고도의 기술을 요구하고 있으며, 시간동기오차에 따른 위치검지의 정밀도가 낮아지는 문제가 있어 이를 극복하기 위하여 위상차를 이용한 새로운 열차검지기법의 제안에 따른 구현을 위하여 무선장치 설계에 대하여 기술하고자 한다. 본 시스템은 전파의 전달 속도($\lambda$)를 응용하여 기준 주파수인 1.5MHz를 송신 시스템과 수신 시스템의 기준 주파수와 비교하여 그 위상의 차이를 비교하여 지연된 시간을 구한 후 이를 거리로 환산하는 시스템으로서 무선장치와 S/W로 구분하여 구현 설계하였다.

• Proceedings of the KIEE Conference
• /
• 2003.07b
• /
• pp.1267-1269
• /
• 2003

경량전철시스템의 개발시험을 수행하기 위하여 시험선을 구축하는데 있어서 각 하부시스템별로 요구사항을 도출하고 이를 설계에 반영하여야 한다. 이에 따라 본 논문에서는 신호제어시스템에 대하여 성능 및 기능시험을 감안한 요구사항을 도술하고 시험환경 및 시험조건에 따라 연차 검지 및 안전제동거리의 시험방법에 대해 검토하여 최적의 시험기법을 제안하고자 한다.

• Korean Journal of Optics and Photonics
• /
• v.7 no.3
• /
• pp.251-259
• /
• 1996

Two-Photon Fluorescence (TPF) experiment measures temporal width of an amplified short laser pulse which has passed through a four-pass Nd: glass amplifier, after selecting a single pulse from pulse train Q-switched and mode-locked(QSML) in Nd:YLF master oscillator. Determination of pulsewidth and pulseshape was also made with detection of autocorrelation trace of CW mode-locked pulse train by using noncollinear type I Second Harmonic Generation (SHG) method. The observed TPF track showed various patterns, depending on pulse-selecting position in QSML pulse train. That is, autocorrelation of a pulse extracted at front of the train displayed smooth pulse shape, while one from the trailing part of the train created many sharp spikes and substructure in the pulse. By TPF method, pulsewidth was measured to be 44.4 ps with contrast ratio of 2.86 which enabled us to find out energy fraction of a pulse to total energy, (sum of pulse and background); we obtain the value of 0.62. Pulsewidth of 46.6ps was also acquired in another SHG experiment with the help of only mode-locked pulse train. On the other hand, we confirmed that shape of the pulse is close to $sech^2$ one as a result of fitting the SHG autocorrelation signal with various functions. With simulation using this $sech^2$ type of pulse, pulsewidth reduction of the beam, having passed through four-pass amplifier, was also verified.