• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.41 no.1
• /
• pp.69-76
• /
• 2004

Recently, The development of DTV receiver and the need of its additional services have been increased vastly. In this paper, we implement new closed caption engine for the deaf and hard of hearing person and languages studying on DTV receiver, The specification of domestic closed caption is almost adopted that of EIA-608A. In this paper, with fully following the specification, we will present how to implement functions of closed caption with new algorithm. the function includes paint-on, Pop-on, roll-up/down, etc. experimental results show that the proposed technique provides satisfactory performance on DTV receiver.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.30 no.3
• /
• pp.236-242
• /
• 2019

Ultra-wideband(UWB) is a technology that can transmit and receive signals at high speeds using a very short signal of wideband of several GHz, and has been recently used in the field of radar technology. Impulse radio(IR)-UWB radar is used in the field of motion recognition with high resolution. In this work, we studied motion recognition using IR-UWB radar. We constructed a development environment to acquire data about motion and implemented a signal processing algorithm for performance enhancement. Based on the signal processing result, the performance was verified through feature extraction and learning of motion.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.15 no.11
• /
• pp.2365-2373
• /
• 2011

This paper propose the hardware IP of real-time face detection system for mobile devices and digital cameras required for high speed, smaller size and lower power. The proposed face detection system is robust against illumination changes, face size, and various face angles as the main cause of the face detection performance. Input image is transformed to LBP(Local Binary Pattern) image to obtain face characteristics robust against illumination changes, and detected the face using face feature data that was adopted to learn and generate in the various face angles using the Adaboost algorithm. The proposed face detection system can be detected maximum 36 faces at the input image size of QVGA($320{\times}240$), and designed by Verilog-HDL. Also, it was verified hardware implementation by using Virtex5 XC5VLX330 FPGA board and HD CMOS image sensor(CIS) for FPGA verification.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.36 no.2B
• /
• pp.131-140
• /
• 2011

As the number of Internet users and applications is increasing, the importance of application traffic classification is growing more and more for efficient network management. While a number of methods for traffic classification have been introduced, such as signature-based and machine learning-based methods, Skype application, which uses encrypted communication on its own P2P network, is known as one of the most difficult traffic to identify. In this paper we propose a novel method to identify Skype application traffic on the fly. The main idea is to setup a list of Skype host information {IP, port} by examining the packets generated in the Skype login process and utilizes the list to identify other Skype traffic. By implementing the identification system and deploying it on our campus network, we proved the performance and feasibility of the proposed method.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• v.9 no.1
• /
• pp.1149-1152
• /
• 2005

Detecting faces in images with complex backgrounds is a difficult task. Our approach, which obtains state of the art results, is based on a generative neural network model: the Constrained Generative Model (CGM). To detect side view faces and to decrease the number of false alarms, a conditional mixture of networks is used. To decrease the computational time cost, a fast search algorithm is proposed. The level of performance reached, in terms of detection accuracy and processing time, allows to apply this detector to a real word application: the indexation of face images on the Web.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.35 no.6
• /
• pp.375-380
• /
• 2022

High-speed railway bridges carry a risk of dynamic response amplification due to resonance caused by train loads, and running safety and riding comfort must therefore be reviewed through dynamic analysis in accordance with design codes. The running safety and ride comfort calculation procedure, however, is time consuming and expensive because dynamic analyses must be performed for every 10 km/h interval up to 110% of the design speed, including the critical speed for each train type. In this paper, a deep-learning-based prediction system that can predict the running safety and ride comfort in advance is proposed. The system does not use dynamic analysis but employs a deep learning algorithm. The proposed system is based on a neural network trained on the dynamic analysis results of each train and speed of the railway bridge and can predict the running safety and ride comfort according to input parameters such as train speed and bridge characteristics. To confirm the performance of the proposed system, running safety and riding comfort are predicted for a single span, straight simple beam bridge. Our results confirm that the deck vertical displacement and deck vertical acceleration for calculating running safety and riding comfort can be predicted with high accuracy.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.4
• /
• pp.125-132
• /
• 2021

As the layers of artificial neural network deepens, and the dimension of data used as an input increases, there is a problem of high arithmetic operation requiring a lot of arithmetic operation at a high speed in the learning and recognition of the neural network (NN). Thus, this study proposes a data dimensionality reduction method to reduce the dimension of the input data in the NN. The proposed Line-segment Feature Analysis (LFA) algorithm applies a gradient-based edge detection algorithm using median filters to analyze the line-segment features of the objects existing in an image. Concerning the extracted edge image, the eigenvalues corresponding to eight kinds of line-segment are calculated, using 3×3 or 5×5-sized detection filters consisting of the coefficient values, including [0, 1, 2, 4, 8, 16, 32, 64, and 128]. Two one-dimensional 256-sized data are produced, accumulating the same response values from the eigenvalue calculated with each detection filter, and the two data elements are added up. Two LFA256 data are merged to produce 512-sized LAF512 data. For the performance evaluation of the proposed LFA algorithm to reduce the data dimension for the recognition of handwritten numbers, as a result of a comparative experiment, using the PCA technique and AlexNet model, LFA256 and LFA512 showed a recognition performance respectively of 98.7% and 99%.

• Journal of IKEEE
• /
• v.17 no.4
• /
• pp.469-477
• /
• 2013

This paper enhances the speed of a pedestrian detection using an estimation of feature information based on integral image. Pedestrian model or input image should be resized to the size of various pedestrians. In case that the size of pedestrian model would be changed, pedestrian models with respect to the size of pedestrians should be required. Reducing the size of pedestrian model, however, deteriorates the quality of the model information. Since various features according to the size of pedestrian models should be extracted, repetitive feature extractions spend the most time in overall process of pedestrian detection. In order to enhance the processing time of feature extraction, this paper proposes the fast extraction of pedestrian features based on the estimate of integral image. The efficiency of the proposed method is evaluated by comparative experiments with the Channel Feature and Adaboost training using INRIA person dataset.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.39 no.8
• /
• pp.801-806
• /
• 2015

Journal bearings support rotors using fluid film between the rotor and the stator. Generally, journal bearings are used in large rotor systems such as turbines in a power plant, because even in high-speed and load conditions, journal bearing systems run in a stable condition. To enhance the reliability of journal-bearing systems, in this paper, we study health-diagnosis algorithms that are based on the supervised learning method. Specifically, this paper focused on defining the unit of features, while other previous papers have focused on defining various features of vibration signals. We evaluate the features of various lengths or units on the separable ability basis. From our results, we find that one cycle datum in the time-domain and 60 cycle datum in the frequency domain are the optimal datum units for real-time journal-bearing diagnosis systems.

• Journal of the Korean Society for Railway
• /
• v.20 no.2
• /
• pp.223-233
• /
• 2017

The railway turnout system is one of the most important systems that set train routes. Turnout system integrity should be guaranteed for robust train operation. To diagnose the turnout system status, LVDT and accelerometers are installed on a turnout system in a high speed line. The LVDT and accelerometers produce signals containing physical meaning of the turnout systems. The LVDT produces the displacement of the rail gauge and vibration when point moving or a train passes on turnout systems and the accelerometer produces impact forces induced by wheel sets. We performed data extraction from the measured signals and parameterized the extracted signals into meaningful quantities. The parameters are used for classifying whether the turnout status is normal. We proposed two methods for the classification, one uses probabilistic distribution and the other artificial neuron networks. The probabilistic distribution is used for the parameter being classified by the quantities and the artificial neuron networks for the form classification. Finally, we show how to learn the normal status of a turnout system.