• Proceedings of the Korea Multimedia Society Conference
• /
• 2012.05a
• /
• pp.340-342
• /
• 2012

Integral Imaging (II) is an attractive technique for three-dimensional (3D) image, video display and recording. Inherently, the high resolution II requires an enormous amount of data for storing and transmitting of 3D scenes. Compression techniques attempt to evade this issue. In this study, we made a comparative performance analysis of popular transforming/compression techniques such as the Discrete Cosine Transform (DCT) and the Discrete Wavelet Transform (DWT) in order to compress 3D-II. The standard baseline JPEG (Joint Photographic Experts Group) using DCT and JPEG 2000 using DWT methods were manipulated in our experiments. In our analysis, we have shown that the DWT based JPEG 2000 compression methodology could be a good alternative for 3D-II.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2005.11a
• /
• pp.107-111
• /
• 2005

It has been reviewed whether it would be suitable that the application of the time-frequency signal analysis techniques to estimate the location of the impact source in plate structure. The STFT(Short Time Fourier Transform), WVD(Wigner-Ville distribution) and CWT(Continuous Wavelet Transform) methods are introduced and the advantages and disadvantages of those methods are described by using a simulated signal component. The essential of the above proposed techniques is to separate the traveling waves in both time and frequency domains using the dispersion characteristics of the structural waves. These time-frequency methods are expected to be more useful than the conventional time domain analyses fer the impact localization problem on a plate type structure. Also it has been concluded that the smoothed WVD can give more reliable means than the other methodologies for the location estimation in a noisy environment.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 1999.10a
• /
• pp.365-373
• /
• 1999

Recently, numerous studies have demonstrated that artificial intelligence such as neural networks can be an alternative methodology for classification problems to which traditional statistical methods have long been applied. In building neural network model, the selection of independent and dependent variables should be approached with great care and should be treated as a model construction process. Irrespective of the efficiency of a learning procedure in terms of convergence, generalization and stability, the ultimate performance of the estimator will depend on the relevance of the selected input variables and the quality of the data used. Approaches developed in statistical methods such as correlation analysis and stepwise selection method are often very useful. These methods, however, may not be the optimal ones for the development of neural network models. In this paper, we propose a genetic algorithms approach to find an optimal or near optimal input variables for neural network modeling. The proposed approach is demonstrated by applications to bankruptcy prediction modeling. Our experimental results show that this approach increases overall classification accuracy rate significantly.

• ETRI Journal
• /
• v.42 no.1
• /
• pp.78-89
• /
• 2020

Human activity recognition (HAR) has become effective as a computer vision tool for video surveillance systems. In this paper, a novel biometric system that can detect human activities in 3D space is proposed. In order to implement HAR, joint angles obtained using an RGB-depth sensor are used as features. Because HAR is operated in the time domain, angle information is stored using the sliding kernel method. Haar-wavelet transform (HWT) is applied to preserve the information of the features before reducing the data dimension. Dimension reduction using an averaging algorithm is also applied to decrease the computational cost, which provides faster performance while maintaining high accuracy. Before the classification, a proposed thresholding method with inverse HWT is conducted to extract the final feature set. Finally, the K-nearest neighbor (k-NN) algorithm is used to recognize the activity with respect to the given data. The method compares favorably with the results using other machine learning algorithms.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.10 no.4
• /
• pp.521-526
• /
• 2015

Research on how to use the biological characteristics of human to confirm the identity of the individual is being actively conducted. Electrocardiogram(: ECG) based biometric system is difficult to counterfeit and does not cause skin irritation on the subject. It can be easily combined with conventional biometrics such as fingerprint and face recognition to give multimodal biometric systems. In this thesis, biometric identification method analysing ECG waveform characteristics from Discrete Wavelet Transform(DWT) coefficients is suggested. Feature selection is performed on the 9 coefficients of DWT using the correlation analysis. The verification is achieved by using the error back propagation neural networks. Using the proposed approach on 24 subjects of MIT-BIH QT Database, 98.88% verification rate has been obtained.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2006.05a
• /
• pp.1465-1469
• /
• 2006

수문순환 과정은 기상현상과 밀접한 관련을 가지고 서로 연관되어 있다. 이러한 연관성을 규명하여 수자원관리에 위험도를 감소시키려는 노력은 많은 분야에서 이루어지고 있으며, 주요 연구 주제가 되고 있다. 이러한 기상현상 중에서 가뭄은 여러 가지 요소가 복합되어 발생되는 것으로 알려지고 있으나 이를 설명하기에는 여전히 부족한 면이 존재한다. 가뭄을 발생시키는 몇 가지 가능한 원인으로는 E1 Nino-Southern Oscillation(ENSO)현상으로 잘 알려져 있는 비정상적인 해수면 온도의 변화나 기후 시스템의 비선형적 거동을 들 수 있다. 특히, 기후 시스템은 대개 경년 변화(inter-annual variability) 및 10년 이상의 주기(decadal variability) 특성을 가지고 있으며 가뭄 또한 경년변화의 주기 특성을 나타내고 있는 것으로 알려지고 있다. 이러한 관점에서 수문시계열을 특정 주파수(frequency)에서 고립시킨 후, 분석이 가능한 분해방법(decomposition method)을 통해 보다 해석적으로 접근하는 것이 가능하다. 이를 위해 본 연구에서는 Wavelet Transform분석을 도입하였으며 통계적으로 유의한 성분을 시계열로부터 추출하여 가뭄과 기상인자와의 변동성 분석을 실시하였다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.9 no.1
• /
• pp.8-15
• /
• 2005

The analysis of Rader signal that have non-linearity variable phase is signal that contact easily in several fields such as radar, telecommunication, seismic, sonar and biomedical applications. In generally, Non-stationary signal means that spectral characteristics are varying with time and instantaneous frequency is only one frequency or narrow range of frequencies varying as a function of time. Therefore, Instantaneous frequency is vary important variable that understanding physical characteristic of signal. This paper was describes continuous wavelet transform to determine instantaneous frequency at non-staionary signal and compare to existing method. When white noise or various frequency is overlapped each other in sign, existing method was can not decide corrected instantaneous frequency, but when used continuous wavelet transform, very well decide correctly frequency regardless of component of signal.

• Transactions of the Korean Society of Machine Tool Engineers
• /
• v.14 no.3
• /
• pp.96-102
• /
• 2005

Acoustic Emission(AE) technique has been applied to not only material characterization evaluation but also on-line monitoring of the structural integrity. The AE source location technique is very important to identify the source, such as crack, leak detection. Since the AE waveforms obtained from sensors are very difficult to distinguish the defect signals, therefore, it is necessary to consider the signal analysis of the transient wave-form. In this study, we have divided the region of interest into a set finite elements, and calculated the arrival time differences between sensors by using the velocities at every degree from 0 to 90. A new technique for the source location of acoustic emission in fiberboard plates has been studied by introducing Wavelet Transform(WT) do-noising technique. WT is a powerful tool for processing transient signals with temporally varying spectra. If the WT de-noising was employed, we could successfully filter out the errors of source location in fiberboard plates by arrival time difference method. The accuracy of source location appeared to be significantly improved.

• Journal of Korea Multimedia Society
• /
• v.22 no.11
• /
• pp.1233-1241
• /
• 2019

The goal of the sound-based mechanical fault diagnosis technique is to automatically find abnormal signals in the machine using acoustic emission. Conventional methods of using mathematical models have been found to be inaccurate due to the complexity of industrial mechanical systems and the existence of nonlinear factors such as noise. Therefore, any fault diagnosis issue can be treated as a pattern recognition problem. We propose an automatic fault diagnosis method using discrete wavelet transform and power spectrum density using multi pattern recognition. First, we perform DWT-based filtering analysis for noise cancelling and effective feature extraction. Next, the power spectral density(PSD) is performed on each subband of the DWT in order to effectively extract feature vectors of sound. Finally, each PSD data is extracted with the features of the classifier using multi pattern recognition. The results show that the proposed method can not only be used effectively to detect faults as well as apply to various automatic diagnosis system based on sound.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.51 no.1
• /
• pp.32-38
• /
• 2002

This paper describes a development of a portable physical activity monitoring system using two accelerometers to quantify physical activity. The system hardware consists of two piezoresistive accelerometers, amplifiers with gain of 30, lowpass filters with cut-off frequency of 15Hz, offset control circuits, one-chip microcontroller and flash memory card. In order to evaluate the performance of the system we acquired 3 channel data at 32 sample/sec from body-fixed accelerometers in chest and right upper leg. And then the acquired data were processed by MatLab on personal computer. We tried to distinguish not only fundamental actions which are steady-state activities such as standing, sitting, and lying but also dynamic activities with walking, up a stairway, down a stairway, and running. Five subjects participated the evaluation process which compare the video data with the measured data. As a result, the activity classification rate of 90.6% on average was obtained. Overall results showed that the steady-state activities could be classified from the low component of 3-axis acceleration signal and dynamic activities could be distinguished from frequency analysis using wavelet transform and FFT. Finally, we could find that this system can be applied to acquire and analyze the static and dynamic physical activity data.