• Journal of KIISE:Databases
• /
• v.30 no.5
• /
• pp.494-506
• /
• 2003

Recently, explosive popularity of multimedia information has triggered the need for retrieving multimedia contents efficiently from the database including audio, video and images. In this paper, we propose an XML-based retrieval scheme and a data model that complement the weak aspects of annotation and conent based retrieval methods. The Property and hierarchy structure of image and video data are represented and manipulated based on the Multimedia Description Schema (MDS) that conforms to the MPEG-7 standard. For audio contents, pitch contours extracted from their acoustic features are converted into UDR string. Especially, to improve the retrieval performance, user's access pattern and frequency are utilized in the construction of an index. We have implemented a prototype system and evaluated its performance through various experiments.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 2006.06a
• /
• pp.505-506
• /
• 2006

PD(partial discharge) occurred from variable PD sources in air may be the cause of breakdown in high voltage equipment which affect huge outage in power system. Identification and localization of PD sources is very important for engineer to cope with huge accident beforhand. PD phenomena can be detected by acoustic emission sensor or electromagnetic sensor like antenna. This paper has investigated the identification method using PCA(principal component analysis) for the PD signals from variable PD sources, for which the electric field distribution and PD inception voltages were simulated by using commercial FEM program. PD signals was detected by ultra wideband antenna. Their own features were extracted as the frequency coefficients transformed with FFT(fast fourier transform) and used to obtain independent pincipal components of each PD signals.

• Journal of Mechanical Science and Technology
• /
• v.19 no.6
• /
• pp.1366-1377
• /
• 2005

Experimental study was carried out in an atmospheric pressure, laboratory-scale dump combustor showing features of combustion instabilities. Flame structure and heat release rates were obtained from OH emission spectroscopy. Qualitative comparisons were made between line-integrated OH chemiluminescence image and Abel-transformed one. Local Rayleigh index distributions were also examined. Mean temperature, normalized standard deviation and temperature fluctuations were measured by coherent anti-Stokes Raman spectroscopy (CARS). To see the periodic behavior of oscillating flames, phase-resolved measurements were performed with respect to the pressure wave in the combustor. Results on system damping and driving characteristics were provided as a function of equivalence ratio. It also could be observed that phase resolved temperatures have been changed in a well-defined manner, while its difference between maximum and minimum reached up to 280K. These results would be expected to play an important role in better understanding of driving mechanisms and thermo-acoustic interactions.

• Journal of Information Processing Systems
• /
• v.13 no.5
• /
• pp.1135-1148
• /
• 2017

Speech recognition is one of the fascinating fields in the area of Computer science. Accuracy of speech recognition system may reduce due to the presence of noise present in speech signal. Therefore noise removal is an essential step in Automatic Speech Recognition (ASR) system and this paper proposes a new technique called combined thresholding for noise removal. Feature extraction is process of converting acoustic signal into most valuable set of parameters. This paper also concentrates on improving Mel Frequency Cepstral Coefficients (MFCC) features by introducing Discrete Wavelet Packet Transform (DWPT) in the place of Discrete Fourier Transformation (DFT) block to provide an efficient signal analysis. The feature vector is varied in size, for choosing the correct length of feature vector Self Organizing Map (SOM) is used. As a single classifier does not provide enough accuracy, so this research proposes an Ensemble Support Vector Machine (ESVM) classifier where the fixed length feature vector from SOM is given as input, termed as ESVM_SOM. The experimental results showed that the proposed methods provide better results than the existing methods.

• Journal of Korean language education
• /
• v.28 no.1
• /
• pp.1-21
• /
• 2017

The purpose of this study is to investigate the aspect of production and perception of Korean vowels /${\Lambda}$/ and /o/ and to discuss the correlation between production and perception of the two vowels. For this purpose, two separate experiments were conducted. 19 Chinese learners and 20 Korean native speakers produced Korean vowels /${\Lambda}$/ and /o/. Production experiments indicated that Koreans and Chinese female groups revealed common features in production, showing that they all pronounced /${\Lambda}$/ and /o/ in a distinguishable manner in the acoustic space. On the other hand, the Chinese male group failed to show that they could pronounce two vowels distinctively. The Chinese male group seemed to be confused in vowel height between the two vowels. A perception experiment was carried out on a continuum consisting of 11 synthesized stimuli. The perceptual judgment from referred Chinese and Korean subjects showed that Koreans and Chinese female groups had the same phonological boundaries (stimulus '04') for the two vowels on the continuum. However, the Chinese male group made perceptual criterion on stimulus '03'. These results confirmed that there was strong correlation between the aspect of production and perception.

• Honam Mathematical Journal
• /
• v.41 no.2
• /
• pp.403-420
• /
• 2019

The Helmholtz equation represents acoustic or electromagnetic scattering phenomena. The Method of Lines are known to have many advantages in simulation of forward and inverse scattering problems due to the usage of angle rays and Bessel functions. However, the method does not account for the jump phenomena on obstacle boundary and the approximation includes many high order Bessel functions. The high order Bessel functions have extreme blow-up or die-out features in resonance region obstacle boundary. Therefore, in particular, when we consider shape reconstruction problems, the method is suffered from severe instabilities due to the logical confliction and the severe singularities of high order Bessel functions. In this paper, two approximation formulas for the Helmholtz equation are introduced. The formulas are new and powerful. The derivation is based on Method of Lines, Huygen's principle, boundary jump relations, Addition Formula, and the orthogonality of the trigonometric functions. The formulas reduce the approximation dimension significantly so that only lower order Bessel functions are required. They overcome the severe instability near the obstacle boundary and reduce the computational time significantly. The convergence is exponential. The formulas adopt the scattering jump phenomena on the boundary, and separate the boundary information from the measured scattered fields. Thus, the sensitivities of the scattered fields caused by the boundary changes can be analyzed easily. Several numerical experiments are performed. The results show the superiority of the proposed formulas in accuracy, efficiency, and stability.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.24 no.1
• /
• pp.22-30
• /
• 2021

The ship wake generated by rotation of the propeller yields changes of characteristics of sound wave such as attenuation and scattering. To develope a battle field environment simulator for military purposes, it is very important to understand acoustical properties of ship wake. Existing research results have limitations in direct application because they performed under simple conditions or model ships were applied. In this study, we developed a ship wake generation model based on the ship's geometric wake distribution theory. The model can provide spatial distribution and void fraction with various marine environments as well as ship size. Through the developed model, geometric distribution features of ship wake according to the ship's maneuvering conditions were successfully simulated. In addition, changes of the bubble void fraction with time at any location within the battle field environment were identified. Therefore, the developed model is expected to be used in the development of a simulator to measure the acoustic characteristics of the ship wake.

• Smart Structures and Systems
• /
• v.29 no.3
• /
• pp.475-484
• /
• 2022

The loss of cable tension for civil infrastructure reduces structural bearing capacity and causes harmful deformation of structures. Currently, most of the structural health monitoring (SHM) approaches for cables rely on contact transducers. This paper proposes a cable tension identification technology using percussion sound, which provides a fast determination of steel cable tension without physical contact between cables and sensors. Notably, inspired by the concept of tensioning strings for piano tuning, this proposed technology predicts cable tension value by deep learning assisted classification of "percussion" sound from tapping a steel cable. To simulate the non-linear mapping of human ears to sound and to better quantify the minor changes in the high-frequency bands of the sound spectrum generated by percussions, Mel-frequency cepstral coefficients (MFCCs) were extracted as acoustic features to train the deep learning network. A convolutional neural network (CNN) with four convolutional layers and two global pooling layers was employed to identify the cable tension in a certain designed range. Moreover, theoretical and finite element methods (FEM) were conducted to prove the feasibility of the proposed technology. Finally, the identification performance of the proposed technology was experimentally investigated. Overall, results show that the proposed percussion-based technology has great potentials for estimating cable tension for in-situ structural safety assessment.

• Journal of the Korean Society of Fisheries and Ocean Technology
• /
• v.58 no.4
• /
• pp.352-358
• /
• 2022

We explored the frequency response of krill target strength (TS) to understand the Antarctic krill (Euphausia superba) and ice krill (Euphausia crystallorophias) using the stochastic distorted-wave Born approximation (SDWBA) model. The results showed that the distribution of orientation and the fatness factor could significantly impact on the frequency response of TS. Krill TS is clearly depended on acoustic properties, which could affect to estimate the biomass of two krill species. The results provide insight into the importance of understanding TS variation to estimate the Antarctic krill and ice krill biomass, and their ecology related to the environmental features in the Southern Ocean.

• The Journal of the Acoustical Society of Korea
• /
• v.43 no.2
• /
• pp.234-242
• /
• 2024

We propose a passive sonar signal classification algorithm using Graph Neural Network (GNN). The proposed algorithm segments spectrograms into image patches and represents graphs through connections between adjacent image patches. Subsequently, Graph Convolutional Network (GCN) is trained using the represented graphs to classify signals. In experiments with publicly available underwater acoustic data, the proposed algorithm represents the line frequency features of spectrograms in graph form, achieving an impressive classification accuracy of 92.50 %. This result demonstrates a 8.15 % higher classification accuracy compared to conventional Convolutional Neural Network (CNN).