• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2013년도 춘계학술발표대회
• /
• pp.204-207
• /
• 2013

Music genre classification is an essential component for music information retrieval system. There are two important components to be considered for better genre classification, which are audio feature extraction and classifier. This paper incorporates two different kinds of features for genre classification, timbral texture and rhythmic content features. Timbral texture contains several spectral and Mel-frequency Cepstral Coefficient (MFCC) features. Before choosing a timbral feature we explore which feature contributes less significant role on genre discrimination. This facilitates the reduction of feature dimension. For the timbral features up to the 4-th order central moments and the covariance components of mutual features are considered to improve the overall classification result. For the rhythmic content the features extracted from beat histogram are selected. In the paper Extreme Learning Machine (ELM) with bagging is used as classifier for classifying the genres. Based on the proposed feature sets and classifier, experiment is performed with well-known datasets: GTZAN databases with ten different music genres, respectively. The proposed method acquires the better classification accuracy than the existing approaches.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제12권2호
• /
• pp.676-692
• /
• 2018

Saliency detection in neurobiology is a vehement research during the last few years, several cognitive and interactive systems are designed to simulate saliency model (an attentional mechanism, which focuses on the worthiest part in the image). In this paper, a bottom up saliency detection model is proposed by taking into account the color and luminance frequency features of RGB, CIE $L^*a^*b^*$ color space of the image. We employ low-level features of image and apply band pass filter to estimate and highlight salient region. We compute the likelihood probability by applying Bayesian framework at pixels. Experiments on two publically available datasets (MSRA and SED2) show that our saliency model performs better as compared to the ten state of the art algorithms by achieving higher precision, better recall and F-Measure.

• 한국정밀공학회지
• /
• 제12권9호
• /
• pp.86-94
• /
• 1995

Bead shape in high frequency electric resistance (HER) pipe welding gives useful information on judging current welding conditon. In most welding process, heat input is controlled by skilled operators observing color and shape of bead. In this paper, a visual monitoring system is designed to observe bead shape in HERW pipe welding process by using structured light beam and a C.I.D(Charge injection device) camera. To avoid some difficul- ties arising in extracting stable features of stripe pattern and classifying the extracted features, Kohonen neural network is used to classify such bead shapes. The experimental results show accurate classification performance of the proposed method.

• 한국전산유체공학회:학술대회논문집
• /
• 한국전산유체공학회 2004년도 춘계 학술대회논문집
• /
• pp.154-159
• /
• 2004

Numerical simulations of supersonic impinging jet flows are carried out using the axisymmetric Navier-Stokes code. This paper focuses on the oscillatory flow features associated with the variation of the nozzle pressure ratio and nozzle-to-plate distance. Frequencies of the surface pressure oscillation from computational results are in accord with the measured impinging tones for various cases of nozzle-to-plate distance. The variation of this frequency with distance show a staging behavior. Computed results for the case of nozzle pressure ratio variation for a fixed nozzle-to-plate distance also demonstrate a staging behavior. These two seemingly different staging behaviors are found to obey the same frequency-distance characteristics when the frequency and the distance are normalized by using the length of the shock cell.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2008년도 제39회 하계학술대회
• /
• pp.2025-2026
• /
• 2008

This study aims to analyze the features of malfunction of an SFC that helps maintain the frequency of static rotation during the operation of a gas turbine in power plants where power generators and controllers have such a complicated structure as a nerve system, by using electric interference with peripheral devices and maintaining acquired data efficiently. Also, in order to track the possibility of malfunction by various surges and noises which may occur in the process of inserting an SFC during the operation of a gas turbine and to prepare protective measures for the possibility, the study intends to offer data for developing a surge protector suited to the features of a section that is supposed to incur the possibility of malfunction.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2003년도 춘계학술대회 논문집 전기기기 및 에너지변환시스템부문
• /
• pp.200-207
• /
• 2003

Today we have excellent motor drive system using high frequency carrier PWM control voltage source inverter with vector control strategy. In the other hand, we have met serious problems caused by high frequency switching. PWM Cyclo-converter called Matrix converter is expected as the new strategy possible to improve these problems and add some more convenient features suitable for new drive system with system integration. In this paper, we will introduce the background of this development and features of this converter from our research, additionally introduce remarkable active promotions for this converter as a survey.

• 한국음향학회지
• /
• 제10권4호
• /
• pp.38-47
• /
• 1991

음성인식에 있어서 문제점의 하나는 발성자에 따른 주파수 변동문제이다. 본 논문에서는 이러한 음성 신호의 주파수 특성의 변동에 따른 영향을 해결하기 위하여 fuzzy 이론을 도입하였다. 여기서 표준패턴은 음성신호의 대표적인 특징들을 포함하고 있어야 하므로, 먼저 여러 화자가 발성한 단어들을 벡터 양자화한 코드북을 생성하였으며, 이코드북으로 부터 추출한 피크 주파수와 피크 에너지를 fuzzy화 패턴으로 작성하였다. 입력 음성신호로 부터 추출한 특징량인 스펙트럼의 피크 주파수와 피크에너지를 각각 멤버쉽 함수로 표현하여 fuzzy 추론에 의한 단어인식을 하였다. 실제 확신도 계산에 있어서는 계산량을 줄이기 위하여 fuzzy 값의 차만으로 확신도를 구하는 개선 확신도를 제안하여 사용하였다. 한국어 숫자음을 인식 실험한 결과 주파수 특성의 변동에 따른 영향을 해결할 수 있음을 확인하였으며, 제안된 개선 확신도 계산방법에 의해서 기억용량과 계산량을 감소 시킬 수 있었다.

• 음성과학
• /
• 제9권2호
• /
• pp.59-68
• /
• 2002

This paper analyzes the prosodic features distinctively associated with Korean topic constructions (marked by nun or its variant un) and subject constructions (marked by ka or its variant i) as a way of explicating the role that prosody plays in differentially constituting their discourse messages. Using both spoken data elicited in controlled settings and spontaneous conversational data, an attempt is made to identify differentiating prosodic features and intonation contours associated with distinct meanings and functions of nun- and ka-constructions evoked in a variety of discourse contexts.

• International Journal of Railway
• /
• 제1권3호
• /
• pp.94-98
• /
• 2008

Static inverters are essential devices onboard of rolling stock. State-of-the-art static inverters have an impact on both rail infrastructure and vehicle maintenance due to their new topology with new features. The paper describes two important aspects as examples of new features available in state-of-the-art static inverters: active input current control and the effects on the rail infrastructure as well as the detection of the state of charge and the state of health of batteries to simplify vehicle maintenance.

• 한국음향학회지
• /
• 제29권6호
• /
• pp.381-387
• /
• 2010

다채널 오디오 코딩 시스템을 평가함에 있어서 음색 요소뿐만 아니라 공간감 요소 역시 큰 중요성을 갖는다. 이러한 이유로 양이 시간차 왜곡 (Interaural Time Difference Distortion, ITDDist), 양이 크기차 왜곡 (Interaural Level Difference Distortion, ILDDist), 양이 상관관계 왜곡 (Interaural Cross Correlation Distortion, IACCDist)과 같은 공간감 요소를 추가하여 기존의 음질 객관 평가 권고안인 ITU-R Rec. BS. 1387-1을 다채널 오디오 시스템에 적용하기 위한 시도가 있었다. 이 다채널 오디오 시스템을 위한 확장 모델에서는 Duplex 이론에 따라, 양이 시간차 왜곡은 1.5kHz 이하의 저주파 영역에 대해서만 계산되었으며, 양이 크기차 왜곡은 2.5kHz 이상의 고주파 영역에 대해서만 계산되었다. 하지만 고주파 영역에 있어서 포락선의 양이 시간차는 공간감 인지에 있어서, 특히 음상 정위에 있어서 중요한 역할을 한다. 이러한 고주파 영역 포락선의 양이 시간차 공간감 인지에 미치는 영향을 정량적으로 분석하기 위해 해당 인자를 계산하는 방법이 본 논문에서 소개된다. 또한, 이렇게 계산된 고주파 영역 포락선의 양이 시간차는 다채널 오디오의 주관평가 결과와 높은 상관관계를 가짐을 확인한다.