• ETRI Journal
• /
• 제30권1호
• /
• pp.89-100
• /
• 2008

We consider the feature recombination technique in a multiband approach to speaker identification and verification. To overcome the ineffectiveness of conventional feature recombination in broadband noisy environments, we propose a new subband feature recombination which uses subband likelihoods and a subband reliable-feature selection technique with an adaptive noise model. In the decision step of speaker recognition, a few very low unreliable feature likelihood scores can cause a speaker recognition system to make an incorrect decision. To overcome this problem, reliable-feature selection adjusts the likelihood scores of an unreliable feature by comparison with those of an adaptive noise model, which is estimated by the maximum a posteriori adaptation technique using noise features directly obtained from noisy test speech. To evaluate the effectiveness of the proposed methods in noisy environments, we use the TIMIT database and the NTIMIT database, which is the corresponding telephone version of TIMIT database. The proposed subband feature recombination with subband reliable-feature selection achieves better performance than the conventional feature recombination system with reliable-feature selection.

• 대한음성학회지:말소리
• /
• 제63호
• /
• pp.125-137
• /
• 2007

Recently, many techniques have been proposed to improve the noise robustness for speaker verification. In this paper, we consider the feature recombination technique in multi-band approach. In the conventional feature recombination for speaker verification, to compute the likelihoods of speaker models or universal background model, whole feature components are used. This computation method is not effective in a view point of multi-band approach. To deal with non-effectiveness of the conventional feature recombination technique, we introduce a subband likelihood computation, and propose a modified feature recombination using subband likelihoods. In decision step of speaker verification system in noise environments, a few very low likelihood scores of a speaker model or universal background model cause speaker verification system to make wrong decision. To overcome this problem, a reliable feature selection method is proposed. The low likelihood scores of unreliable feature are substituted by likelihood scores of the adaptive noise model. In here, this adaptive noise model is estimated by maximum a posteriori adaptation technique using noise features directly obtained from noisy test speech. The proposed method using subband-based reliable feature selection obtains better performance than conventional feature recombination system. The error reduction rate is more than 31 % compared with the feature recombination-based speaker verification system.

• 한국음향학회지
• /
• 제36권2호
• /
• pp.144-150
• /
• 2017

음악 장르는 음악 검색 및 분류 등의 정보 처리 시스템 구현에 있어서 필수적인 요소이다. 일반적으로 장르 분류를 위한 스펙트럼 특징은 음악의 화음 및 강약 구조를 표현하기 위해 부밴드로 분해하여 구해진다. 본 논문은 음악 장르 분류 성능 개선을 위한 특징 추출을 위한 부밴드 분해 방법에 관해 연구하였다. 또한 부밴드 음악 특징의 차수를 줄일 수 있는 방법에 대해서도 연구하였다. 널리 사용되고 있는 장르 데이터셋들에서 실험을 수행하여 널리 사용되고 있는 옥타브 스케일보다 세분화된 부밴드 분해가 장르 분류 성능을 향상시킬 수 있으며, 특징 차수 축소를 결합하여 분류기의 계산량도 줄일 수 있음을 보였다.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제4권3호
• /
• pp.322-326
• /
• 2004

This paper describes a new iris recognition method based on a shift-invariant wavelet sub-images. For the feature representation, we first preprocess an iris image for the compensation of the variation of the iris and for the easy implementation of the wavelet transform. Then, we decompose the preprocessed iris image into multiple subband images using a shift-invariant wavelet transform. For feature representation, we select a set of subband images, which have rich information for the classification of various iris patterns and robust to noises. In order to reduce the size of the feature vector, we quantize. each pixel of subband images using the Lloyd-Max quantization method Each feature element is represented by one of quantization levels, and a set of these feature element is the feature vector. When the quantization is very coarse, the quantized level does not have much information about the image pixel value. Therefore, we define a new similarity measure based on mutual information between two features. With this similarity measure, the size of the feature vector can be reduced without much degradation of performance. Experimentally, we show that the proposed method produced superb performance in iris recognition.

• 한국콘텐츠학회:학술대회논문집
• /
• 한국콘텐츠학회 2004년도 춘계 종합학술대회 논문집
• /
• pp.392-395
• /
• 2004

최근 보안 시스템 분야에서 컴퓨터 기술의 발전으로 얼굴 인식에 대한 관심이 높아지고 있다. 얼굴 인식은 기하학적 특징을 이용하는 방법과 통계적 특징을 이용하는 방법이 있다. 본 연구는 정면 얼굴에 대한 대수적인 방법이다. 제안 방식은, 웨이브릿 변환을 통한 k 단계의 LL, LH, HL 부대역을 구하고, 이를 PCA/LDA를 적용하여 얼굴 인식을 하였다. 전체 영상에 대한 얼굴 인식률에 비해 웨이브릿 변환을 이용한 부대역 영상에 대한 얼굴 인식률이 더 좋음을 보여준다.

• 한국음향학회지
• /
• 제28권4호
• /
• pp.375-382
• /
• 2009

본 논문에서는 압축도메인 특징을 이용한 오디오 핑거프린팅 방법을 제안하였다. 압축도메인을 이용함으로써 계산량과 시간을 크게 줄일 수 있는 장점이 있다. 특히 오디오 압축에 널리 쓰이고 있는 MDCT 도메인을 이용하였으며, MDCT 도메인을 부밴드로 나누고 대표적인 모멘트 특징인 에너지, 무게중심, 평탄도로 부터 각각 핑거프린트를 얻었다. 추출된 특징을 차분 필터링하고 부호를 취하여 이진 핑거프린트를 얻었다. 실험을 통해서 고려한 MDCT 도메인 특징들로부터 얻은 핑거프린트들의 인식 성능을 비교하였다. 수 천곡 규모의 오디오에 대해서 다양한 변환에 대한 인식 성능을 고려하였으며, 실험결과 부밴드 에너지가 가장 우수한 핑거프린팅 성능을 보였다.

• 음성과학
• /
• 제14권2호
• /
• pp.127-135
• /
• 2007

Feature vectors of speech are generally extracted from whole frequency domain. The inherent character of a speaker is located in the low band or high band frequency. However, if the speech is corrupted by narrowband noise with concentrated energy, speaker verification performance is reduced as the individual characteristic is removed. In this paper, we propose a PCA Covariance Model based on the multiband to extract the robust feature vectors against the narrowband noise. First, we divide the overall frequency band into several subbands. Second, the correlation of feature vectors extracted independently from each subband is removed by PCA. The distance obtained from each subband has different distribution. To normalize against the different distribution, we moved the value into the normalized distribution through the mapping function. Finally, the represented value applying the weighting function is used for speaker verification. In the experiments, the proposed method shows better performance of the speaker verification and reduces the computation.

• 한국소음진동공학회논문집
• /
• 제16권1호
• /
• pp.27-33
• /
• 2006

Fault diagnosis and condition monitoring for rotating machines are important for efficiency and accident prevention. The process of fault diagnosis is to extract the feature of signals and to classify each state. Conventionally, fault diagnosis has been developed by combining signal processing techniques for spectral analysis and pattern recognition, however these methods are not able to diagnose correctly for certain rotating machines and some faulty phenomena. In this paper, we add a minimum detection error algorithm to the previous method to reduce detection error rate. Vibration signals of the induction motor are measured and divided into subband signals. Each subband signal is processed to obtain the RMS, standard deviation and the statistic data for constructing the feature extraction vectors. We make a study of the fault diagnosis system that the feature extraction vectors are applied to K-means clustering algorithm and minimum detection error algorithm.

• ETRI Journal
• /
• 제23권4호
• /
• pp.190-198
• /
• 2001

The performance of Motion Compensated Discrete Cosine Transform (MC-DCT) video coding is improved by using the region adaptive subband image coding [18]. On the assumption that the video is acquired from the camera on a moving platform and the distance between the camera and the scene is large enough, both the motion of camera and the motion of moving objects in a frame are compensated. For the compensation of camera motion, a feature matching algorithm is employed. Several feature points extracted using a Sobel operator are used to compensate the camera motion of translation, rotation, and zoom. The illumination change between frames is also compensated. Motion compensated frame differences are divided into three regions called stationary background, moving objects, and newly emerging areas each of which is arbitrarily shaped. Different quantizers are used for different regions. Compared to the conventional MC-DCT video coding using block matching algorithm, our video coding scheme shows about 1.0-dB improvements on average for the experimental video samples.

• Communications for Statistical Applications and Methods
• /
• 제10권2호
• /
• pp.457-469
• /
• 2003

In this paper, we have proposed the efficient algorithm that can segment the video scene change using a various statistical characteristics obtained from by applying the wavelet transformation for each frames. Our method firstly extracts the histogram features from low frequency subband of wavelet-transformed image and then uses these features to detect the abrupt scene change. Second, it extracts the edge information from applying the mesh method to the high frequency subband of transformed image. We quantify the extracted edge information as the values of variance characteristic of each pixel and use these values to detect the gradual scene change. And we have also proposed an algorithm how extract the proper key frame from segmented video scene. Experiment results show that the proposed method is both very efficient algorithm in segmenting video frames and also is to become the appropriate key frame extraction method.