• The Journal of the Acoustical Society of Korea
• /
• v.20 no.7
• /
• pp.83-88
• /
• 2001

In this paper, matched field processing algorithm robust to environmental mismatches in an ocean waveguide based on feature extraction is summarized. However, in applying this processor to localize a source there are two preliminary issues to be resolved. One is the number of eigenvectors to be extracted and the other is the number of environmental samples to be used. To determine these issues, the relation between the number of dominant modes propagating in a given ocean waveguide and that of eigenvectors to be extracted is analyzed. Then, the analysis results are confirmed by the subspace analysis. This analysis quantifies the similarity between the subspace spanned by the signal vectors and that spanned by the eigenvectors to be extracted. The error index is defined as a relative difference between the location estimated by the current processor and the real source location. It is identified that in the case of extracting the largest eigenvectors equal to the number of dominant modes in a given environment, the processor localizes the source successfully. From the numerical simulations, it is shown that use of at least 30 environmental samples guarantee stable performance of the proposed processor.

• Proceedings of the KIEE Conference
• /
• 2009.07a
• /
• pp.1857_1858
• /
• 2009

본 논문에서는 PCA와 입자 군집 최적화 알고리즘을 이용한 얼굴이미지에서 특징선택 방법에 대하여 제안한다. 2차원 얼굴이미지의 히스토그램 분표값에서 정규화합 연산을 이용한 히스토그램 평활화 기법을 거쳐 대비효과를 주어 화질을 개선시켜 준다. PCA는 2차원 얼굴이미지를 이용하여 공분산 행렬을 구한 후 그것의 고유값에 따른 고유벡터를 구하여 얼굴인식에 사용될 특징 벡터들을 추출한다. 또한 추출된 특징벡터 중에서 얼굴인식 성능에 중요한 요소가 되는 특징 벡터들을 입자 군집 최적화 알고리즘을 이용하여 최적화한다. 다항식 기반 RBF 신경회로망을 사용하여 얼굴인식 성능을 평가한다. 본 논문에서 제안된 방법을 통해 최적화된 특징벡터와 얼굴인식률과의 관계를 알 수 있다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.39B no.10
• /
• pp.700-707
• /
• 2014

In this paper, we propose the noise reduction method by using the eigenfilter in cyclic prefix system based on SNR. To obtain the signal eigenvectors for the eigenfiltering, we propose a method of obtaining the autocorrelation matrix by exploiting the circulant property of the received block which results from the cyclic extension of the OFDM symbol. Since the structures of the transmitter and the receiver are not changed, the proposed method is easy to apply to the conventional OFDM system. To verify the proposed method, we evaluate the persistency of excitation (POE) criterion for the input and demonstrate the effectiveness of the proposed method in the simulation results.

• Proceedings of the Korean Information Science Society Conference
• /
• 2012.06b
• /
• pp.414-416
• /
• 2012

고차 주성분 방법으로는 텐서 분석이 있었다. Electroencephalography(EEG) 데이터나 Social Network 데이터에 텐서 분석이 적용되어 주요한 성분들을 찾는 연구들이 있었다. 그러나 텐서 분석은 직관적으로 이해하기에 어려움이 있으며 중요한 노드를 찾는데에는 다소 어려움이 있다. 본 논문에서는 고차 하이퍼에지로 이차원 행렬을 만들고 주성분분석법을 이용하여 중요한 노드를 찾는 새로운 방법론을 제시한다. 데이터로는 Multimodal Memory Game(MMG) 수행시 촬영한 EEG 데이터를 사용하였다. MMG는 TV 드라마 기반의 기억인출게임이다. 베타파의 Power Spectrum Density(PSD)는 각 위치의 채널들의 활성도를 나타내는 지표이다. 우리는 Random Sampling을 바탕으로 PSD 상위 50%의 채널들간의 전이행렬을 구하였다. 그 후 고유치와 고유벡터를 구하였다. 가장 큰 고유치의 고유벡터는 주성분을 나타내며 고유벡터의 각 원소들은 중요도를 나타내는 centrality 이다. 세 명의 피험자에 대한 centrality 상위 30개의 중요한 채널들을 구하였고 세명에 공통적으로 포함되는 채널을 확인하였다.

• Proceedings of the IEEK Conference
• /
• 2001.09a
• /
• pp.829-832
• /
• 2001

본 논문에서는 2차 통계치를 이용하여 패널추징 및 등화 기법을 제안하였다. 기존의 채널 추정 알고리듬은 잡음이 없는 환경에서 LS방법을 이용하기 때문에 잡음이 강한 패널에서는 원하는 성능을 얻을 수 없는 단점이 있다. 수신신호의 상관행렬의 최소 고유값에 대응하는 고유벡터는 채널의 임펄스 응답에 관한 정보를 포함하고 있다. 이러한 고유 벡터를 매시간마다 갱신시키면서 구하는 적응 알고리듬을 제안하고 이를 이용하여 블라인드 채널 추정 및 등화기 파라미터를 추정하였다. 제안한 알고리듬은 잡음에 강인한 특성을 보일 뿐 아니라 기존의 알고리듬들 보다 우수한 채널 추정 및 등화 성능을 모의 실험을 통하여 검증하였다.

• Journal of Korea Multimedia Society
• /
• v.6 no.3
• /
• pp.429-435
• /
• 2003

In this paper, we propose improved binary tree vector quantization with consideration of spatial sensitivity which is one of the human visual properties. We combine weights in consideration with the responsibility of human visual system according to changes of three primary color in blocks of images with the process of splitting nodes using eigenvector in binary tree vector quantization. Also we propose the novel quality measure of the quantization images that applies MTF(modulation transfer function) to luminance value of quantization error of color image. The test results show that the proposed method generates the quantized images with fine color and performs better than the conventional method in terms of clustering the similar regions. Also the proposed method can get less quantized level images and can reduce the resource occupied by the quantized image.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2007.06a
• /
• pp.597-600
• /
• 2007

Face is considered to be one of the biometrics in person identification. But Face recognition is a high dimensional pattern recognition problem. Even low-resolution face images generate huge dimensional feature space. The aim of this paper is to present a fast feature extraction method for real time human face recognition. first, It compute eigen-vector and eigen-value by Principle component analysis on inputed human face image, and propose method of feature extraction that make feature vector by apply gabor filter to computed eigen-vector. And it compute feature value which multiply by made eigen-value. This study simulations performed using the ORL Database.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.19 no.12
• /
• pp.2303-2312
• /
• 1994

A high resolution algorithm is presented for resolving multiple narrowband plane waves that are incident on an equispaced linear array. To overcome the deleterious effects due to coherent sources, a number of noise-eigenvector-based approaches have been proposed for narrowband signal processing. For differing reasons, each f these methods provide a less than satisfactory resolution of the coherency problem. The proposed algorithm makes use of fundamental property possessed by those eigenvectors of the spatial covariance matrix that are associated with eigenvalues that are larger than the sensor noise level. This property is then used to solve the incoherent and coherent sources incident on an equispaced linear array. Simulation results are shown to illustrate the high resolution performance achieved with this new approach relative to that obtained with MUSIC and spatial smoothed MUSIC.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.45 no.6
• /
• pp.67-76
• /
• 2008

Reliable face recognition under various illumination environments is essential for successful commercialization. Feature-based face recognition relies on a good choice of feature vectors. Gabor feature vectors are known to be more robust to variations of pose and illumination than any other feature vectors so that they are popularly adopted for face recognition. However, they are not completely independent of illuminations. In this paper, we propose an illumination-robust face recognition method based on the Gabor feature vector illumination PCA model. We first construct the Gabor feature vector illumination PCA model where Gator feature vector space is rendered to be decomposed into two orthogonal illumination subspace and face identity subspace. Since the Gabor feature vectors obtained by projection into the face identity subspace are separated from illumination, the face recognition utilizing them becomes more robust to illumination. Through experiments, it is shown that the proposed face recognition based on Gabor feature vector illumination PCA model performs more reliably under various illumination and Pose environments.

• Journal of Korea Multimedia Society
• /
• v.9 no.1
• /
• pp.33-40
• /
• 2006

This paper presents an efficient face recognition method using both first moment of image and eigenvector. First moment is a method for finding centroid of image, which is applied to exclude the needless backgrounds in the face recognitions by shitting to the centroid of face image. Eigenvector which are the basis images as face features, is extracted by principal component analysis(PCA). This is to improve the recognition performance by excluding the redundancy considering to second-order statistics of face image. The proposed methods has been applied to the problem for recognizing the 60 face images(15 persons *4 scenes) of 320*243 pixels. The 3 distances such as city-block, Euclidean, negative angle are used as measures when match the probe images to the nearest gallery images. In case of the 45 face images, the experimental results show that the recognition rate of the proposed methods is about 1.6 times and its the classification is about 5.6 times higher than conventional PCA without preprocessing. The city-block has been relatively achieved more an accurate classification than Euclidean or negative angle.