• 대한의용생체공학회:의공학회지
• /
• 제31권6호
• /
• pp.456-463
• /
• 2010

In the present study, we proposed a new subspace scanning algorithm to enhance the spatial resolution of electroencephalography (EEG) and magnetoencephalography(MEG) source localization. Subspace scanning algorithms, represented by the multiple signal classification (MUSIC) algorithm and the first principal vector (FINE) algorithm, have been widely used to localize asynchronous multiple dipolar sources in human cerebral cortex. The conventional MUSIC algorithm used principal component analysis (PCA) to extract the noise vector subspace, thereby having difficulty in discriminating two or more closely-spaced cortical sources. The FINE algorithm addressed the problem by using only a part of the noise vector subspace, but there was no golden rule to determine the number of noise vectors. In the present work, we estimated a non-orthogonal signal vector set using independent component analysis (ICA) instead of using PCA and performed the source scanning process in the signal vector subspace, not in the noise vector subspace. Realistic 2D and 3D computer simulations, which compared the spatial resolutions of various algorithms under different noise levels, showed that the proposed ICA-MUSIC algorithm has the highest spatial resolution, suggesting that it can be a useful tool for practical EEG/MEG source localization.

• 전자공학회논문지B
• /
• 제33B권10호
• /
• pp.107-119
• /
• 1996

In this paper, we propose multiple component neural network(MCNN) which learn partitioned patterns in each multiple component neural networks by reducing dimensions of input pattern vector using PCA (principal component analysis). Procesed neural network use Oja's rule that has a role of PCA, output patterns are used a slearning patterns on small component neural networks and we call it CBP. For simply not solved patterns in a network, we solves it by regenerating new CBP neural networks and by performing dynamic partitioned pattern learning. Simulation results shows that proposed MCNN neural networks are very small size networks and have very fast learning speed compared with multilayer neural network EBP.

• International Journal of Automotive Technology
• /
• 제4권2호
• /
• pp.101-108
• /
• 2003

There are a lot of factors that influence automotive fuel economy such as average trip time per kilometer, average trip speed, the number of times of vehicle stationary, and so forth. These factors depend on road conditions and traffic environment. In this study, various driving data were measured and recorded during road tests in Seoul. The accumulated road test mileage is around 1,300 kilometers. The objective of the study is to identify the driving patterns of the Seoul metropolitan area and to analyze the fuel economy based on these driving patterns. The driving data which was acquired through road tests was analysed statistically in order to obtain the driving characteristics via modal analysis, speed analysis, and speed-acceleration analysis. Moreover, the driving data was analyzed by multivariate statistical techniques including correlation analysis, principal component analysis, and multiple linear regression analysis in order to obtain the relationships between influencing factors on fuel economy. The analyzed results show that the average speed is around 29.2 km/h, and the average fuel economy is 10.23 km/L. The vehicle speed of the Seoul metropolitan area is slower, and the stop-and-go operation is more frequent than FTP-75 test mode which is used for emission and fuel economy tests. The average trip time per kilometer is one of the most important factors in fuel consumption, and the increase of the average speed is desirable for reducing emissions and fuel consumption.

• 환경영향평가
• /
• 제23권3호
• /
• pp.187-196
• /
• 2014

The objective of this study is to evaluate and analyze Osu stream watershed water environment system. The data were collected from January 2009 to December 2011 including water temperature, pH, DO, EC, BOD, COD, TOC, SS, T-N, T-P and discharge. The data were used for principle component analysis and factor analysis. The results are as followes. The primary factors obtained from both the principal component analysis and the factor analysis were BOD, COD, TOC, SS and T-P. Once principal component analysis and factor analysis have been performed with the collected data and then the results will be applied to both simple regression model and multiple regression model. The regression model was developed into case 1 using concentrations of water quality parameters and case 2 using delivery loads. The value of the coefficient of determination on case 1 fell between 0.629 and 0.866; this was lower than case 2 value which fell between 0.946 and 0.998. Therefore, case 2 model would be a reliable choice.The coefficient of determination between the estimated figure using data which was developed to the regression model in 2012 and the actual measurement value was over 0.6, overall. It can be safely deduced that the correlation value between the two findings was high. The same model can be applied to get TOC concentrations in future.

• 한국멀티미디어학회논문지
• /
• 제4권1호
• /
• pp.37-45
• /
• 2001

다대역(multi-band) 카메라 시스템으로 물체의 분광 반사율을 추정하여 피사체의 고유한 색자극을 기록하기 위한 연구가 활발하다. 그러나 다대역 카메라 시스템은 대역 수에 따라 추가적인 색필터가 필요하며 중복촬영으로 인하여 시스템의 복잡성이 증가하는 단점이 있다. 따라서 본 논문에서는 기존의 3대역 RGB 카메라를 이용하여 분광 반사율의 추정 오차를 개선하는 방법을 제안한다. 제안된 방법에서는 화소 단위로 반사광의 모집단을 갱신하여 각 입력색에 대해 적응적인 주성분 벡터를 구하였으며 이를 이용하여 해당 화소의 분광반사율 추정시 오차를 줄였다. 제안된 반사율 추정 방법의 유용성을 평가하기 위하여 제안된 방법과 3대역 주성분 분석(principal component analysis) 방법 및 5대역 위너 추정(Wiener estimation) 방법에 대하여 각각 Macbeth ColorChecker에 대한 분광 반사율 추정 실험을 하였다. 결과, 제안한 방법은 색차 및 분광 반사율 평균자승오차가 기존의 3대역 주성분 분석 방법보다 적었으며 5대역 카메라를 이용한 분광 반사율 추정 방법과 근사하건, 개선되었음을 확인하였다.

• 한국대기환경학회지
• /
• 제8권3호
• /
• pp.179-190
• /
• 1992

This study was carried out to characterize sources of atmospheric total suspended particulates (TSP) in urban and sub--urban areas of metropolitan taegu. The sources were tentatively identified by a multivariate technique, i.e. principal component analysis (PCA), and the source contributions to the atmospheric concentrations of TSP were further estimated by stepwise multiple regression analysis. A total of 5 sources was identified in the urban area of Taegu (soil dust resuspension, fuel combustion, secondary aerosol, traffic related aerosol, and refuge burning), while 4 sources were found to be significant in the sub--urban area as following: fuel combustion/secondary aerosol, soil dust resuspension, traffic related aerosol, and wood/agricultural burning. The largest contributor to the atmospheric TSP appeared to be the soil dust resuspension in both areas. The source apportionment of the extractable organic matter (EOM) was also carried out for the Taegu data. The EOM was determined with respect to the solvent polarity, i.e. cyclohexane (non-polar), dichloromethane (semi--polar), and acetone (polar). In addition, the source profiles for the TSP in Taegu area were estimated using a PCA-based algorithm, and the validity was evaluated tentatively by comparing the data in the literature.

• 한국멀티미디어학회논문지
• /
• 제18권6호
• /
• pp.701-709
• /
• 2015

Image stitching method is a technique for obtaining an high-resolution image by combining two or more images. In X-ray image for clinical diagnosis, the size of the imaging region taken by one shot is limited due to the field-of-view of the equipment. Therefore, in order to obtain a high-resolution image including large regions such as a whole body, the synthesis of multiple X-ray images is required. In this paper, we propose a rapid stitching method of digital X-ray images using template-based registration. The proposed algorithm use principal component analysis(PCA) and k-nearest neighborhood(k-NN) to determine the location of input images before performing a template-based matching. After detecting the overlapping position using template-based matching, we synthesize input images by alpha blending. To improve the computational efficiency, reduced images are used for PCA and k-NN analysis. Experimental results showed that our method was more accurate comparing with the previous method with the improvement of the registration speed. Our stitching method could be usefully applied into the stitching of 2D or 3D multiple images.

• Applied Biological Chemistry
• /
• 제45권1호
• /
• pp.18-24
• /
• 2002

본 연구는 보리등겨로 제조한 간장 맛의 특성을 찾기 위해서 수행되었다. 맛성분은 기기분석으로, 관능검사는 panel로, 그 외 통계적 처리의 방법 등을 이용하였다. 보리간장 맛성분은 유기산, 유리당 및 유리아미노산으로 분류하였으며, 이들과 관능검사 성적과의 단순상관으로 보리간장 맛의 품질을 결정하는 것은 불가능하였다. 중상관계수는 절대값의 대수 변환에서 가장 높게 나타났으며, 따라서 단계적 중회귀분석은 가장 설명력이 높으며, 표준오차가 적은 절대값의 대수 변환을 이용하여 실시하였다. 단계적 중회귀분석 결과, 보리간장 맛의 좋고 나쁨에 기여를 하는 성분은 짠맛, 구수한 맛 및 쓴맛을 내는 성분 순이었다.

• 대한화학회지
• /
• 제36권1호
• /
• pp.113-124
• /
• 1992

Chemometrics의 한 분야인 패턴인지(pattern recognition)법을 한국산 고대 유리시료 94종의 중성자방사화분석으로부터 얻은 다변수데이타에 적용하였다. unsupervised learning의 방법인 주성분분석과 비선형도시법으로 시료를 분류한 결과 유리시료는 6개의 군을 형성하였다. 6개의 참조시료셋트와 시험시료셋트에 supervised learning의 SIMCA법을 적용시켰다. 그 결과 참조시료셋트는 주성분분석법 및 비선형도시법의 결과와 일치하였고 시험시료셋트에서 33개의 시료 중 17개 시료에 대해 시료가 속한 군을 판정할 수 있었다.

• 한국CDE학회논문집
• /
• 제16권4호
• /
• pp.249-259
• /
• 2011

In this paper we present our work on the parameterized construction of virtual drivers' reach motion to seat belt, by using motion capture data. A user can generate a new reach motion by controlling a number of parameters. We approach the problem by using multiple sets of example reach motions and learning the relation between the labeling parameters and the motion data. The work is composed of three tasks. First, we construct a motion database using multiple sets of labeled motion clips obtained by using a motion capture device. This involves removing the redundancy of each motion clip by using PCA (Principal Component Analysis), and establishing temporal correspondence among different motion clips by automatic segmentation and piecewise time warping of each clip. Next, we compute motion blending functions by learning the relation between labeling parameters (age, hip base point (HBP), and height) and the motion parameters as represented by a set of PC coefficients. During runtime, on-line motion synthesis is accomplished by evaluating the motion blending function from the user-supplied control parameters.