• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 1998년도 추계종합학술대회 논문집
• /
• pp.891-894
• /
• 1998

This paper introduces an automatic face region extraction method. This method consists of two part: face recognition and extraction of facial organs which are eye, eyebrow, nose and mouth. In first stage, we use genetic algorithms(GAs) to get face region in complex background. In second stage, we use Geometrical Face Model to textract eye, eyebrow, nose and mouth. In both stage, stochastic component is used to deal with the problems caused by had lighting condition. According to this value, blurring number is determined. Average Computation time is less than 1 sec, and using this method we can extract facial feature efficiently from several images which has different lightning condition.

• 한국컴퓨터산업학회논문지
• /
• 제6권2호
• /
• pp.365-372
• /
• 2005

일반적으로 응용프로그램에서 병렬성 추출에 대한 핵심 부분은 루프이고 루프내의 첨자변수들 사이에는 자료종속성이 존재한다. 특히 문장들 사이에 가변 및 불변종속거리를 갖는 종속관계는 매우 복잡하다. 따라서 이 경우 컴파일 시 병렬성 추출은 매우 어렵다. 따라서 본 논문에서는 기존에 제시되어있는 병렬성 추출 방법들 중에서 unimodular방법과 non-unimodular방법에 대하여 분석하고 이들의 장단점을 파악하여 향후 효율적인 종속성 제거방법을 제안하는데 도움이 되고자한다.

• 수산해양기술연구
• /
• 제35권4호
• /
• pp.435-447
• /
• 1999

It is main objective of this approach to present a method to analyse stochastic design sensitivity for problems of structural dynamics with randomness in design parameters. A combination of the adjoint variable approach and the second order perturbation method is used in the finite element approach. An alternative form of the constant functional that holds for all times is introduced to consider the time response of dynamic sensitivity. The terminal problem of the adjoint system is solved using equivalent homogeneous equations excited by initial velocities. The numerical procedures are shown to be much more efficient when based on the fold superposition method: the generalized co-ordinates are normalized and the correlated random variables are transformed to uncorrelated variables, whereas the secularities are eliminated by the fast Fourier transform of complex valued sequences. Numerical algorithms have been worked out and proved to be accurate and efficient : they can be readily adapted to fit into the existing finite element codes whose element derivative matrices can be explicitly generated. The numerical results of two cases -2 dimensional portal frame for the comparison with reference and 3-dimensional frame structure - for the deterministic sensitivity analysis are presented.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2001년도 하계학술대회 논문집 B
• /
• pp.904-906
• /
• 2001

This paper presents a magnetic circuit design procedure by using Response Surface Methodology(RSM) to determine initial and detail design parameters for reducing torque ripple in BLDC motor of Electric Power Steering (EPS). RSM is achieved through using the experiment design method in combination with Finite Element Method and well adapted to make analytical model for a complex problem considering a lot of interaction of design variable Moreover, Sequential Quadratic Problem (SQP) method is used to solve the relsulting of constrained nonlinear optimization problem.

• Structural Engineering and Mechanics
• /
• 제5권4호
• /
• pp.369-383
• /
• 1997

The classical theory of thin-walled members is unable to reflect the shear lag phenomenon since it is based on the assumption of no shearing strains in the middle surface of the walls. In this paper, an energy equation for the lateral buckling of thin-walled members has been derived which includes the effects of torsion, warping and, especially, the shearing strains which reflect the shear lag phenomenon. A numerical analysis for the lateral buckling of thin-walled members with openings by using Galerkin's method of weighted residuals has been presented. The proposed numerical values and the predictions by experiment for the lateral buckling loads are to agree closely in the paper. The results from these comparisons show that the proposed method here is capable of predicting the lateral buckling of thin-walled members with openings. The fast convergence of the results indicates the numerical stability of the method. By the study, a very complex practical eigenvalue problem is transformed into a very simple one of solving only a linear equation with one variable.

• 한국컴퓨터정보학회논문지
• /
• 제3권2호
• /
• pp.131-138
• /
• 1998

통상 통계적인 데이터 해석에서 취급되는 데이터는 확정된 값으로서 통계 처리를실시한다. 그러나 복잡˙대규모화하는 현대의 시스템에 있어서는 정확하게 측정된 데이터만을 취급하는 것은 곤란하며 인간의 주관적인 판단에 따른 데이터를 수집하는 경우가 발생하게 된다. 본 연구에서는 이러한 인간의 주관적인 판단에 따른 데이터를 퍼지 관측 데이터로하여(언어 변수에 의해 Membership 함수를 정의한다.) 최대 엔트로피 원리를 이용한 새로운 분석 방법을 제안한다. 또한 보다 현실적인 상황 아래 시뮬레이션을 실시함으로서 제안모델의 유효성을 검증한다.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1999년도 하계학술대회 논문집 B
• /
• pp.507-509
• /
• 1999

In this paper, we propose a new fuzzy model identification method that can yield a successful fuzzy rule base for fundamental approximations. The method in this paper uses a set of input-output data and is based on a hybrid messy genetic algorithm (mGA) with a fine-tuning scheme. The mGA processes variable-length strings, while standard GAs work with a fixed-length coding scheme. For successfully identifying a complex nonlinear system, we first use the mGA, which coarsely optimizes the structure and the parameters of the fuzzy inference system, and then the gradient descent method which tine tunes the identified fuzzy model. In order to demonstrate the superiority and efficiency of the proposed scheme, we finally show its application to a nonlinear approximation.

• 소성∙가공
• /
• 제6권6호
• /
• pp.482-488
• /
• 1997

New carbon equivalent equation for the better prediction for the better prediction of roll force in a continuous hot strip mill has been formulated by applying a neural network method. In predicting roll force of steel strip, carbon equivalent equation which normalize the effects of various alloying elements by a carbon equivalent content is very critical for the accurate prediction of roll force. To overcome the complex relationships between alloying elements and operational variables such as temperature, strain, strain rate and so forth, a neural network method which is effective for multi-variable analysis was adopted in the present work as a tool to determine a proper carbon equivalent equation. The application of newly formulated carbon equivalent equation has increased prediction accuracy of roll force significantly and the effectiveness of neural network method is well confirmed in this study.

• Journal of Information Technology Applications and Management
• /
• 제13권1호
• /
• pp.77-94
• /
• 2006

Case-based Reasoning (CBR) is widely used in business and industry prediction. It is suitable to solve complex and unstructured business problems. Recently, the prediction accuracy of CBR has been enhanced by not only various machine learning algorithms such as genetic algorithms, relative weighting of Artificial Neural Network (ANN) input variable but also data mining technique such as feature selection, feature weighting, feature transformation, and instance selection As a result, CBR is even more widely used today in business area. In this study, we investigated the usefulness of the CBR method in forecasting success in implementing ERP systems. We used a CBR method based on the feature weighting technique to compare the performance of three different models : MDA (Multiple Discriminant Analysis), GECBR (GEneral CBR), FWCBR (CBR with Feature Weighting supported by Analytic Hierarchy Process). The study suggests that the FWCBR approach is a promising method for forecasting of successful ERP implementation in Small and Medium sized Enterprises.

• 한국통신학회논문지
• /
• 제22권3호
• /
• pp.469-480
• /
• 1997

For moving image coding, the variable size of region coding based on local motion is more efficient than fixed size of region coding. It can be applied well to complex motions and is more stable for wide motions because images are segmented according to local motions. In this paper, new image coding method using the segmentation of motion vectors is proposed. First, motion vector field is smoothed by filtering and segmented by smoothed motion vectors. The region growing method is used for decomposition of regions, and merging of regions is decided by motion vector and prediction errors of the region. Edge of regions is excluded because of the correlation of image, and neighbor motion vectors are used evaluation of current block and construction of region. The results of computer simulation show the proposed method is superior than the existing methods in aspect of coding efficiency.