• 한국멀티미디어학회논문지
• /
• 제16권11호
• /
• pp.1338-1347
• /
• 2013

Face recognition is a science of automatically identifying individuals based their unique facial features. In order to avoid overfitting and reduce the computational reduce the computational burden, a new face recognition algorithm using PCA-fisher linear discriminant (PCA-FLD) and fuzzy radial basis function neural network (RBFNN) is proposed in this paper. First, face features are extracted by the principal component analysis (PCA) method. Then, the extracted features are further processed by the Fisher's linear discriminant technique to acquire lower-dimensional discriminant patterns, the processed features will be considered as the input of the fuzzy RBFNN. As a widely applied algorithm in fuzzy RBF neural network, BP learning algorithm has the low rate of convergence, therefore, an improved learning algorithm based on Levenberg-Marquart (L-M) for fuzzy RBF neural network is introduced in this paper, which combined the Gradient Descent algorithm with the Gauss-Newton algorithm. Experimental results on the ORL face database demonstrate that the proposed algorithm has satisfactory performance and high recognition rate.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2008년도 하계종합학술대회
• /
• pp.717-718
• /
• 2008

Fisher Linear Discriminant(FLD) is a sample and intuitive linear feature extraction method in pattern recognition. But in some special cases, such as un-separable case, one class data dispersed into several clustering case, FLD doesn't work well. In this paper, a new discriminant named K-means Fisher Linear Discriminant, which combines FLD with K-means clustering is proposed. It could deal with this case efficiently, not only possess FLD's global-view merit, but also K-means' local-view property. Finally, the simulation results also demonstrate its advantage against K-means and FLD individually.

• Communications for Statistical Applications and Methods
• /
• 제25권4호
• /
• pp.411-430
• /
• 2018

The Bayesian approach is a suitable alternative in constructing appropriate models for observed record values because the number of these values is small. This paper provides an objective Bayesian analysis method for upper record values arising from the Rayleigh distribution. For the objective Bayesian analysis, the Fisher information matrix for unknown parameters is derived in terms of the second derivative of the log-likelihood function by using Leibniz's rule; subsequently, objective priors are provided, resulting in proper posterior distributions. We examine if these priors are the PMPs. In a simulation study, inference results under the provided priors are compared through Monte Carlo simulations. Through real data analysis, we reveal a limitation of the appropriate confidence interval based on the maximum likelihood estimator for the scale parameter and evaluate the models under the provided priors.

• International Journal of Reliability and Applications
• /
• 제18권2호
• /
• pp.21-44
• /
• 2017

Accelerated life tests (ALTs) are frequently used in manufacturing industries to evaluate the reliability of products within a reasonable amount of time and cost. Test units are subjected to elevated stresses which yield quick failures. Most of the previous works on designing ALT plans are focused on tests that involve a single stress. Many times more than one stress factor influence the product's functioning. This paper deals with the design of optimum modified ramp-stress ALT plan for Burr type XII distribution with Type-I censoring under two stress factors, viz., voltage and switching rate each at two levels- low and high. It is assumed that usage time to failure is power law function of switching rate, and voltage increases linearly with time according to modified ramp-stress scheme. The cumulative exposure model is used to incorporate the effect of changing stresses. The optimum plan is devised using D-optimality criterion wherein the ${\log}_{10}$ of the determinant of Fisher information matrix is maximized. The method developed has been explained using a numerical example and sensitivity carried out.

• 응용통계연구
• /
• 제25권2호
• /
• pp.269-277
• /
• 2012

마이크로어레이 분석은 특이 발현하는 개별적인 유전자보다 유전자 온톨로지(Gene Ontology)와 같이 기능적 분류나 생물학적 경로(pathway)와 관련된 유전자군을 찾아내는 것이 그 해석의 용이성 때문에 최근 더욱 많은 연구가 진행되고 있다. 약물 처리에 의한 생물학적 반응을 연구할 때, 한 유전자군에 속하는 유전자들 각각의 특이 발현 여부의 유의성을 나타내는 $p$-value들을 취합하여 그 유전자군의 유의성을 결정하는 통계 검증 방법을 본 논문에서 소개하였다. 본 논문에 제시된 유전자군 분석(Gene group analysis) 방법은 Fisher's exact test나 permutation test와 같은 기존의 대표적인 방법들보다 더 정확하고 적용범위가 넓음을 실재 생물학 실험 자료의 분석을 통해 보였다. 제시된 유전자군 분석 방법은 SAS 프로그램으로 구현되었고 저자의 홈페이지(http://cafe.daum.net/go.analysis)에서 내려 받아 사용할 수 있다.

• 비파괴검사학회지
• /
• 제21권1호
• /
• pp.46-53
• /
• 2001

음향방출 신호를 이용하여 분류기를 설계하는 과정에서의 특징값 선택법에 관해 연구하였다. 분류기는 역전파법을 이용한 신경망 분류기를 사용하였다. Fisher's criterion, class mean scatter criterion, eigenvector analysis와 함께 본 논문에서 새로 제안하는 특징값 공간에서의 특징값 좌표사이의 차이를 이용하는 2-D criterion, 3-D criterion을 이용해서 특징값을 선택하고 각각에 대해 분류기를 설계하여, 인식률과 수렴속도를 비교하였다. 분류를 위한 자료를 얻기 위하여 용접부 결함시편과 로터리 압축기 금속 접촉부 결함시편을 사용하였다. 인식률 면에서 2-D criterion과 3-D criterion이 우수한 결과를 나타내었다.

• 정보과학회 논문지
• /
• 제42권2호
• /
• pp.203-212
• /
• 2015

본 논문은 도로주행 영상에서 도로표지판을 인식하는 방법을 제안한다. 지능형 차량에서 얻어지는 도로표지판 영상은 일반적인 사물 영상과는 다른 두 가지 특징이 있다. 첫째는 대상이 되는 사물들은 종류가 제한적이고 형태가 단순한 도형인 경우가 대부분이다. 둘째는 일반적인 도로주행 영상은 다양한 조명 환경과 날씨 상태로 인해서 선명한 영상을 취득하기 어려운 점이다. 본 논문에서는 조명 변화가 심한 도로주행 영상에 대해서 효과적으로 특징을 추출하기 위해서 Modified Census Transform(MCT)을 개선한 특징추출 방법을 제안한다. 추출된 특징들은 히스토그램으로 쌓여지고 영상 전반에 걸쳐 아주 고차원의 기술자(Descriptor)로 변환되며, 변환된 수많은 기술자들은 가우시안 혼합 모델(Gaussian Mixture Model)을 활용한 Fisher-vector 방법에 의해서 저차원으로 변형하여 특징으로 사용한다. 본 논문에서 제안하는 방법은 일반적인 표지판 인식 방법에 비해서 조명변화에 강한 검출 결과를 보여주었으며, 실시간 검출 및 인식도 가능하였다.

• Geomechanics and Engineering
• /
• 제13권5호
• /
• pp.861-879
• /
• 2017

Coal and gas outburst is a serious dynamic disaster that occurs during coal mining and threatens the lives of coal miners. Currently, coal and gas outburst is commonly predicted using single indicator and its critical value. However, single indicator is unable to fully reflect all of the factors impacting outburst risk and has poor prediction accuracy. Therefore, a more accurate prediction method is necessary. In this work, we first analyzed on-site impacting factors and precursors of coal and gas outburst; then, we constructed a Fisher discriminant analysis (FDA) index system using the gas adsorption index of drilling cutting ${\Delta}h_2$, the drilling cutting weight S, the initial velocity of gas emission from borehole q, the thickness of soft coal h, and the maximum ratio of post-blasting gas emission peak to pre-blasting gas emission $B_{max}$; finally, we studied an FDA-based multiple indicators discriminant model of coal and gas outburst, and applied the discriminant model to predict coal and gas outburst. The results showed that the discriminant model has 100% prediction accuracy, even when some conventional indexes are lower than the warning criteria. The FDA method has a broad application prospects in coal and gas outburst prediction.

• 한국지능시스템학회논문지
• /
• 제17권4호
• /
• pp.460-465
• /
• 2007

본 논문에서는 규칙의 커플링 문제를 최소화하기 위해 주어진 데이터의 통계적 특성과 퍼지-러프집합을 기반으로 한 새로운 패턴분류 방법을 제안한다. 제안한 방법 하에서 주어진 데이터의 통계적 특성은 입력부 퍼지집합의 파티션 개수를 결정하고, 생성된 규칙의 커플링문제를 최소화하기 위한 선택기준으로 사용하였다. 또한 러프집합은 수치적인 데이터로부터 생성된 규칙들 간의 불필요한 속성들을 제거하기 위한 도구로서 이용하였다. 제안된 방법의 타당성을 검증하기 위하여 Fisher의 IRIS 데이터를 사용하여 기존의 패턴분류 방법과 분류 정확도를 비교하였다. 실험결과, 제안한 방법이 기존의 학습에 의한 방법들보다 비교적 좋은 성능을 가진다는 것을 알 수 있었다.

• Journal of Korean Neurosurgical Society
• /
• 제30권12호
• /
• pp.1375-1380
• /
• 2001

Objective : The authors analyzed the incidence, the cause and the prognosis of hydrocephalus following aneurysmal subarachnoid hemorrhage to evaluate the risk factors of hydrocephalus and to provide the proper treatment method for hydrocephalus following aneurysmal subarachnoid hemorrhage. Methods : The 505 cases of subarachnoid hemorrhage followed by aneurysmal surgery from January 1990 to May 1999, were divided into shunt group and shunt-free group and we were reviewed for the clinical status, Fisher's grade, brain CT findings and prognosis. Results : The incidence of acute hydrocephalus was 37.2% of patients and 18.9% to developed chronic hydrocephalus. Shunt surgery due to chronic hydrocephalus was required in 6.5% of patients. We found following variables were significantly related to shunt-dependent hydrocephalus : high Hunt-Hess and Fisher grade, initial CT findings of intraventricular hemorrahge, posterior circulation aneurysm, preoperative rebleeding, delayed ischemic deficits, and initial high ventricular size index. There were no statistically significant relationships between shunt-dependent hydrocephalus and patient age or sex, timing of operation. The previous hypertension was not related to shunt dependent hydrocephalus. Prognosis in shunt group showed poor result. Conclusion : The risk factors of hydrocephalus following aneurysmal subarachnoid hemorrhage are high Hunt-Hess grade, high Fisher's grade, aneurysms of posterior circulations, preoperative aneurysmal rebleeding, delayed ischemic deficits, initial CT findings of intraventricular hemorrahge and initially increased ventricular size. The patients with these factors should the carefully observed and managed accordingly due to poor prognosis related to hydrocephalus requiring shunt operation.