• Communications for Statistical Applications and Methods
• /
• 제10권1호
• /
• pp.1-9
• /
• 2003

In this article, we consider a Bayesian estimation method for the geometric mean of $textsc{k}$ exponential parameters, Using the Tibshirani's orthogonal parameterization, we suggest an invariant prior distribution of the $textsc{k}$ parameters. It is seen that the prior, probability matching prior, is better than the uniform prior in the sense of correct frequentist coverage probability of the posterior quantile. Then a weighted Monte Carlo method is developed to approximate the posterior distribution of the mean. The method is easily implemented and provides posterior mean and HPD(Highest Posterior Density) interval for the geometric mean. A simulation study is given to illustrates the efficiency of the method.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2001년도 ICCAS
• /
• pp.24.3-24
• /
• 2001

In cases where strong a priori knowledge about the object being analyzed is available, it can be embedded into the formulation of the snake model. When prior knowledge of shape is available for a specific application, information concerning the shape of the desired objects can be incorporated into the formulation of the snake model as an active contour model. In this paper we show Five points algorithm can be applied to design invariant energy.

• Communications for Statistical Applications and Methods
• /
• 제30권3호
• /
• pp.311-330
• /
• 2023

The present work describes simulation studies to compare the performances in terms of averaged mean squared error of bayesian wavelet shrinkage methods in estimating component curves from aggregated functional data. Five bayesian methods available in the literature were considered to be compared in the studies: The shrinkage rule under logistic prior, shrinkage rule under beta prior, large posterior mode (LPM) method, amplitude-scale invariant Bayes estimator (ABE) and Bayesian adaptive multiresolution smoother (BAMS). The so called Donoho-Johnstone test functions, logit and SpaHet functions were considered as component functions and the scenarios were defined according to different values of sample size and signal to noise ratio in the datasets. It was observed that the signal to noise ratio of the data had impact on the performances of the methods. An application of the methodology and the results to the tecator dataset is also done.

• 제어로봇시스템학회논문지
• /
• 제5권2호
• /
• pp.228-236
• /
• 1999

In this paper, we propose visual-based self-localization and obstacle detection algorithms for indoor mobile robots. The algorithms do not require calibration, and can be worked with only single image by using the projective invariant relationship between natural landmarks. We predefine a risk zone without obstacles for a robot, and update the image of the risk zone, which will be used to detect obstacles inside the zone by comparing the averaging image with the current image of a new risk zone. The positions of the robot and the obstacles are determined by relative positioning. The method does not require the prior information for positioning robot. The robustness and feasibility of our algorithms have been demonstrated through experiments in hallway environments.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권3호
• /
• pp.993-1014
• /
• 2021

This paper proposes a new image retrieval method called the 3D integrated multi-index to fuse SIFT (Scale Invariant Feature Transform) visual words with other features at the indexing level. The advantage of the 3D integrated multi-index is that it can produce finer subdivisions in the search space. Compared with the inverted indices of medium-sized codebook, the proposed method increases time slightly in preprocessing and querying. Particularly, the SIFT, contour and colour features are fused into the integrated multi-index, and the joint cooperation of complementary features significantly reduces the impact of false positive matches, so that effective image retrieval can be achieved. Extensive experiments on five benchmark datasets show that the 3D integrated multi-index significantly improves the retrieval accuracy. While compared with other methods, it requires an acceptable memory usage and query time. Importantly, we show that the 3D integrated multi-index is well complementary to many prior techniques, which make our method compared favorably with the state-of-the-arts.

• 전자공학회논문지SC
• /
• 제40권3호
• /
• pp.117-126
• /
• 2003

본 연구에서는 월쉬함수를 이용하여 샘플링 구간내에서 데이터를 처리할 수 있는 온라인 월쉬변환과 샘플링 구간내에서 미분 방정식을 계산할 수 있는 새로운 온라인 월쉬함수 미분연산법을 제안하였다. 스케일링 인자의 도입과 다음구간에서의 초기조건을 계산하여 줌으로써 임의의 샘플링 시간을 취할 수 있게 하였으며 제안된 월쉬함수 온라인 알고리즘을 이용하여 기존의 직교함수가 가지는 최종시간까지의 신호를 모두 알고 있어야 그 적용이 가능하다는 단점을 제거하였다. 유사변환법을 이용하여 유도된 동적 시스템에 대한 Luenberger관측기를 월쉬함수를 이용하여 새롭게 변환함으로써 관측기에 포함된 출력의 미분항을 없애기 위한 불필요한 관측기 방정식의 분할을 피하였으며 대수적으로 상태 및 미지입력을 추정하였다. 제안된 방법을 이용하면 실시간 처리를 요하는 시스템에 유용하게 적용될 것으로 기대된다.

• 호남수학학술지
• /
• 제5권1호
• /
• pp.45-69
• /
• 1983

The hermitian structures on complex manifolds have been studied by several mathematicians ([1], [2], and [3]), and the Kähler structure on hermitian manifolds have been so much too ([6], [12], and [15]). There has been some gradual progress in studying the invariant forms on Grassmann manifolds ([17]). The purpose of this dissertation is to prove the Theorem 3.4 and the Theorem 4.7, with relation to the nature of complex Grassmann manifolds. In $\S$ 2. in order to prove the Theorem 4.7, which will be explicated further in $\S$ 4, the concepts of the hermitian structure, connection and curvature have been defined. and the characteristic nature about these were proved. (Proposition 2.3, 2.4, 2.9, 2.11, and 2.12) Two characteristics were proved in $\S$ 3. They are almost not proved before: particularly. we proved the Theorem 3.3 : $G_{k}(C^{n+k})=\frac{GL(n+k,C)}{GL(k,n,C)}=\frac{U(n+k)}{U(k){\times}U(n)}$ In $\S$ 4. we explained and proved the Theorem 4. 7 : i) Complex Grassmann manifolds are Kahlerian. ii) This Kähler form is $\pi$-fold of curvature form in hyperplane section bundle. Prior to this proof. some propositions and lemmas were proved at the same time. (Proposition 4.2, Lemma 4.3, Corollary 4.4 and Lemma 4.5).

• Smart Structures and Systems
• /
• 제3권4호
• /
• pp.423-437
• /
• 2007

The empirical mode decomposition (EMD) method is well-known for its ability to decompose a multi-component signal into a set of intrinsic mode functions (IMFs). The method uses a sifting process in which local extrema of a signal are identified and followed by a spline fitting approximation for decomposition. This method provides an effective and robust approach for decomposing nonlinear and non-stationary signals. On the other hand, the IMF components do not automatically guarantee a well-defined physical meaning hence it is necessary to validate the IMF components carefully prior to any further processing and interpretation. In this paper, an attempt to use the EMD method to identify properties of nonlinear elastic multi-degree-of-freedom structures is explored. It is first shown that the IMF components of the displacement and velocity responses of a nonlinear elastic structure are numerically close to the nonlinear normal mode (NNM) responses obtained from two-dimensional invariant manifolds. The IMF components can then be used in the context of the NNM method to estimate the properties of the nonlinear elastic structure. A two-degree-of-freedom shear-beam building model is used as an example to illustrate the proposed technique. Numerical results show that combining the EMD and the NNM method provides a possible means for obtaining nonlinear properties in a structure.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권10호
• /
• pp.4214-4230
• /
• 2020

This paper introduces a pilot allocation scheme for massive MIMO systems based on deep convolutional neural network (CNN) learning. This work is an extension of a prior work on the basic deep learning framework of the pilot assignment problem, the application of which to a high-user density nature is difficult owing to the factorial increase in both input features and output layers. To solve this problem, by adopting the advantages of CNN in learning image data, we design input features that represent users' locations in all the cells as image data with a two-dimensional fixed-size matrix. Furthermore, using a sorting mechanism for applying proper rule, we construct output layers with a linear space complexity according to the number of users. We also develop a theoretical framework for the network capacity model of the massive MIMO systems and apply it to the training process. Finally, we implement the proposed deep CNN-based pilot assignment scheme using a commercial vanilla CNN, which takes into account shift invariant characteristics. Through extensive simulation, we demonstrate that the proposed work realizes about a 98% theoretical upper-bound performance and an elapsed time of 0.842 ms with low complexity in the case of a high-user-density condition.

• 한국컴퓨터정보학회논문지
• /
• 제14권4호
• /
• pp.47-57
• /
• 2009

최근 컴퓨터비젼 분야에서 이벤트 검출 및 인식이 활발히 연구되고 있으며, 도전적인 주제들 중 하나이다. 본 논문에서는 사무실 환경에서 발생할 수 있는 이벤트의 검출 및 인식을 위한 방법을 제안한다. 제안된 방법은 MHI(Motion History Image) 시퀀스(sequence)를 이응한 인간의 모션을 분석하며, 사람의 처형과 착용한 옷의 종류와 색상, 그리고 카메라로부터의 위치관계에 불변한 특성을 가진다. 제안된 방법은 기존의 방법들 중, 칼라 정보를 이용한 방법에 비해 조명의 변화에 민감하지 않은 장점이 있으며, 관심의 대상이 되는 객체의 외형과 같은 사전지식에 의존하는 방법에 비해 스케일에 민감하지 않은 장점이 있다. 에지검출 기술을 HMI 순서 영상 정보와 결합하여 사람 모션의 기하학적 특징을 추출한 후, 이벤트 인식의 기본정보로 활용한다. 제안된 방법은 단순한 이벤트 검출 프레임웍을 사용하기 때문에 검출하고자 하는 이벤트의 설명만을 첨가하는 것으로 확장이 가능하다. 또한, 제안된 방법은 컴퓨터비젼 기술에 기반한 많은 감시시스템 뿐 아니라 상황인식 기반의 이벤트 검출 시스템에 핵심기술이다.