• ETRI Journal
• /
• v.31 no.4
• /
• pp.438-447
• /
• 2009

In this paper, a novel image projection technique for face recognition application is proposed which is based on linear discriminant analysis (LDA) combined with the relevance-weighted (RW) method. The projection is performed through 2-directional and 2-dimensional LDA, or $(2D)^2$LDA, which simultaneously works in row and column directions to solve the small sample size problem. Moreover, a weighted discriminant hyperplane is used in the between-class scatter matrix, and an RW method is used in the within-class scatter matrix to weigh the information to resolve confusable data in these classes. This technique is called the relevance-weighted $(2D)^2$LDA, or RW$(2D)^2$LDA, which is used for a more accurate discriminant decision than that produced by the conventional LDA or 2DLDA. The proposed technique has been successfully tested on four face databases. Experimental results indicate that the proposed RW$(2D)^2$LDA algorithm is more computationally efficient than the conventional algorithms because it has fewer features and faster times. It can also improve performance and has a maximum recognition rate of over 97%.

• 제어로봇시스템학회:학술대회논문집
• /
• 1993.10a
• /
• pp.70-75
• /
• 1993

The performance of the BEWE(Bearing Estimation Without Eigendecomposition) algorithm depends on the sensor outputs which are selected to construct the projection matrix. In this paper, we construct the covariance matrix of the bearing estimates for two targets and propose the criterion to select the sensor outputs which minimize the covariance matrix. The computer simulation conforms that the estimation error is smallest when the sensor outputs are selected based on the proposed criterion.

• Korean Journal of Optics and Photonics
• /
• v.11 no.6
• /
• pp.441-446
• /
• 2000

Wigner functions and density matrices are computer simulated for various quantum mechanical states of light. Wigner function and density matrices are evaluated by filtered back projection which includes inverse Radon transform from the distribution function of the photocurrents, which are calculated in the balanced homodyne detection scheme. The density matrix is also directly obtained by using the pattern function from the simulated phase independent photocurrent distribution function. ction.

• Proceedings of the Korean Nuclear Society Conference
• /
• 1998.05b
• /
• pp.705-710
• /
• 1998

We describe a multigrid wavelet-based natural pixel (WNP) method for image reconstruction in emission computed tomography (ECT). The ECT is used to identify the tagged radioactive material's position in the body for detection of abnormal tissue such as tumor or cancer, as in SPECT and PET. With ECT methodology in parallel beam mode, we formulate a matrix-based reconstruction method for radionuclide sources in the human body. The resulting matrix for a practical problem is very large and nearly singular. To overcome this ill-conditioning, wavelet transform is considered in this study. Wavelets have inherent de-noising and multiscale resolution properties. Therefore, the multigrid wavelet-based natural pixel (WNP) method is very efficient to reconstruct image from projection data that is noisy and incomplete. We test this multigrid wavelet natural pixel (WNP) reconstruction method with the MCNP generated projection data for diagnosis of the simulated cancerous tumor.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.13 no.5
• /
• pp.606-613
• /
• 2003

As a technical method for controlling chaotic dynamics, this paper presents a predictive control for chaotic systems based on radial basis function networks(RBFNs). To control the chaotic systems, we employ an on-line identification unit and a nonlinear feedback controller, where the RBFN identifier is based on a suitable NARMA real-time modeling method and the controller is predictive control scheme. In our design method, the identifier and controller are most conveniently implemented using a gradient-descent procedure that represents a generalization of the least mean square(LMS) algorithm. Also, we introduce a projection matrix to determine the control input, which decreases the control performance function very rapidly. And the effectiveness and feasibility of the proposed control method is demonstrated with application to the continuous-time and discrete-time chaotic nonlinear system.

• Proceedings of the IEEK Conference
• /
• 2007.07a
• /
• pp.353-354
• /
• 2007

This paper proposes the method which estimates the pointing region at the real world. This paper uses the technique to easily calibrate a camera of Z. Zhang. First, we calculate the projection matrix of each camera by the technique. Next, we estimate the location of the shoulder and the fingertip. Then we compute the pointing region in 3D real world by using projection matrix of each camera. Experiment result showed that the error between estimated point and the plane center point is less than 5cm.

• Nuclear Engineering and Technology
• /
• v.50 no.5
• /
• pp.698-708
• /
• 2018

This article introduces a methodology of designing a wavelet operator suitable for multiscale modeling. The operator matrix transforms states of a multivariable system onto projection space. In addition, it imposes a specific structure on the system matrix in a multiscale environment. To be specific, the article deals with a diagonalizing transform that is useful for decoupled control of a system. It establishes that there exists a definite relationship between the model in the measurement space and that in the projection space. Methodology for deriving the multirate perfect reconstruction filter bank, associated with the wavelet operator, is presented. The efficacy of the proposed technique is demonstrated by modeling the point kinetics nuclear reactor. The outcome of the multiscale modeling approach is compared with that in the single-scale approach to bring out the advantage of the proposed method.

• Journal of the Korean Mathematical Society
• /
• v.56 no.4
• /
• pp.1001-1016
• /
• 2019

We are concerned with optimization methods for the $L^2$-Wasserstein least squares problem of Gaussian measures (alternatively the n-coupling problem). Based on its equivalent form on the convex cone of positive definite matrices of fixed size and the strict convexity of the variance function, we are able to present an implementable (accelerated) gradient method for finding the unique minimizer. Its global convergence rate analysis is provided according to the derived upper bound of Lipschitz constants of the gradient function.

• Journal of Electrical Engineering and Technology
• /
• v.13 no.2
• /
• pp.868-874
• /
• 2018

In this paper, we propose a new jammer suppression algorithm that uses orthogonal waveform space projection (OWSP) processing for a multiple input multiple output (MIMO) radar system exposed to a jamming signal. Generally, a conventional suppression algorithm based on adaptive beamforming (ABF) needs a covariance matrix composed of the jammer and noise only. By exploiting the orthogonality of the transmitting waveforms of MIMO, we can construct a transmitting waveform space (TWS). Then, using the OWSP processing, we can build a space orthogonal to the TWS that contains no SOI. By excluding the SOI from the received signal, even in the case that contains the SOI and jamming signal, the proposed algorithm makes it possible to evaluate the covariance matrix for ABF. We applied the proposed OWSP processing to suppressing the jamming signal in bistatic MIMO radar. We verified the performance of the proposed algorithm by comparing the SINR loss to that of the ideal covariance matrix composed of the jammer and noise only. We also derived the computational complexity of the proposed algorithm and compared the estimation of the DOD and DOA using the SOI with those using the generalized likelihood ratio test (GLRT) algorithm.

• Journal of Sensor Science and Technology
• /
• v.28 no.3
• /
• pp.182-186
• /
• 2019

When a camera is employed for 3D sensing, accurate camera calibration is vital as it is a prerequisite for the subsequent steps of the sensing process. Camera calibration is usually performed by complex mathematical modeling and geometric analysis. On the other contrary, data learning using an artificial neural network can establish a transformation relation between the 3D space and the 2D camera image without explicit camera modeling. However, a neural network requires a large amount of accurate data for its learning. A significantly large amount of time and work using a precise system setup is needed to collect extensive data accurately in practice. In this study, we propose a two-step neural calibration method that is effective when only a small amount of learning data is available. In the first step, the camera projection transformation matrix is determined using the limited available data. In the second step, the transformation matrix is used for generating a large amount of synthetic data, and the neural network is trained using the generated data. Results of simulation study have shown that the proposed method as valid and effective.