• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.42 no.6
• /
• pp.167-176
• /
• 2005

This paper describes how the state tying model based on the decision tree which is one of Acoustic models used for speech recognition optimizes the model by reducing the number of mixture Gaussians of the output probability distribution. The state tying modeling uses a finite set of questions which is possible to include the phonological knowledge and the likelihood based decision criteria. And the recognition rate can be improved by increasing the number of mixture Gaussians of the output probability distribution. In this paper, we'll reduce the number of mixture Gaussians at the highest point of recognition rate by clustering the Gaussians. Bhattacharyya and Euclidean method will be used for the distance measure needed when clustering. And after calculating the mean and variance between the pair of lowest distance, the new Gaussians are created. The parameters for the new Gaussians are derived from the parameters of the Gaussians from which it is born. Experiments have been performed using the STOCKNAME (1,680) databases. And the test results show that the proposed method using Bhattacharyya distance measure maintains their recognition rate at $97.2\%$ and reduces the ratio of the number of mixture Gaussians by $1.0\%$. And the method using Euclidean distance measure shows that it maintains the recognition rate at $96.9\%$ and reduces the ratio of the number of mixture Gaussians by $1.0\%$. Then the methods can optimize the state tying model.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2014.11a
• /
• pp.130-133
• /
• 2014

본 논문은 Laplacian of Gaussians (LoG)에 캐니 에지 검출 기법을 접목한 새로운 색상 보간 알고리듬을 제안한다. 캐니 에지 검출 기법은 영상 스무딩, 기울기 크기와 각도 계산, 세션화, 이중 문턱치 처리 과정으로 이루어진다. 이때 앞의 두 과정을 LoG를 이용하여 처리함으로써 기존의 캐니 애지 검출법보다 정확한 방향 정보를 얻을 수 있다. 실험결과를 통해 기존의 색상 보간 알고리듬에 비해 Peak Signal to Noise Ratio (CPSNR)이 상승함을 확인하였으며, 에지 영역 주변에서 발생하였던 무지개 에러가 현저히 감소하였음을 확인할 수 있었다.

• Steel and Composite Structures
• /
• v.47 no.5
• /
• pp.569-585
• /
• 2023

This paper proposes an efficient approach for the structural topology optimization of bi-directional functionally graded structures by incorporating popular radial basis functions (RBFs) into an implicit level set (ILS) method. Compared to traditional element density-based methods, a level set (LS) description of material boundaries produces a smoother boundary description of the design. The paper develops RBF implicit modeling with multiquadric (MQ) splines, thin-plate spline (TPS), exponential spline (ES), and Gaussians (GS) to define the ILS function with high accuracy and smoothness. The optimization problem is formulated by considering RBF-based nodal densities as design variables and minimizing the compliance objective function. A LS-RBF optimization method is proposed to transform a Hamilton-Jacobi partial differential equation (PDE) into a system of coupled non-linear ordinary differential equations (ODEs) over the entire design domain using a collocation formulation of the method of lines design variables. The paper presents detailed mathematical expressions for BiDFG beams topology optimization with two different material models: continuum functionally graded (CFG) and mechanical functionally graded (MFG). Several numerical examples are presented to verify the method's efficiency, reliability, and success in accuracy, convergence speed, and insensitivity to initial designs in the topology optimization of two-dimensional (2D) structures. Overall, the paper presents a novel and efficient approach to topology optimization that can handle bi-directional functionally graded structures with complex geometries.

• Bulletin of the Korean Chemical Society
• /
• v.25 no.7
• /
• pp.997-1002
• /
• 2004

A hydrogenperoxide sensor containing peroxidase extracted from horseradish was constructed and pH effect on its sensing ability was investigated. Current profiles of the biosensor with pH and the electrophoretic analysis showed that horseradish peroxidase consists of two isozymes. Assuming that it is a hypothetical twoisozyme mixture, the current profiles were deconvoluted into two Gaussians. Application of the new Michaelis-Menten equation connoting pH concept to this system enabled to find all the related dissociation constants of the isozyme-substrates and the isozyme-proton complexes and to determine pHs for the maximal isozyme activities.

• Analytical Science and Technology
• /
• v.16 no.6
• /
• pp.504-508
• /
• 2003

A carbon paste electrode was constructed with peroxidase extracted from Horseradish and the variation of the response of the sensor with pH was investigated. Current profiles showed two highest sensitivities at two pH values respectively. In addition, two bands were observed in the electrophoretic expansion. A coincidence of the two experimental results added support to the possibility that the biosensor has two different isozymes. Assuming that current profiles are the sum of two gaussians, we deconvoluted them and determined the optimum pH of peroxidase isozymes.

• Journal of applied mathematics & informatics
• /
• v.23 no.1_2
• /
• pp.435-443
• /
• 2007

An important approach towards solving the scattered data problem is by using radial basis functions. However, for a large class of smooth basis functions such as Gaussians, the existing theories guarantee the interpolant to approximate well only for a very small class of very smooth approximate which is the so-called 'native' space. The approximands f need to be extremely smooth. Hence, the purpose of this paper is to study approximation by a scattered shifts of a radial basis functions. We provide error estimates on larger spaces, especially on the homogeneous Sobolev spaces.

• The KIPS Transactions:PartB
• /
• v.12B no.4 s.100
• /
• pp.395-402
• /
• 2005

Object extraction is needed to track objects in color traffic image sequence. To extract objects, we use background differencing method based on MOG(Mixture of Gaussians). In extracted objects, shadows may be included. Due to shadows, we may not find exact location of objects and sometimes we find adjacent objects are glued together. Many methods have been proposed to remove shadows. Conventional methods usually assume that color and texture information are preserved under the shadow. Thus these methods do not work well if these assumptions do not hold. In this paper, we propose a new robust shadow removal method which works well in those situations. First we extract shadow pixel candidates by analysing color information and compute the ratio of shadow pixel candidates over the total number of Pixels. W the ratio is reasonable, we remove shadow candidate Pixels and if not, we use data in history array containing Previous removal records. We applied the method to real color traffic image sequences and obtained good results.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.30B no.1
• /
• pp.90-98
• /
• 1993

A Gaussian is formed by diffusing a spot excitation. In this paper, a diffusion neural network model is derived from the diffusion equation. And it is shown that a difference of two Gaussians(DOG) may have the same shape as a Laplacian of Gaussian(LOG), A neural network model executing a DOG convolution by diffusing an external excitation is proposed. By this model intensity changes of image may be detected. This model may be implemented economically because each neuron has only four fixed-valued synapes.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.8
• /
• pp.3911-3924
• /
• 2016

Compared to the classical cryptography, lattice-based cryptography is more secure, flexible and simple, and it is believed to be secure against quantum computers. In this paper, an efficient signature scheme is proposed from the ring learning with errors (R-LWE), which avoids sampling from discrete Gaussians and has the characteristics of the much simpler description etc. Then, the scheme is implemented in C/C++ and makes a comparison with the RSA signature scheme in detail. Additionally, a linearly homomorphic signature scheme without trapdoor is proposed from the R-LWE assumption. The security of the above two schemes are reducible to the worst-case hardness of shortest vectors on ideal lattices. The security analyses indicate the proposed schemes are unforgeable under chosen message attack model, and the efficiency analyses also show that the above schemes are much more efficient than other correlative signature schemes.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.32B no.1
• /
• pp.120-126
• /
• 1995

The diffusion neural network can be cfficiently applied to the Gaussian processing. For example, a difference of two Gaussians(DOG) is performed by this network with ease. In this paper, we model a neural network to perform the function /t(.del.${\Delta}^{2}$G) by using the diffusion neural network. This model is used to detect the edges of moving target in image. By this model not only moving target is separated from stationary background but also their trajectories are obtained using accumulated past information in the diffusion neural network. Furthermore this model needs a small number of connections per cell and the connection weights are fixed-valued. Therefore its hardware can be easily implemented with simple structure.