• Proceedings of the IEEK Conference
• /
• 2000.09a
• /
• pp.783-786
• /
• 2000

본 논문에서는 웨이브렛 변환을 적용한 광대역 음성부호화 알고리즘을 제안하였다. 제안한 음성부호화 알고리즘은 split-band 구조를 가지며, 16 kHz로 sampling된 입력신호를 QMF를 이용해서 동일한 대역폭을 갖는 두 개의 subband 신호로 나누고 이를 8kHz의 sampling율을 갖도록 downsampling 한다. 그리고 저대역 신호는 GSM-EFR 음성부호화 알고리즘을 이용하여 부호화하고, 고대역 신호는 DWT(Discrete Wavelet Transform)을 적용하여 subband로 나누어 부호화하였다. 각 subband에서 양자화 된 파라미터는 IDWT(Inverse DWT)과정을 거쳐서 upsampling되고 합성 QMF를 통과시켜 최종 합성음을 구하였다. 제안한 음성부호화기는 저대역 신호의 GSM-EFR 부호화에 12.2 kbps, 웨이브렛 변환을 이용한 고대역 신호의 부호화에 7.8 kbps로 전체 20 kbps의 전송율을 가지면서 G.722 표준안의 56 kbps에서의 합성음과 비슷한 음질을 나타내었다.

• Speech Sciences
• /
• v.9 no.1
• /
• pp.39-47
• /
• 2002

This paper proposes an effective method to remedy the acoustic modeling problem inherent in the usual log-normal Parallel Model Composition intended for achieving robust speech recognition. In particular, the Gaussian kernels under the prescribed log-normal PMC cannot sufficiently express the corrupted speech distributions. The proposed scheme corrects this deficiency by judiciously selecting the 'fairly' corrupted component and by re-estimating it as a mixture of two distributions using data-driven PMC. As a result, some components become merged while equal number of components split. The determination for splitting or merging is achieved by means of measuring the similarity of the corrupted speech model to those of the clean model and the noise model. The experimental results indicate that the suggested algorithm is effective in representing the corrupted speech distributions and attains consistent improvement over various SNR and noise cases.

• International Journal of Fluid Machinery and Systems
• /
• v.8 no.3
• /
• pp.209-219
• /
• 2015

A robust multi-fidelity optimization methodology has been developed, focusing on efficiently handling industrial runner design of hydraulic Francis turbines. The computational task is split between low- and high-fidelity phases in order to properly balance the CFD cost and required accuracy in different design stages. In the low-fidelity phase, a physics-based surrogate optimization loop manages a large number of iterative optimization evaluations. Two derivative-free optimization methods use an inviscid flow solver as a physics-based surrogate to obtain the main characteristics of a good design in a relatively fast iterative process. The case study of a runner design for a low-head Francis turbine indicates advantages of integrating two derivative-free optimization algorithms with different local- and global search capabilities.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.29 no.12 s.243
• /
• pp.1567-1573
• /
• 2005

This paper introduces a new active magnetic bearing(AMB) that can provide both radial and axial control functions in one bearing unit without axial disk. It has a structure of double four-pole AMB or a four-pole AMB where each core is split into two axially. The cores have two kinds of coil winding; they independently generate fluxes on the planes perpendicular or parallel to the shaft. For the radial control action, it works just like a conventional four-pole AMB. Meanwhile, for the axial control, it uses the Lorentz force generated by the interaction of the bias flux for radial control and the axial control flux. In this paper, the proposed structure, principle, and design process based on magnetic flux analysis are introduced, and its feasibility is experimentally verified by using a simple PD control algorithm with a feedforward loop to compensate the coupled flux effect.

• Journal of the Korean Society of Combustion
• /
• v.6 no.2
• /
• pp.15-22
• /
• 2001

A pressure-based, unstructured finite volume method has been applied to couple the chemical kinetics and fluid dynamics and to capture effectively and accurately the steep gradient flame field. The pressure-velocity coupling is handled by two methodologies including the pressure-correction algorithm and the projection scheme. A stiff, operator-split projection scheme for the detailed nonequilibrium chemistry has been employed to treat the stiff reaction source terms. The conservative form of the governing equations are integrated over a cell-centered control volume with collocated storage for all transport variables. Computations using detailed chemistry and variable transport properties were performed for two laminar reacting flows: a counterflow hydrogen-air diffusion flame and a lifted methane-air triple flame. Numerical results favorably agree with measurements in terms of the detailed flame structure.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.22 no.4
• /
• pp.115-131
• /
• 1997

The problem addressed in this paper is to minimize the deviations of workloads assigned to worker and to maximize the total utilizations(degree of skill) between workers and machines simultaneously. Each worker can handle the set of machines with the different degree of skill and each machine has the workloads needed to be processed. Also, each worker has to be assigned to at least one machine with the minimum workload deviation. This problem can be formulated as a preemptive goal programming with generalized assignment constraints. The proposed algorithm consists of two phases. First, a capacitated circulation network is constructed to assign the workers to machines with the maximum total utilizations while considering workloads balance. Then, a refinement process is applied to the split machines to satisfy the feasibility condition. The real industrial application in a plastic extrusion manufacturer is included along with several computational experiments.

• Journal of Multimedia Information System
• /
• v.6 no.4
• /
• pp.173-178
• /
• 2019

Mongolian historical and cultural monuments on settlement areas of stone inscriptions, stone images, rock-drawings, remains of cities, architecture are still telling us their stories. These monuments depict the understanding of the word, philosophical and artistic outlook, beliefs, religion, national art, language, culture and traditions of Mongols [1]. Nowadays computer science, especially computer vision is applying in the other science fields. The main problem is how to apply and which algorithm can detect and classify the objects correctly. In this paper, we propose a method to detect object from Mongolian nomadic environment images. This work proposes a method for object detection that is the combination of the binary operations in the edge detection results. We found out the best method and parameters of state-of-the-art machine learning algorithms. In experimental result, we evaluate our results with 10-fold cross validation and split 66% strategies.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.31B no.5
• /
• pp.50-57
• /
• 1994

In the non-destructive evaluation techniques using ultrasonic signal, backscattering noise from grain interface decreases the SNR of received signal. In this paper, SSP(split-spectrum processing) based on the constant FBR decomposition method has been applied to enhance the SNR. This algorithm helps to find optimal parameters of filter bank through a simple theory and has an advantage that reduce the signal processing time compared with the conventional constant bandwidth decomposition method. In this experiment, the 304 stainless steel sample is heat-treated and received ultrasonic signal is processed by SSP using the constand bandwidth decomposition method and the constand FBR decomposition method enhanced the SNR by 1.4 dB and reduced the required number of filters by 4 compared with the constant bandwidth decomposition method.

• Earthquakes and Structures
• /
• v.13 no.3
• /
• pp.279-288
• /
• 2017

Having the ability to keep on yielding stable solutions in problems involving high potential of instability, composite time integration methods have become very popular among scientists. These methods try to split a time step into multiple sub-steps so that each sub-step can be solved using different time integration methods with different behaviors. This paper proposes a new composite time integration in which a time step is divided into two sub-steps; the first sub-step is solved using the well-known Newmark method and the second sub-step is solved using Simpson's Rule of integration. An unconditional stability region is determined for the constant parameters to be chosen from. Also accuracy analysis is perform on the proposed method and proved that minor period elongation as well as a reasonable amount of numerical dissipation is produced in the responses obtained by the proposed method. Finally, in order to provide a practical assessment of the method, several benchmark problems are solved using the proposed method.

• Proceedings of the Korea Contents Association Conference
• /
• 2004.05a
• /
• pp.234-240
• /
• 2004

Now that the e-commerce users are explosively increasing, there is always followed by a sharp increase in load from the e-commerce system and the heavy traffic on the network, where leads to the delayed service for the client's request. The natural consequences contain decreasing customer satisfaction and weakening the business's competitive position in markets. Therefore, we'll need to study the e-commerce system in due consideration of the operational efficiency and response speed. This study includes a design of e-commerce system, with a hierarchical structure based on the local server, which has the capability of caching necessary to distribute the system load, makes a proposal concerning a split web-cache algorithm devoted to the local web server to finally give an analysis of the performance through the appropriate trial.