• 한국HCI학회:학술대회논문집
• /
• 2008.02a
• /
• pp.433-438
• /
• 2008

본 논문은 비주얼 헐 오브젝트를 이용한 움직이는 오브젝트에 대한 실시간 영상기반 라이팅 기술을 제안한다. 본 논문에서는 특히 서로 다른 공간상의 조명 환경을 일치시키는 기술에 중점을 두고, 실시간으로 움직이는 오브젝트의 실시간 비디오 기반 재조명 기술로서 3가지 핵심 내용을 소개한다. 첫째는 비주얼 헐 데이터를 기반으로 기존에 벡터의 외적을 사용하던 방법을 개선하여 수식을 근사화시켜 연산량을 줄여서 고속으로 노말 벡터를 추출하는 방법이고, 둘째는 사용자 주변 조명 환경 정보를 효과적으로 샘플링하여 라이팅에 사용하는 점광원의 개수를 줄였으며, 세 번째는 CPU와 GPU의 연산량을 분배하여 효과적으로 병렬 고속 연산이 가능하도록 하였다. 종래의 영상기반 라이팅 기술이 정지된 환경맵 영상을 사용하거나 정지된 객체를 라이팅하였던 연구를 한 반면에 본 논문은 실시간에서 라이팅을 구현하기 위한 기술로서 고속 라이팅 연산을 위한 방법을 제시하고 있다. 본 연구의 결과를 이용하면 영상기반 라이팅 연구의 실제적이고도 폭넓은 작용이 가능할 것으로 사료되며 고화질의 콘텐츠 양산에도 기여할 것으로 사료된다.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.48 no.5
• /
• pp.123-128
• /
• 2011

In this paper, we design the CA-CFAR processor using a root-square approximation approach and a fixed-point operation to improve hardware complexity and reduce computational effort. We also propose CA-CFAR processor with multi-window, which is capable of concurrent parallel processing. The proposed architecture is synthesized and implemented into the FPGA and the performance is compared with the conventional processor designed by root-square libarary licensed by FPGA corporation.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.33 no.5
• /
• pp.761-773
• /
• 2023

This paper proposes a method to increase the power-analysis resistance of the neural network model's feedforward process by replacing the exponential-based activation function, used in the deep-learning field, with an approximated function especially at the multi-layer perceptron model. Due to its nature, the feedforward process of neural networks calculates secret weight and bias, which already trained, so it has risk of exposure of internal information by side-channel attacks. However, various functions are used as the activation function in neural network, so it's difficult to apply conventional side-channel countermeasure techniques, such as masking, to activation function(especially, to exponential-based activation functions). Therefore, this paper shows that even if an exponential-based activation function is replaced with approximated function of simple form, there is no fatal performance degradation of the model, and than suggests a power-analysis resistant feedforward neural network with exponential-based activation function, by masking approximated function and whole network.

• Journal of the Korean Data and Information Science Society
• /
• v.24 no.3
• /
• pp.495-504
• /
• 2013

Evolutionary algorithms have been applied to multi-objective optimization problems by approximation methods using computational intelligence. Those methods have been improved gradually in order to generate more exactly many approximate Pareto optimal solutions. The paper introduces a new method using support vector machine to find an approximate Pareto frontier in multi-objective optimization problems. Moreover, this paper applies an evolutionary algorithm to the proposed method in order to generate more exactly approximate Pareto frontiers. Then a decision making with two or three objective functions can be easily performed on the basis of visualized Pareto frontiers by the proposed method. Finally, a few examples will be demonstrated for the effectiveness of the proposed method.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.21 no.4
• /
• pp.317-324
• /
• 2008

Optimum design for general or complex structures are required to the need of many numbers of structural analyses. However, current computational environment with single processor is not capable of generating a high-level efficiency in structural analysis and design process for complex structures. In this paper, a virtual parallel computing system communicated by an internet of personal computers and workstation is constructed. In addition, a routine executing Pro/E, ANSYS and optimization algorithm automatically are adopted in the distributed process technique of sequential approximate optimization for the purpose of enhancing the flexibility of application to general structures. By employing the distributed processing technique during structural analysis using commercial application, total calculation time could be reduced, which will enhance the applicability of the proposed technique to the general complex structures.

• Journal of the Korean Institute of Telematics and Electronics T
• /
• v.35T no.1
• /
• pp.33-39
• /
• 1998

This paper discusses approximation modelling of discrete-time linear time-varying system(LTVS). The wavelet transform is considered as a tool for representing and approximating a LTVS. The joint time-frequency properties of wave analysis are appropriate for describing the LTVS. Simulation results is included to illustrate the potential application of the technique.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.25 no.6A
• /
• pp.825-832
• /
• 2000

광섬유 통신 시스템이 고속화되고 장거리를 전송하게 될 수록 편광모드 분산의 중요성은 더욱 부각되어 있다. 따라서 본 논문에서는 복굴절 광섬유에서 비선형 광펄스의 전파특성을 편광 모드 분산의 영향을 고려하여 시뮬레이션하였으며 이러한 현상이 발생되는 것을 알 수 있었다. 그리고 광섬유 비션형성에 의해서 GVD(Group Velocity Dispersion)와 마찬가기로 PMD(Polarization Mode Dispersion)에서도 부분적인 보상 현상이 나타남을 수치 결과를 통해 알 수 있었다. 이러한 광 전송 시뮬레이션을 구현하기 위해서 기존의 단계분할 푸리에 방식 (SS-FM, Split-Step Fourier Method)보다 장거리 전송시 오차의 발생이 적은 단계 분할 유한 요소법)SS-FEM, Split-Step Finite Element Method)을 적용하였으며, 또한 그 단점인 수행 속도를 개선한 희귀 행렬 근사 단계 분할 유한 요소법을 제안하였다. 그 결과 제안된 방법이 기존의 푸리에 연산법이나 일반적인 유한 요소법과 비교하여 더 빠른 수행 속도를 나타내는 것을 알 수 있었다.

• Annual Conference of KIPS
• /
• 2011.04a
• /
• pp.374-375
• /
• 2011

보행자 탐지를 위해 많은 알고리즘들이 제안되었고 그 중 HOG 알고리즘은 가장 좋은 성능을 보이는 알고리즘으로 알려져 있다. 하지만 HOG(Histogram of Oriented Gradients) 알고리즘은 연산량이 많아 계산 속도가 느려 실시간 시스템에 적용하기는 힘들다. 본 논문은 HOG 알고리즘으로 얻어진 특징 벡터를 이용해 보행자를 인식하는 방법의 속도 개선에 대하여 연구하였다. 기존 HOG 알고리즘에서 계산량이 많은 곳이 어느 부분인지 분석하고, 그 중 기울기와 방향을 계산하는 부분의 근사화를 통해 계산 속도를 높이는 방법을 제안한다.

• Journal of IKEEE
• /
• v.26 no.4
• /
• pp.736-741
• /
• 2022

Lattice-based cryptography is the most practical post-quantum cryptography because it enjoys strong worst-case security, relatively efficient implementation, and simplicity. Ring learning with errors (R-LWE) is a public key encryption (PKE) method of lattice-based encryption (LBC), and the most important operation of R-LWE is the modular polynomial multiplication of rings. This paper proposes a method for optimizing modular multipliers based on approximate computing (AC) technology, targeting the medium-security parameter set of the R-LWE cryptosystem. First, as a simple way to implement complex logic, LUT is used to omit some of the approximate multiplication operations, and the 2's complement method is used to calculate the number of bits whose value is 1 when converting the value of the input data to binary. We propose a total of two methods to reduce the number of required adders by minimizing them. The proposed LUT-based modular multiplier reduced both speed and area by 9% compared to the existing R-LWE modular multiplier, and the modular multiplier using the 2's complement method reduced the area by 40% and improved the speed by 2%. appear. Finally, the area of the optimized modular multiplier with both of these methods applied was reduced by up to 43% compared to the previous one, and the speed was reduced by up to 10%.

• Journal of KIISE:Computer Systems and Theory
• /
• v.36 no.6
• /
• pp.543-556
• /
• 2009

Recently, NAND flash memory is being used as the swap device of virtual memory as well as the file storage of mobile systems. Since temporal locality is dominant in page references of virtual memory, LRU and its approximated CLOCK algorithms are widely used. However, cost of a write operation in flash memory is much larger than that of a read operation, and thus a page replacement algorithm should consider this factor. This paper analyzes virtual memory read/write reference patterns individually, and observes the ranking inversion problem of temporal locality in write references which is not observed in read references. With this observation, we present a new page replacement algorithm considering write frequency as well as temporal locality in estimating write reference behaviors. This new algorithm dynamically allocates memory space to read/write operations based on their reference patterns and I/O costs. Though the algorithm has no external parameter to tune, it supports optimized implementations for virtual memory systems, and also performs 20-66% better than CLOCK, CAR, and CFLRU algorithms.