• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2001.12a
• /
• pp.303-306
• /
• 2001

패턴 인식에서 선형 분류 가능한 경계면을 찾아 패턴을 분류하는 방법 중 가장 기본적인 방법은 퍼셉트론이라고 볼 수 있다. 하지만 선형 분류 불가능한 패턴에 대해서는 유용한 결과를 보여주지 못하였다. 먼저 제안된 퍼지 퍼셉트론은 베타영역 설정에 의해 수렴하지 못하는 특성을 보완하였다. 그러나 패턴의 순수한 전형성을 고려해 주지 못하는 단점이 있다. 이에 Crisp의 선형분류 특성과 퍼지의. 수렴특성을 합성하고자 Possibilistic 퍼셉트론을 제시한다.

• Journal of KIISE
• /
• v.44 no.3
• /
• pp.253-260
• /
• 2017

In computer graphics, high-quality lens flares have been generated using costly offline rendering. A recent matrix-based approximation has enabled generation of high-quality lens flares suitable for real-time applications, but its quality degrades due to the lack of nonlinear patterns of lens flares. This paper introduces a method for high-quality lens-flare rendering, which includes blending of both nonlinear as well as linear patterns. The nonlinear patterns are pre-rendered or photo-graphically captured offline and stored in a look-up table. The online stage reads only the pattern by looking up the table using a light angle, hence making its performance drop negligible while greatly improving the quality.

• Proceedings of the Korean Information Science Society Conference
• /
• 2003.10b
• /
• pp.70-72
• /
• 2003

대용량 데이터베이스로부터 순차 패턴을 발견하는 문제는 지식 발견 또는 데이터 마이닝(Data Mining) 분야에서 주요한 패턴 추출 문제이다. 순차 패턴은 추출 기법에 있어 연관 규칙의 Apriori 알고리즘과 비슷한 방식을 사용하며 그 과정에서 시퀀스는 해쉬 트리 구조를 통해 다루어 진다. 이러한 해쉬 트리 구조는 항목들의 정렬과 데이터 시퀀스의 지역성을 무시한 저장 구조로 단순 검색을 통한 다수의 복잡한 포인터 연산수행을 기반으로 한다. 본 논문에서는 이러한 해쉬 트리 구조의 단정을 보완한 다단게 선형 배치 트리(MLLT, Multi-level Linear Location Tree)를 제안하고, 다단계 선형 배치 트리를 이용한 효율적인 마이닝 메소드(MLLT-Join)를 소개한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2015.10a
• /
• pp.1467-1469
• /
• 2015

일반적인 홍채 인식 방법은 홍채 근육이 형성하는 고유 패턴에 기초하여 사용자를 식별한다. 홍채의 주요 기능은 자연광의 양에 따라 동공의 크기를 제어하는 것이다. 그러므로 홍채 패턴은 조명 조건의 변화에 따라 팽창 및 수축한다. 홍채 패턴의 변형에도 불구하고 강력한 인식의 정확도를 보장하기 위해 이전의 연구에서는 홍채 근육은 선형적으로 변형된다고 가정하여, 홍채의 트랙을 선형적이고 균일하게 할당하였다. 그러나 홍채 패턴은 실제로 특정 홍채 근육의 복잡한 움직임으로 인해 비선형적으로 변형된다. 이러한 문제점을 극복하기 위해 Dynamic Warping 방법을 사용하여 비선형성을 고려한 새로운 매칭 방법을 제안한다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.27 no.4A
• /
• pp.324-329
• /
• 2002

In the high-speed/low-power digital filter applications like wireless communication systems, canonical signed digit(CSD) linear phase finite impulse response(FIR) filter structures are widely investigated. In this paper, we propose a high-speed/low-power CSD linear phase FIR filter structure using vertical common sub-expression. In the conventional linear phase CSD filter, horizontal common sub-expressions are utilized due to the inherent horizontal common sub-expression of symmetrical filter coefficients. We use the fact that their MSBs are also equal since adjacent filter coefficients have similar values in the linear phase filter Through the examples, it is shown that our proposed structure is more efficient in case that precision of implementation is lower, and tap length are longer.

• The Korean Journal of Financial Management
• /
• v.15 no.2
• /
• pp.369-399
• /
• 1998

Most of the studies on stock price predictability using the linear model conclude that there are little possibility to predict the future price movement. But some anomalous patterns may be generated by remaining market inefficiency or regulation, market system that is facilitated to prevent the market failure. And these anomalous pattern, if exist, make them difficult to predict the stock price movement with linear model. In this study, I try to find the anomalous pattern using the ANN model. And by comparing the predictability of ANN model with the predictability of correspondent linear model, I want to show the importance of recognitions of anomalous pattern in stock price prediction. I find that ANN model could have the superior performance measured with the accuracy of prediction and investment return to correspondent linear model. This result means that there may exist the anomalous pattern that can't be recognized with linear model, and it is necessary to consider the anomalous pattern to make superior prediction performance.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.12 no.1
• /
• pp.14-18
• /
• 2002

The perceptron algorithm, which is one of a class of gradient descent techniques, has been widely used in pattern recognition to determine linear decision boundaries. However, it may not give desirable results when pattern sets are nonlinerly separable. A fuzzy version was developed to male up for the weaknesses in the crisp perceptron algorithm. This was achieved by assigning memberships to the pattern sets. However, still another drawback exists in that the pattern memberships do not consider class typicality of the patterns. Therefore, we propose a possibilistic approach to the crisp perceptron algorithm. This algorithm combines the linearly separable property of the crisp version and the convergence property of the fuzzy version. Several examples are given to show the validity of the method.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.31 no.12C
• /
• pp.1184-1193
• /
• 2006

Image interpolation has been widely used and studied in the various fields of image processing. There are many approaches of varying complexity and robustness. In this paper, a new distance weight is proposed for the conventional linear interpolation. In comparison with the conventional linear weight, the new distance weight uses a quadratic or cubic polynomial equation to reflect that the interpolated value should be influenced more by the value of closer pixels in an input image. In this paper, the new adaptive linear (NAL) interpolation, which considers patterns near the interpolated value, is also proposed. This algorithm requires a pattern weight, which is used to determine the ratio of reflection on local patterns, to obtain an interpolated image that exhibits better quality at various magnification factors (MF). In the computer simulation, not only did the NAL interpolation exhibit much lower computational complexity than conventional bicubic interpolation, it also improved peak signal-to-noise ratios (PSNR).

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.14 no.3
• /
• pp.707-714
• /
• 2010

Since an LFSR(linear feedback shift register) as a pattern generator has solely linear dependency in itself, it generates sequences by moving the bit positions for pattern generation. So the correlation between the generated patterns is high and thus reduces the possibility of fault detection. To overcome these problems many researchers studied to have goodness of randomness between the output test patterns. In this paper, we propose the new and effective method to construct phase shifter as PRPG(pseudo random pattern generator).

• Journal of the Korea Society of Computer and Information
• /
• v.18 no.7
• /
• pp.149-155
• /
• 2013

Commonly, one seeks a particular pattern suitable for stock cutting and the number of such patterns through linear programming. However, since the number of the patterns increases exponentially, it is nearly impossible to predetermine all the existing patterns beforehand. This paper thus proposes an algorithm whereby one could accurately predetermine the number of existing patterns by applying Suliman's feasible pattern method. Additionally, this paper suggests a methodology by which one may obtain exact polynomial-time solutions for feasible patterns without applying linear programming or approximate algorithm. The suggested methodology categorizes the feasible patterns by whether the frequency of first occurrence of all the demands is distributed in 0 loss or in various losses. When applied to 2 data sets, the proposes algorithm is found to be successful in obtaining the optimal solutions.