• Proceedings of the Korea Information Processing Society Conference
• /
• 2001.10b
• /
• pp.1025-1028
• /
• 2001

인터넷의 확산과 이용자의 급증으로 웜바이러스에 대한 문제가 최근에 크게 대두되고 있다. 스크립트 형 웜바이러스의 경우 제작이 쉬워 누구나 몇 시간의 학습을 통하여 바이러스를 제작찬 수 있다. 이러한 문제의 심각성은 최근 7개월 동안의 바이러스 통계에서도 나타나는데 전체 바이러스 중 평균 22.5%를 차지하고 있다. 이러한 웜바이러스를 차단하기 위해서 여러 가지 방법들이 사용되고 있으나 E-mail로 급속하게 퍼지는 웜바이러스의 확산을 차단하기 위해서 네트워크 기반의 시스템 보호방법이 요구되어지고 있다. 이에 본 논문에서는 알려지지 않은 웜바이러스로부터 내부 네트워크를 방어하기 위한 면역시스템을 제안한다. 자동화된 면역 시스템은 분산된 각각의 웜바이러스 탐지 시스템들이 새로운 바이러스 정보를 동적으로 공유할 수 있도록 하여 새로운 바이러스로부터 해당 시스템과 그 시스템이 속해 있는 내부 네트워크를 바이러스로부터 보호할 수 있도록 한다.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2005.04a
• /
• pp.161-165
• /
• 2005

In order to produce an optimal traffic cycle. We must first check how many waiting cars are at the lower intersection, because waiting queue is bigger than the length of upper traffic intersection. Start up delay time and vehicle waiting time occurs. To reduce vehicle waiting time, in this paper, we present an optimal green time algorithm using fuzzy neural network. Through computer simulation has been proven to be improved average vehicle speed than fixed traffic signal light which do not consider different intersection conditions.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2001.10a
• /
• pp.755-758
• /
• 2001

본 논문에서는 서로 다른 크기의 클러스터에 대해서 효과적으로 데이터를 분류할 수 있는 내부클러스터를 이용한 개선된 FCM 알고리즘을 제안하였다. 내부클러스터는 평균내부거리 안쪽에 속하는 데이터 집합으로 클러스터의 크기와 밀도에 비례한다. 그러므로 이를 이용한 개선된 FCM 알고리즘은 기존의 FCM 알고리즘이 클러스터 크기가 다를 경우 퍼지분할과 중심탐색을 제대로 하지 못하는 문제점을 개선한 수 있다. 실험을 통하여 개선된 FCM 알고리즘이 분류 엔트로피에 의해 기존의 FCM 알고리즘 보다 더 좋은 결과를 나타냄을 알 수 있었다.

• Proceedings of the KIEE Conference
• /
• 1993.07a
• /
• pp.383-386
• /
• 1993

In this paper, we propose a sensory data combination method by a fuzzy number approach for multisensor data fusion. Generally, the weighting of one sensory data with respect to another is derived from measures of the relative reliabilities of the two sensory modules. But the relative weight of two sensory data can be approximately determined through human experiences or insufficient experimental data without difficulty. We represent these relative weight using appropriate fuzzy numbers as well as sensory data itself. Using the relative weight, which is subjective valuation, and a fuzzy-numbered sensor data, the fuzzy weighted average method is used for a representative sensory data. The manipulation and calculation of fuzzy numbers can be carried out using the Zadeh's extension principle which can be approximately implemented by the $\alpha$-cut representation of fuzzy numbers and interval analysis.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2016.10a
• /
• pp.517-519
• /
• 2016

본 논문에서는 적응적 이진화 기법을 적용하여 흠집 영역을 검출한다. 제안된 방법은 안경 렌즈 영상에서 명암 대비를 적용하여 렌즈의 명암을 강조한다. 명암이 강조된 영상에서 렌즈 밖의 배경 영역은 흠집 검출에 불필요하므로 개선된 평균 이진화 기법을 적용한 후에 렌즈의 윤곽선을 검출하여 렌즈 이외의 배경을 제거한다. 렌즈 이외의 배경이 제거된 렌즈 영상에서 렌즈 내부에 명암대비를 적용하여 렌즈 내부의 배경과 흠집의 명암을 강조한다. 명암이 강조된 렌즈 내부 영역에서 적응적 이진화 기법을 적용하여 흠집과 잡음을 검출한다. 잡음은 중간값 필터를 적용하여 제거한 후에 흠집 영역을 추출한다. 추출된 흠집 영역을 렌즈의 중심으로부터의 거리와 흠집의 크기를 퍼지 추론 규칙을 적용하여 눈에 미치는 영향 정도를 분석한다. 본 논문에서 제안된 방법의 성능을 분석하기 위해 CHEMI, MID, HL, HM과 같은 시력 보정용 렌즈 영상 6장을 대상으로 실험한 결과, 제안된 방법이 기존 렌즈 흠집 추출 방법보다 흠집 영역이 정확하게 추출되었고 눈에 미치는 영향을 효과적으로 분석할 수 있는 가능성을 확인하였다.

• Journal of Broadcast Engineering
• /
• v.23 no.1
• /
• pp.104-114
• /
• 2018

Recently, as HTTP-based video streaming traffic continues to increase, Dynamic Adaptive Streaming over HTTP(DASH), which is one of the HTTP-based adaptive streaming(HAS) technologies, is receiving attention. Accordingly, many video quality control techniques have been proposed to provide a high quality of experience(QoE) to clients in a DASH environment. In this paper, we propose a new quality control method using ANFIS(Adaptive Network based Fuzzy Inference System) which is one of the neuro-fuzzy system structure. By using ANFIS, the proposed scheme can find fuzzy parameters that selects the appropriate segment bitrate for clients. Also, considering the characteristic of VBR video, the next segment download time can be more accurately predicted using the actual size of the segment. And, by using this, it adjusts video quality appropriately in the time-varying network. In the simulation using NS-3, we show that the proposed scheme shows higher average segment bitrate and lower number of bitrate-switching than the existing methods and provides improved QoE to the clients.

• Journal of Convergence for Information Technology
• /
• v.8 no.5
• /
• pp.131-135
• /
• 2018

The heart disease taking the second place of the cause of the death of modern people is a terrible disease that makes sudden death without noticing. To judge the aortic valve disease of heart diseases a name of disease was diagnosed using psychological data provided from physioNet. Aortic valve is a valve of the area that blood is spilled from left ventricle to aorta. Aortic stenosis of heart troubles is a disease when the valve does not open appropriately in contracting the left ventricle to aorta due to narrowed aortic valve. In this paper, 3126 samples of cardiac sound data were used as an experiment data composed of 180 characteristics including normal people and aortic valve stenosis patients. To diagnose normal and aortic valve stenosis patients, NEWFM was utilized. By using an average method of weight as an feature selection method of NEWFM, the result shows 91.0871% accuracy.

• The Journal of the Korea Contents Association
• /
• v.5 no.5
• /
• pp.287-295
• /
• 2005

This paper describes user context awareness system, which is one of the most essential technologies in various application services of ubiquitous computing. The proposed system used two-akial accelerometer, embedded in SenseWear(R)PRO2 Armband (BodyMedia). When it was worn on the right upper arm of the experiment subjects, MAD (mean of absolute difference) value of the sensor data was calculated to quantify the amount of the wear's activity. Using this data, PC-based fuzzy inference system was realized to distinguish human motion states, such as, lying, sitting, walking and running and to recognize the restricted emergency situations. In laboratory experiment, the amount of activities for tying, sitting, walking and running were 0.204 g/s, 0.373 g/s, 2.808 g/s and 16.243 g/s respectively. The recognition rates of human motion states were 96.7 %, 93.0 %, 95.2 % and 98.4 % respectively for lying, sitting, walking and running. The recognition rate of restricted emergency situation was 100%.

• Journal of the Korean GEO-environmental Society
• /
• v.9 no.5
• /
• pp.83-89
• /
• 2008

This study is to figure out the speed of soil diffusion and shapes as well when landslide occurred. It was measured the speed and shapes of soil diffusion using installed indoor-simulation of a land slide. Standard sand was used to measure it and this measurement is following the moisture content of standard sand. The result of this experiment follows the moisture content; as the moisture content goes up, the area and speed of soil diffusion goes faster as well. As the average of records by the moisture content was analyzed, the area and speed of land slide are increased proportionally. The shape of diffusion is formed as fanwise. It depends on the moisture content as above; the speed was fast when the moisture content is high, and it was getting narrower and longer as the speed of diffusion was faster.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.9 no.11
• /
• pp.1221-1226
• /
• 2014

In an effort to improve the efficiency of the speech recognition, we investigate the effect of the number of phoneme clusters. For this purpose, codebooks of varied number of phoneme clusters are prepared by modified k-means clustering algorithm. The subsequent processing is fuzzy vector quantization (FVQ) and hidden Markov model (HMM) for speech recognition test. The result shows that there are two distinct regimes. For large number of phoneme clusters, the recognition performance is roughly independent of it. For small number of phoneme clusters, however, the recognition error rate increases nonlinearly as it is decreased. From numerical calculation, it is found that this nonlinear regime might be modeled by a power law function. The result also shows that about 166 phoneme clusters would be the optimal number for recognition of 300 isolated words. This amounts to roughly 3 variations per phoneme.