• Proceedings of the Korea Information Processing Society Conference
• /
• 2007.05a
• /
• pp.305-308
• /
• 2007

본 논문은 교통표지판을 자동으로 인식하는 방법에 관한 연구로 기존의 교통표지판 인식시스템에서는 인식하는데 걸리는 시간이 길고 잡음환경에서 인식률이 저하되며 변경된 교통표지판은 인식하지 못하는 문제점이 있다. 본 논문에서는 이와 같은 문제점을 해결하기위해 컬러정보를 이용하여 교통표지판 영역을 추출하고 추출된 이미지를 인식하는데 다층퍼셉트론 신경망 알고리즘을 적용하여 교통표지판 인식시스템을 제안한다. 제안된 방법은 교통표지판의 컬러를 분석하여 영상에서 교통표지판 영역을 추출한다. 영역을 추출하는 방법은 RGB 컬러 공간으로부터 YUV, YIQ, CMYK 컬러 공간이 가지는 특성을 이용한다. 형태처리는 교통표지판의 기하학적 특성을 이용하여 군집화한다. 교통표지판 인식은 학습이 가능한 다층퍼셉트론의 오류역전파알고리즘을 적용하여 인식한다. 다층퍼셉트론 신경망 알고리즘은 패턴인식 분야에서 우수한 성능이 입증 되었다.

• Proceedings of the KIEE Conference
• /
• 2008.07a
• /
• pp.2065-2066
• /
• 2008

GIS 예방진단 시스템을 현장에 적용할 경우 예상치 못한 노이즈와 전기외란으로 예측 일치율에 저하가 발생한다. 본 논문에서는 각 현장에 설치된 온라인 GIS예방진단 시스템의 기록정보를 받아들여 새롭게 학습을 시킴으로써 현장 적응형 진단을 구축할 수 있는 방안을 제시한다. 또한 현장의 각종 노이즈 패턴도 학습시킴으로써 진단엔진의 성능향상을 도모한다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.26 no.8B
• /
• pp.1141-1151
• /
• 2001

본 논문에서는 효과적으로 선택된 초기 탐색 위치를 이용한 움직임 추정의 고속 알고리즘을 제안하였다. 제안된 알고리즘에서는 2$\times$2화소 블록 평균으로 부표본화 영상에서 움직임 벡터를 얻어 원영상 비율로 확대하고, 주위 블록의 움직임으로부터 예측 움직임 벡터를 구하여, 이 중에서 정합오차가 작은 것을 초기 탐색 위치로 선택한다. 그리고 선택된 초기 탐색 위치를 중심으로 회전 탐색을 시작하여, 연속 소거 알고리즘으로 탐색할 후보 블록을 선택하고, 부분 정합 왜곡 소거법을 사용하여 블록간 정합오차 계산량을 줄이면서, 고속으로 움직임 벡터를 추정한다. 알고리즘의 실제 적용에 있어서는 선택된 초기 탐색 위치를 중심으로 회전 탐색 패턴의 탐색 범위를 조절하거나, 매크로 블록 당 복잡도를 제한하여 계산량을 줄일 수 있다. 실험 결과, 제안된 알고리즘은 전역탐색 블록정합 알고리즘에 대하여 0.2dB 이하의 미소한 평균 PSNR 저하만을 발생하면서, FBMA 복잡도의 3% 이하의 평균 복잡도를 소요하였다. 이것은 3단계 탐색법에 대하여 40% 이하의 계산량이다. 그리고 실험 영상들의 각 프레임에 대해서도 비슷한 성능을 보임을 확인하였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2002.04b
• /
• pp.911-914
• /
• 2002

최근 들어 네트워크 보안 사고가 잦아지고 있다. 이는 내부/외부로부터 정보시스템에 대한 불법침입과 공격시도 증가로 알려지고 있다. 이러한 이유는 기존의 침입탐지시스템의 신뢰도와 성능 저하를 야기시켰다. 에이전트는 자율성과 독립적 특성을 갖는다. 에이전트 종류는 Mobile agent, Multiagent, Assistant agent, User Interface agent, Intelligent agent가 있다. 기존의 침입탐지시스템은 실시간 침입탐지를 제공하지 못한다. 그래서 본 논문에서는 에이전트를 기반으로 하는 다양한 침입 패턴에 대응하는 침입탐지시스템의 연구 결과들을 분석하여 새로운 침입 탐지시스템 설계의 발판을 제공한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2012.04a
• /
• pp.533-536
• /
• 2012

인터넷의 급격한 성장과 함께 컴퓨터 통신 이용률이 폭발적으로 증가함에 따라 여러 악성코드가 등장하게 되었다. 이러한 악성코드는 시스템의 비정상 동작 유발, 네트워크 성능 저하, 개인정보유출의 문제를 발생시킨다. 현재의 악성코드 분석은 Signature 분석이 대부분이며, Signature 분석은 특정 패턴의 악성코드는 빠르게 탐지하나, 변조된 코드는 탐지하지 못하며, 이미 피해가 널리 퍼진 뒤 분석 및 차단이 가능하다는 단점을 가진다. 따라서 본 논문은 NDIS(Network Driver Interface Specification)를 이용하여 악성코드에 대해 수동적인 Signature 분석의 단점을 보완 하는 시스템 및 네트워크 상태 분석모델을 제시 하여 보다 능동적인 탐지 및 차단 프로세스를 정의하고, 모델 구현을 위한 방법을 제시한다.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.36S no.1
• /
• pp.29-37
• /
• 1999

In general, the information processing capability of a neural network is determined by its architecture and efficient training patterns. However, there is no systematic method for designing neural network and selecting effective training patterns. Evolutionary Algorithms(EAs) are referred to as the methods of population-based optimization. Therefore, EAs are considered as very efficient methods of optimal system design because they can provide much opportunity for obtaining the global optimal solution. In this paper, we propose a new method for finding the optimal structure of neural networks based on competitive co-evolution, which has two different populations. Each population is called the primary population and the secondary population respectively. The former is composed of the architecture of neural network and the latter is composed of training patterns. These two populations co-evolve competitively each other, that is, the training patterns will evolve to become more difficult for learning of neural networks and the architecture of neural networks will evolve to learn this patterns. This method prevents the system from the limitation of the performance by random design of neural networks and inadequate selection of training patterns. In co-evolutionary method, it is difficult to monitor the progress of co-evolution because the fitness of individuals varies dynamically. So, we also introduce the measurement method. The validity and effectiveness of the proposed method are inspected by applying it to the visual servoing of robot manipulators.

• Journal of Advanced Navigation Technology
• /
• v.24 no.6
• /
• pp.593-600
• /
• 2020

The equipment for marine IoT service have to overcome the effect of seawater. Furthermore, the free floating transmitter in seawater will be less affected by the seawater environment. The results of the design and fabrication of antenna, which is embedded in buoy, are shown in this research. The proposed antenna is used to supervise the states of fishing gears in monitoring system for real-name system of electric fishing gear. The selected frequency band of the proposed antenna is 920 MHz, and PCB pattern type is selected for subminiature and light weight. It is confirmed that RF characteristics is more degraded, however, the radiation is gradually upward as the contact surface of buoy with seawater is more broaden through the simulation results. That is, the RF performance of the proposed antenna is more deteriorated but beam radiation characteristics is more suited the marine IoT, the seawater effect is more increased. It is expected that the proposed antenna will contribute the implementation of IoT network based on low power wide area (LPWA) when the degradation of RF performance will be settled.

• Journal of KIISE:Computing Practices and Letters
• /
• v.16 no.4
• /
• pp.482-486
• /
• 2010

Recently, gesture recognition methods based on dynamic time warping (DTW) have been actively investigated as more mobile devices have equipped the accelerometer. DTW has no additional training step since it uses given samples as the matching templates. However, it is difficult to apply the DTW on mobile environments because of its computational complexity of matching step where the input pattern has to be compared with every templates. In order to address the problem, this paper proposes a gesture recognition method based on DTW that uses localized subset of templates. Here, the k-means clustering algorithm is used to divide each class into subclasses in which the most centered sample in each subclass is employed as the localized template. It increases the recognition speed by reducing the number of matches while it minimizes the errors by preserving the diversities of the training patterns. Experimental results showed that the proposed method was about five times faster than the DTW with all training samples, and more stable than the randomly selected templates.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.51 no.6
• /
• pp.228-236
• /
• 2014

Towed array SONAR is deformed because it operates in fluid such as an ocean. It especially undergoes significant change in shape as a towing vessel takes a turn. In this case, beam pattern synthesis of the line array is limited, resulting in degradation in quality such as signal-to-noise ratio. This paper presents a modified two-step least squares algorithm based on the two-step least squares method. The shape of the sea-operated line array formation with the towing vessel changing course(angle) was modeled and the algorithm was subsequently applied. While changing course and location of the main lobe in beam pattern was altered, signal-to-noise ratio of steering beam pattern synthesis was analyzed by algorithm (proposed and others). As a result, the proposed algorithm presented improvement in performance by 2dB compared to other algorithms while forming relatively constant beam pattern.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.29 no.10
• /
• pp.752-757
• /
• 2018

This study presents the performance evaluation results of a 77-GHz waveguide slot array antenna that was fabricated by attaching a patterned printed circuit board(PCB) on a metal block. The 77-GHz waveguide was divided into a top plate and a bottom structure. The top plate was fabricated using a patterned PCB that can implement a fine slot at low cost. The top cover was then bonded to the bottom metal structure with a waveguide trough using anisotropic conductive film. For evaluating the antenna performance, a $1{\times}16$ slot array antenna was fabricated using our proposed method and the gain and pattern were measured and compared with the simulation results. Though the measurement results demonstrate a reduction in gain of around 2.3~3.5 dB compared to the simulation results assuming ideal bonding conditions, the pattern hardly changed and the slot antenna with a gain of approximately 17 dBi at 77 GHz can be easily manufactured at a low cost using the proposed method.