• Proceedings of the IEEK Conference
• /
• 2005.11a
• /
• pp.1189-1192
• /
• 2005

본 논문에서는 스케치 연산자를 적용하여 견실한 얼굴인식 방법을 제안한다. 제안된 방법은 인식 대상의 중요한 특성인 에지(edge), 벨리(valley) 및 질감(texture) 성분을 효과적으로 표현하기 위한 방법으로써, BDIP(block difference of inverse probabilities)를 사용하여 얼굴의 특징을 스케치 영상과 같이 나타내는 얼굴 영상을 획득한다. 그리고, BDIP 처리된 얼굴 영상은 입력 데이터의 차원 축소 및 얼굴 특징 벡터의 추출을 위해 PCA(Principal Component Analysis)를 수행한 후, Nearest Neighbor 분류기를 통해 인식을 수행한다. 제안된 방법의 성능을 평가하기 위하여, 일반적으로 많이 사용되는 HE(Histogram equalization)을 사용한 얼굴 인식 방법과의 비교를 수행한다. 실험결과, 본 논문에서 제안한 방법이 고유값이 적은 경우에 가장 높은 인식률을 나타내는 것을 알 수 있었다.

• Proceedings of the IEEK Conference
• /
• 2001.09a
• /
• pp.747-750
• /
• 2001

The paper presents an algorithm for estimating the variance of additive zero mean Gaussian noise in an image. The algorithm uses the wavelet transform which is a good tool for energy compaction. The algorithm consists of three steps. At first, high frequency components, wavelet coefficients in HH band, are generated from a noisy image by the wavelet transform. In a second step, high frequency components which are out of the noise range ate eliminated. Finally, if the image has many components eliminated in the previous step, then its noise estimated value is reduced. Experimental results show that the wavelet filter has better performance than the other high pass filters such as a Laplacian filter, residual from a median filter, residual from a mean filter, and a difference operator. In various images, the algorithm reduces 50% of estimated error on an average.

• The Journal of Korea Robotics Society
• /
• v.15 no.1
• /
• pp.32-38
• /
• 2020

This paper describes a method for tracking attitude and position of underwater robots. Underwater work with underwater robots is subject to differences in work efficiency depending on the skill of the operator and the utilization of additional sensors. Therefore, this study developed an underwater robot that can operate autonomously and maintain a certain attitude when working underwater to reduce difference of work efficiency. The developed underwater robot uses 8 thrusters to control 6 degrees of freedom motion, IMU (Inertial Measurement Unit), DVL (Doppler Velocity Log) and PS (Pressure Sensor) to measure attitude and position. In addition, the thruster allocation algorithm was designed to follow the control desired value using 8 thrusters, and the motion control experiments were performed in the engineering water basin using the thruster allocation method.

• Bulletin of the Korean Mathematical Society
• /
• v.36 no.3
• /
• pp.543-564
• /
• 1999

Consider a symmetric bilinear form defined on $\prod\times\prod$ by $_{\lambda\mu}$ = $<\sigma,fg>\;+\;\lambdaL[f](a)L[g](a)\;+\;\muM[f](b)m[g](b)$ ,where $\sigma$ is a quasi-definite moment functional, L and M are linear operators on $\prod$, the space of all real polynomials and a,b,$\lambda$ , and $\mu$ are real constants. We find a necessary and sufficient condition for the above bilinear form to be quasi-definite and study various properties of corresponding orthogonal polynomials. This unifies many previous works which treated cases when both L and M are differential or difference operators. finally, infinite order operator equations having such orthogonal polynomials as eigenfunctions are given when $\mu$=0.

• Bulletin of the Korean Mathematical Society
• /
• v.40 no.4
• /
• pp.565-576
• /
• 2003

In this paper, we determine the general solution of the quartic equation f(x+2y)+f(x-2y)+6f(x) = 4[f(x+y)+f(x-y)+6f(y)] for all x, $y\;\in\;\mathbb{R}$ without assuming any regularity conditions on the unknown function f. The method used for solving this quartic functional equation is elementary but exploits an important result due to M. Hosszu [3]. The solution of this functional equation is also determined in certain commutative groups using two important results due to L. Szekelyhidi [5].

• Journal of KIISE:Computer Systems and Theory
• /
• v.26 no.6
• /
• pp.627-635
• /
• 1999

물체의 3차원 모델을 음영처리하기 위해서는 물체 표면의 각 점에서 법선 벡터를 계산해야 한다. 복섹 기반의 볼륨 데이터는 표면에 대한 기하학적 정보가 없기 때문에 이웃 점들의 상대적인 위치나 데이터 값의 차이로부터 법선 벡터를 추정할 수 밖에 없다. 기존에 고안된 법선 벡터추정 연산자는크기가 고정되어 있기 때문에 제한된 영역에서만 법선 벡터를 정확하게 계산하고 나머지 영역에서는 오류를 유발한다. 이 논문에서는 표면을 구성하는 점들의 공간적 배치나 그 점들의 데이터값에 따라 크기가 변하는 가변 중심 편차 연산자와 이를 이용한 법선 벡터 추정 방법을 제안한다. 이 연산자를 사용하면 기존연산자들보다 정확하게 법선 벡터를 추정할 수 있으며, 동일한 화질인 경우 계산 시간이 상당히 단축된다.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.1524-1529
• /
• 2005

Particle filter is a kind of conditional density propagation model. Its similar characteristics to both selection and mutation operator of evolutionary strategy (ES) due to its Bayesian inference rule structure, shows better performance than any other tracking algorithms. When a new object is entering the region of interest, particle filter sets which have been swarming around the existing objects have to move and track the new one instantaneously. Moreover, there is another problem that it could not track multiple objects well if they were moving away from each other after having been overlapped. To resolve reinitialization problem, we use competitive-AVQ algorithm of neural network. And we regard interfarme difference (IFD) of background images as potential field and give priority to the particles according to this IFD to track multiple objects independently. In this paper, we showed that the possibility of real-time object tracking as intelligent interfaces by simulating the deformable contour particle filters.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.1863-1866
• /
• 2005

In this paper we propose a platform that is applicable to low altitude remote sensing. Basic idea of the platform is based on the model helicopter. On big difference from the conventional model helicopter is that our platform has four main rotors. Furthermore, vision control strategy is introduced so that operator can use the platform without any specialized intensive knowledge

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1997.10a
• /
• pp.413-416
• /
• 1997

Recently, dynamic recurrent neural networks(DRNN) for identification of nonlinear dynamic systems have been researched extensively. In general, dynamic backpropagation was used to adjust the weights of neural networks. But, this method requires many complex calculations and has the possibility of falling into a local minimum. So, we propose a new approach to identify nonlinear dynamic systems using DRNN. In order to adjust the weights of neurons, we use evolution strategies, which is a method used to solve an optimal problem having many local minimums. DRNN trained by evolution strategies with mutation as the main operator can act as a plant emulator. And the fitness function of evolution strategies is based on the difference of the plant's outputs and DRNN's outputs. Thus, this new approach at identifying nonlinear dynamic system, when applied to the simulation of a two-link robot manipulator, demonstrates the performance and efficiency of this proposed approach.

• ETRI Journal
• /
• v.31 no.4
• /
• pp.387-392
• /
• 2009

This paper introduces a novel parallel shift operator finite-difference time-domain (SO-FDTD) method for plasma in the dispersive media. We calculate the interaction between the electromagnetic wave of various frequencies and non-magnetized plasma by using the parallel SO-FDTD method. Then, we compare the results，which are calculated with serial and parallel SO-FDTD executions to obtain the speedup ratio and validate the parallel execution. We conclude that the parallel SO method has almost the same precision as the serial SO method, while the parallel approach expands the scope of memory and reduces the CPU time.