• 한국HCI학회:학술대회논문집
• /
• 한국HCI학회 2009년도 학술대회
• /
• pp.146-149
• /
• 2009

실시간 영상에서 사람의 얼굴 검출은 얼굴 인식분야에 있어서 주요한 관심 분야 중의 하나이다. 본 논문에서는 실시간 입력되는 영상에서 피부 색상과 Haar-like feature를 이용한 얼굴 검출 및 추적 알고리듬을 제안한다. 제안된 알고리듬은 컬러 색 공간에서 피부색상과 특징점을 가지고 얼굴 영역 및 추적하였다. 실험 결과 실시간 영상에 대해 조명 변화 및 가림 현상에서 강건한 추적 결과를 얻을 수 있었다.

• 대한수학회지
• /
• 제36권6호
• /
• pp.1145-1168
• /
• 1999

The purpose of this paper is to study the complete submanifolds with restricted space-like and time-like holomorphic bisectional curvatures in an indefinite locally symmetric K hler manifold.

• 정보처리학회논문지A
• /
• 제19A권4호
• /
• pp.175-180
• /
• 2012

본 논문은 Haar-like 특징을 이용한 고성능 보행자 및 차량 인식 회로를 제안한다. 제안하는 회로는 영상의 매 프레임 마다 슬라이딩 윈도우를 적용하여 Haar-like 특징을 추출하고 보행자 및 차량을 인식한다. Haar-like 특징 추출 회로는 슬라이딩 윈도우 당 200개의 Haar-like 특징을 추출하며, 추출된 특징들은 AdaBoost 인식 회로에서 사용된다. 제안하는 회로는 속도 향상을 위해 병렬 회로 구조를 적용하였으며 두 개의 슬라이딩 윈도우가 동시에 보행자 또는 차량을 인식한다. 제안하는 고성능 보행자 및 차량 인식 회로는 Verilog HDL로 설계하였으며 130nm 표준 셀 라이브러리를 이용하여 게이트 수준의 회로로 합성하였다. 합성된 회로는 1,388,260개의 게이트로 구성되며 최대 동작 주파수는 203MHz이다. 제안하는 회로는 $640{\times}480$ 영상을 초당 약 47.8장 처리할 수 있기 때문에 보행자와 차량을 실시간으로 인식하기 위해 사용될 수 있다.

• 제어로봇시스템학회논문지
• /
• 제9권3호
• /
• pp.186-195
• /
• 2003

This paper deals with eigenvalue assignment techniques for linear time-varying systems as a way of achieving feedback stabilization. For this, the novel eigenvalue concepts, which are the time-varying counterparts of the conventional (time-invariant) eigenvalue notions, are introduced. Then, the Ackermann-like formulae for SISO/MIMO linear time-varying systems are proposed. It is believed that these techniques are the generalized versions of the Ackermann formulae for linear time-invariant systems. The advantages of the proposed Ackermann-like formulae are that they neither require the transformation of the original system into the phase-variable form nor compute the eigenvalues of the original system. Two examples are given to demonstrate the capabilities of the proposed techniques.

• Smart Structures and Systems
• /
• 제21권2호
• /
• pp.195-205
• /
• 2018

Research on Lamb wave-based damage identification in plate-like structures depends on precise knowledge of dispersive wave velocity. However, boundary reflections with the same frequency of interest and greater amplitude contaminate direct waves and thus compromise measurement of Lamb wave dispersion in different materials. In this study, non-reflecting boundaries were proposed in both numerical and experimental cases to facilitate time-frequency characterization of Lamb wave dispersion. First, the Lamb wave equations in isotropic and laminated materials were analytically solved. Second, the non-reflecting boundaries were used as a series of frames with gradually increased damping coefficients in finite element models to absorb waves at boundaries while avoiding wave reflections due to abrupt property changes of each frame. Third, damping clay was sealed at plate edges to reduce the boundary reflection in experimental test. Finally, the direct waves were subjected to the slant-stack and short-time Fourier transformations to calculate the dispersion curves of phase and group velocities, respectively. Both the numerical and experimental results suggest that the boundary reflections are effectively alleviated, and the dispersion curves generated by the time-frequency analysis are consistent with the analytical solutions, demonstrating that the combination of non-reflecting boundary and time-frequency analysis is a feasible and reliable scheme for characterizing Lamb wave dispersion in plate-like structures.

• Steel and Composite Structures
• /
• 제21권3호
• /
• pp.501-515
• /
• 2016

Seismic analysis for steel frame structure with brace configuration using topology optimization based on truss-like material model is studied. The initial design domain for topology optimization is determined according to original steel frame structure and filled with truss-like members. Hence the initial truss-like continuum is established. The densities and orientation of truss-like members at any point are taken as design variables in finite element analysis. The topology optimization problem of least-weight truss-like continuum with stress constraints is solved. The orientations and densities of members in truss-like continuum are optimized and updated by fully-stressed criterion in every iteration. The optimized truss-like continuum is founded after finite element analysis is finished. The optimal bracing system is established based on optimized truss-like continuum without numerical instability. Seismic performance for steel frame structures is derived using dynamic time-history analysis. A numerical example shows the advantage for frame structures with brace configuration using topology optimization in seismic performance.

• 한국염색가공학회지
• /
• 제25권3호
• /
• pp.223-231
• /
• 2013

This research has a main focus on providing fundamental data for on-the-spot industrial fields by comparing and contrasting physical properties of fluffy spun-like material. The fluffy spun-like yarn is developed as fluffy yarn similar to natural spun-like yarn by treating polyester(FDY and + type shaped DTY) with ATY machine. In this experiment, using ATY machine for raw material texturing, we produced two fluffy yarns: (i) + type shaped(50d/36f, DTY) as core yarn and 100d/192f FDY as effect yarn[ATY(D)], (ii) FDY(75/36) as core yarn and 100d/192f FDY [ATY(F)] as effect yarn. After producing thous yarns, we twisted them with 500T/M, 700T/M, 1000T/M, respectively. produced yarns through this process were used as the samples for this experiment. Even though the shrinkage of fluffy yarn ATY(F) and ATY(D) becomes high as treated temperature rises and treated time lengthens, it is more affected by treated temperature then by treated time. In this experiment, produced fluffy yarn[ATY(D)] shows a little high values for temperature, but almost same values for higher temperatures. When we compare ATY(F) with ATY(D) fluffy yarn shows more natural fluffy yarn surface structure like natural cotton. The shrinkage of 700T/M twisted ATY(D) fluffy yarn show about 11% under treated temperature $180^{\circ}C$ and treated time 30min, and about 7% under $120^{\circ}C$ and 30min, respectively. But the shrinkage of 1000T/M fluffy yarn shoes about 9% and 6% under same conditions. Regarding treated time, tenacity and initial modulus of ATY(D) fluffy yarn rise high until 30min, but do not show much increase above 30min. Regarding treated temperature, tenacity and initial modulus of it rise high aboyer $140^{\circ}C$.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2001년도 ICCAS
• /
• pp.125.1-125
• /
• 2001

This paper deals with the eigenvalue assignment technique for linear time-varying systems to achieve feedback stabilization. For this, we introduce the novel eigenvalue concepts. Then, we propose the Ackerman-like formula for linear time-varying systems. It is believed that this technique is the generalized version of the Ackerman formula forlinear time-invariant systems. The advantages of the proposed technique are that it does not require the transformation of the original system into the phase-variable form nor the computation of eigenvalues of the original system.

• 한국정보시스템학회지:정보시스템연구
• /
• 제10권2호
• /
• pp.129-150
• /
• 2001

In this study, we model a decentralized supply chain system which is managed by four types of centers, sequentially located: Retailer, Wholesaler, Distributor, and Factory Each center contributes to enhancing the performance of the supply chain system individually with its own local inventory information. Through experiments which are performed with a programmed simulation (like the MIT beer game), we investigate how the information lead time improvement in each center affects the whole system. And we show that the impact of the lead time improvement in the downstream, like retailers, affects more to the system than the one in the upstream, like factories, in a cost-effective way. Moreover, by using information lead time for each center, we analyze how much the extent of the improvement affects the whole system, especially for the total cost and the order level.

• 한국컴퓨터정보학회논문지
• /
• 제10권4호
• /
• pp.113-121
• /
• 2005

실시간 영상에서 사람의 얼굴을 검출하는 것은 얼굴 인식 분야에 있어서 주요한 관심 분야 중의 하나이다. 본 본문에서는 실시간 입력되는 영상에서 피부색과 Haar-like feature를 이용한 얼굴 검출 알고리즘을 제안하였다. 제안된 알고리즘은 YCbCr 색 공간에서의 차 연산 기법을 이용하여 이동 물체의 움직임 영역을 ROI(region of interest)로 선정하고 Haar-like feature를 이용하여 얼굴 후보영역을 선정한 다음 피부색 정보를 이용하여 얼굴을 검출하였다. 특히, 가변적으로 선정되는 ROI 영역에 대하여 피부색 정보와 특징 정보를 이용함으로서 실시간 영상에 대하여 처리 속도의 향상과 비슷한 특징 또는 색상을 가진 영상이 얼굴로 검출되는 오류를 방지하였다. 실험 결과는 기존의 연구에 비해 30%의 처리 속도 향상과 96.8%의 검출 성공률을 보였다.