• 한국컴퓨터정보학회논문지
• /
• 제25권2호
• /
• pp.49-58
• /
• 2020

본 연구는 온라인 다중 객체 추적 환경에서 모든 객체의 상태(예. 위치 및 크기) 및 identifications (IDs)를 추적하는 문제를 다룬다. 프레임들 간 검출 결과들을 연관하여 객체들의 궤도를 점진적으로 완성하는 tracking-by-detection 접근법을 기반으로 온라인 다중 객체 추적 문제를 해결하고자 한다. 정확한 온라인 연관을 수행하기 위해 이산 푸리에 변환과 부분 최소 제곱법(partial least square, PLS) 분석을 기반으로 하는 새로운 온라인 외형 학습 방법을 제안한다. 즉, 먼저 주파수 도메인에서 추적에 용이한 객체 특징량을 추출하기 위해 추적 객체에 대한 이미지를 푸리에 이미지로 변환한다. 나아가 객체간의 주파수 특징을 보다 잘 구별할 수 있도록 PLS기반 부분 공간을 학습한다. 제안된 외형 학습을 최신 신뢰도 기반 연관 기법과 결합하였고, 다중 객체 추적평가 분야에서 국제적으로 공인된 MOT 벤치마크 챌린지 데이터 셋에서 최신 다중 객체 추적 알고리즘과 비교평가를 수행하였다.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2003년도 ICCAS
• /
• pp.1994-1999
• /
• 2003

This paper discusses dynamic characteristics of motion of a pair of multi-degrees of freedom robot fingers executing grasp of a rigid object and controlling its orientation with the aid of rolling contacts. In particular, the discussions are focused on a problem of gain-tuning of sensory feedback signals proposed from the viewpoint of sensorymotor coordination, which consist of a feedforward term, a feedback term for controlling rotational moment of the object, and another term for controlling its rotational angle. It is found through computer simulations of the overall fingersobject dynamics subject to rolling contact constraints that some dynamic characteristics of torque-angular velocity relation may play an important role likely as reported by experimental results in muscle physiology and therefore selection of damping gains in angular velocity feedback depending on the guess of object mass is crucial. Finally, a guidance of gain-tuning in each feedback term is suggested and its validity is discussed by various computer simulations.

• 천문학회보
• /
• 제38권2호
• /
• pp.70.1-70.1
• /
• 2013

We present long-term optical to NIR data of the tidal disruption object, Swift J1644+57. The data were obtained with CQUEAN, UKIRT WFCAM observations. We analyze the morphology of the host galaxy of this object and decompose the bulge component using high resolution HST WFC3 images. We conclude that the host galaxy is bulge dominant. We also estimate the multi-band fluxes of the host galaxy through the light curves based on the long-term observational data. We fit the SED models to the multi-band fluxes of the host galaxy and determine its stellar mass. Finally, we estimate the mass of the central super massive black hole which is thought to be the main role of the tidal disruption event. The estimated stellar mass and black hole mass are $10^{9.1}M_{\odot}$, $10^{6.8}M_{\odot}$ respectively. We compare our results to other results that have studied before.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1996년도 한국자동제어학술회의논문집(국내학술편); 포항공과대학교, 포항; 24-26 Oct. 1996
• /
• pp.539-542
• /
• 1996

The forces exerted on an object by the end-effectors of multi-manipulators are decomposed into the motion-inducing force and the internal force. Motion-inducing force effects the motion of an object and internal force can't effect it. The motion of an object can't track exactly the desired motion because of internal force component, therefore internal force component must be considered. In this paper using the resolved acceleration control method and the fact that internal force lies in the null space of jacobian matrix, we construct independently the position, motion-inducing force and internal force controller. Secondly we construct the robust controller to preserve the robustness with respect to the uncertainty of manipulator parameters.

• 한국컴퓨터정보학회:학술대회논문집
• /
• 한국컴퓨터정보학회 2015년도 제51차 동계학술대회논문집 23권1호
• /
• pp.137-138
• /
• 2015

본 논문에서는 연속된 영상에서 다중해상도 에지정보의 차이를 이용하여 이동하는 물체를 탐지하고 계수하는 시스템을 제안한다. 연속적으로 입력되는 영상에 대하여 이산 웨이블릿 연산을 수행하여 다중해상도 에지를 추출하고, 인접한 프레임 사이의 다중해상도 에지 차이를 이용하여 이동물체를 추출한다. 가중치가 부여된 유클리디언 거리를 이용하여 물체를 추적한 다음, 칼만필터를 이용하여 물체 궤적의 위치 정보를 보정한다. 마지막으로, 관심영역에 대한 물체 궤적의 상대적인 위치를 고려하여 이동물체를 계수한다.

• 전자공학회논문지S
• /
• 제35S권1호
• /
• pp.80-87
• /
• 1998

In this study, a multi-element ultrasound transducer has been developed aiming at basic experiment of three-dimension endovascular ultrasound endscopy for clinical diagnos, and experimental results of two-dimensional object imaging in the water are presented by the ultrasound tranducer and neural network. Each ultrasound echo received by thirty-six angular transducer elements is inputed to the eural network, and then backpropagation is used as a learning algorithm. A three-layer artificial neural network is used for learning and imaging of targetw placed in front of the transducer. The object shape of imaging is restricted to rectangular shapes by considering experimental restraint conditions. As a result, rough visualization can be realized even for objects with unlearned shapes through the training by primitive patterns of a various sized rectangular targets.

• Journal of Computing Science and Engineering
• /
• 제8권2호
• /
• pp.65-77
• /
• 2014

To address the problem of multi-target retrieval (MTR) of remote sensing images, this study proposes a new object-level feature representation model. The model provides an enhanced application image representation that improves the efficiency of MTR. Generating the model in our scheme includes processes, such as object-oriented image segmentation, feature parameter calculation, and symbolic image database construction. The proposed model uses the spatial representation method of the extended nine-direction lower-triangular (9DLT) matrix to combine spatial relationships among objects, and organizes the image features according to MPEG-7 standards. A similarity metric method is proposed that improves the precision of similarity retrieval. Our method provides a trade-off strategy that supports flexible matching on the target features, or the spatial relationship between the query target and the image database. We implement this retrieval framework on a dataset of remote sensing images. Experimental results show that the proposed model achieves competitive and high-retrieval precision.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2006년도 하계종합학술대회
• /
• pp.409-410
• /
• 2006

We propose a surface modeling method using multi-laser lining. We can acquire 3D surface by analyzing projected laser lines on object. The laser lines reflect the surface curvature. In this paper, we show two novel approaches to reconstruct 3D shape of moving object. One is used for robust laser line extraction, and the other for registration between 3D shapes.

• 천문학회보
• /
• 제39권1호
• /
• pp.48.2-48.2
• /
• 2014

We present long-term optical to NIR data of the tidal disruption object, Swift J1644+57. The data were obtained with CQUEAN, UKIRT WFCAM observations. We analyze the morphology of the host galaxy of this object and decompose the bulge component using high resolution HST WFC3 images. We conclude that the host galaxy is bulge dominant. We also estimate the multi-band fluxes of the host galaxy through the light curves based on the long-term observational data. We fit the SED models to the multi-band fluxes of the host galaxy and determine its stellar mass. Finally, we estimate the mass of the central super massive black hole which is responsible for the tidal disruption event. The estimated stellar mass and black hole mass are ${\sim}10^{9.1}M_{\odot}$, ${\sim}10^{6.8}M_{\odot}$, respectively. We compare our results to other previous estimates.

• 제어로봇시스템학회논문지
• /
• 제16권2호
• /
• pp.120-125
• /
• 2010

This paper proposes an intelligent PID/Fuzzy control system for two humanoid robots to transport objects stably. When a robot transports an object while walking, a whole body system of a robot may not be stable due to vibration or external factors from a different departure speed error and a body movement of walking robots. Therefore, it is necessary to measure the horizontal and vertical locations and speeds of object, then calibrate the difference of departure speed between robots with PID/Fuzzy control. The results of simulation with two robots indicated that a proposed controller makes robots to transport an object stably.