• 한국멀티미디어학회논문지
• /
• 제24권6호
• /
• pp.796-804
• /
• 2021

In computer vision, visual tracking method addresses the problem of localizing an specific object in video sequence according to the bounding box. In this paper, we propose a tracking method by introducing the feature correlation comparison into the siamese network to increase its matching identification. We propose a way to compute location of object to improve matching performance by a correlation operation, which locates parts for solving the searching problem. The higher layer in the network can extract a lot of object information. The lower layer has many location information. To reduce error rate of the object center point, we built a siamese network that extracts the distribution and location information of target objects. As a result of the experiment, the average center error rate was less than 25%.

• 한국건축시공학회:학술대회논문집
• /
• 한국건축시공학회 2010년도 춘계 학술논문 발표대회 1부
• /
• pp.231-234
• /
• 2010

Recently, the construction industry has been changed more highly and complicacy. According to this, several informational systems of maintenance are introduced suitable the work and the purpose of each maintenance. However there are some problems in the manage the information of maintenance management by without the standization and systematic management of variable information from building life-cycle. This paper purposed the process of creation or management of information which is used for present maintenance through IFC-based facilities object library development is suggested for work efficiency of maintenance management. In addition, it can minimize the reproduce the same information unnecessarily. Fainally, it makes information compatibility through the building life-cycle highly by define the IFC-based object information and standardize the format of information.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2007년도 하계종합학술대회 논문집
• /
• pp.313-314
• /
• 2007

In this paper, we suggest robust object extraction algorithm taking advantage of efficient Belief Propagation method. It does not get a disparity information because of uniform region and occlusion region etc. on initial depth map that use forward direction disparity information although is object area. Therefore, We run parallel backward disparity information and brightness information for certain object extraction.

• 한국산학기술학회논문지
• /
• 제9권1호
• /
• pp.100-107
• /
• 2008

본 논문에서는 다중객체 환경에서 특정한 객체를 추출과 추적을 하기 위하여 입력되는 영상의 처리과정을 간소화하여 실시간으로 수행할 수 있는 방법을 제시하였다. 실내환경의 많은 객체들 중 특정영역에 입력되는 객체를 추출하고 추출된 객체를 지속적으로 추적하기 위하여 상단의 카메라와 전면의 카메라의 상호정보교환을 통하여 특정 객체를 지속적으로 추적할 수 있음을 실험을 통하여 검증하였다. 상단의 카메라에서 입력받은 영상에서 x 축의 정보를 전면의 카메라에 전송하여 전면의 카메라는 객체의 x축 정보를 계산하지 않고 빠르게 객체영역을 획득할 수 있으며, 전면카메라에서는 객체들 간의 겹침현상을 해결할 수 없지만 상단의 카메라와 정보를 공유하므로 객체들 간의 겹침현상을 해결할 수 있었다. 실험결과 기존의 객체추적 시스템보다 객체를 추적하기위한 연산량이 줄어들었으며, 객체추출 및 추적의 정확성이 향상되었다.

• 방송공학회논문지
• /
• 제7권3호
• /
• pp.248-261
• /
• 2002

본 논문에서는 스테레오 비젼 시스템의 3차원 정보에 의해 가변되는 윈도우 마스크와 광 BPEJTC(binary phase extraction joint transform correlator)를 이용하여 스테레오 카메라를 제어하는 새로운 3차원 물체추적 시스템을 제안하였다. 즉, 스테레오 비젼 시스템의 구성 요소에 의해 3차원 정보인 추적 물체까지의 거리 정보를 쉽게 구할 수 있고, 이 거리 정보로 윈도우 마스크를 가변 시켜 추적물체 영역을 추출할 수 있다. 이 추적물체 영역은 다음 기준영상으로 갱신하여 사용된다. 그리고 이 기준영상과 스테레오 입력 영상간에 광 BPEJTC를 실행하여 추적 물체의 위치 값을 구하고. 이 값으로 스테레오 카메라의 주시각과 팬/틸트를 제어하여 3차원 물체추적이 이루어진다. 실험 결과 제안한 알고리즘은 스테레오 입력 영상에서 배경잡음과 관계없이 추적 물체영역을 추출하여 3차원 물체추적이 가능하고, 이의 구현으로 3차원 원격작업 시스템이나 적응적인 3차원 물체 추적기 등의 구현 가능성을 제시하였다.

• 한국정보처리학회논문지
• /
• 제6권11호
• /
• pp.3034-3045
• /
• 1999

For efficiently providing multimedia services, distributed computing environments are specified the requirements of various services and distributed object platforms applied an object-oriented technology by TINA Consortium and OMG CORBA. Because multimedia service applications are becoming large and distributing, their servicing managing interfaces among objects are being complicated. In order to solve these defects, it is necessary to suggest a new object grouping model and specify object service/management requirements can be introduced under the object groups. We have been developed the distributed object group platform that can group all individual objects by the relating services and can supply trading functions for interconnecting between distributed objects or object groups. In this paper, we designed and implemented the Virtual Drawing Board for remote equational services on the distributed object group platform we mentioned above. As results, we designed a basic structure and service interfaces, and showed execution procedures of VDB system consisted of distributed objects and objects groups for educational services. For supporting distributed services of VDB system, we used three kinds of tools as follows; IONA orbix 2.2 of CORBA compliance as an object middleware, OrbixTrader 1.0 for interconnection of distributed objects, and the OGTG we developed for interconnection of distributed object groups and checking access rights of objects included in an arbitrary object group.

• 디지털산업정보학회논문지
• /
• 제14권2호
• /
• pp.35-47
• /
• 2018

The 4th industrial revolution, digital image technology has developed beyond the limit of multimedia industry to advanced IT fusion and composite industry. Particularly, application technology related to HCI element algorithm in 3D image object recognition field is actively developed. 3D image object recognition technology evolved into intelligent image sensing and recognition technology through 3D modeling. In particular, image recognition technology has been actively studied in image processing using object recognition recognition processing, face recognition, object recognition, and 3D object recognition. In this paper, we propose a research method of human factor 3D image recognition technology applying human factor algorithm for 3D object recognition. 1. Methods of 3D object recognition using 3D modeling, image system analysis, design and human cognitive technology analysis 2. We propose a 3D object recognition parameter estimation method using FACS algorithm and optimal object recognition measurement method. In this paper, we propose a method to effectively evaluate psychological research techniques using 3D image objects. We studied the 3D 3D recognition and applied the result to the object recognition element to extract and study the characteristic points of the recognition technology.

• 한국전산구조공학회:학술대회논문집
• /
• 한국전산구조공학회 2008년도 정기 학술대회
• /
• pp.518-523
• /
• 2008

In this study, we analyzed applicability of object database, designed the concept model based on object-oriented idea for measurement data management, and applied the design model to object database. The concept model composes three sub models Infrastructure managing information model, Infrastructure measurement data model, and Measurement unit model. The process to expand measurement data of new type was executed easily without changing database schema in object database. The process to expand measurement data of new type was executed easily without changing database schema in object database. Therefore, applicability of new technology to infrastructures for building a safety management network of road bridges could be increased with object database system.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2001년도 하계종합학술대회 논문집(4)
• /
• pp.183-186
• /
• 2001

This paper for the method that automatically extracts moving object of the video image is presented. In order to extract moving object, it is that velocity vectors correspond to each frame of the video image. Using the estimated velocity vector, the position of the object are determined. the value of the coordination of the object is initialized to the seed, and in the image plane, the moving object is automatically segmented by the region growing method and tracked by the range of intensity and information about Position. As the result of an application in sequential images, it is available to extract a moving object.

• 한국통신학회논문지
• /
• 제26권10B호
• /
• pp.1419-1427
• /
• 2001

본 논문은 연속영상에서 non-rigid object를 자동으로 분할하고 알고리즘을 제안하였다. Non-rigid object는 형태의 변화가 일정하기 않은 object로서 기존의 분할 알고리즘과는 다른 새로운 분할 알고리즘을 필요로 한다. 본 논문에서는 특히 구름이나 연기와 같이 형태의 변화가 큰 non-rigid object를 자동으로 분할하는 알고리즘을 제안하였다. 제안된 알고리즘은 공간분할, 시간분할, 그리고 공간분할과 시간분할의 결합의 세 가지 단계로 구성되어 있다. 공간분할은 영상에서 픽셀의 intensity를 마코프 랜덤 필드로 가정하고 에너지 최소화를 통해 영상을 분할한다. 시간분할은 속도벡터를 기반으로 하여 움직임이 있는 영역만을 분할한다. 마지막으로 공간분할과 시간분할을 결합하여 non-rigid object의 최종적인 분할을 수행한다. 실험결과, 제안된 알고리즘은 연속영상에서 non-rigid object를 자동으로 분할함을 확인할 수 있었다.