• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제10권11호
• /
• pp.5530-5546
• /
• 2016

Automatic analysis and understanding of human activities is a challenging task in computer vision, especially for the surveillance scenarios which typically contains crowds, complex motions and occlusions. To address these issues, a Bag-of-words representation of videos is developed by leveraging information including crowd positions, motion directions and velocities. We infer the crowd activity in a motion field using Category Constrained Correlated Topic Model (CC-CTM) with latent topics. We represent each video by a mixture of learned motion patterns, and predict the associated activity by training a SVM classifier. The experiment dataset we constructed are from Crowd_PETS09 bench dataset and UCF_Crowds dataset, including 2000 documents. Experimental results demonstrate that accuracy reaches 90%, and the proposed approach outperforms the state-of-the-arts by a large margin.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제8권6호
• /
• pp.1946-1963
• /
• 2014

Most background subtraction methods focus on dynamic and complex scenes without considering robustness against noise. This paper proposes a background subtraction algorithm based on dictionary learning and sparse coding for handling low light conditions. The proposed method formulates background modeling as the linear and sparse combination of atoms in the dictionary. The background subtraction is considered as the difference between sparse representations of the current frame and the background model. Assuming that the projection of the noise over the dictionary is irregular and random guarantees the adaptability of the approach in large noisy scenes. Experimental results divided in simulated large noise and realistic low light conditions show the promising robustness of the proposed approach compared with other competing methods.

• 한국통신학회논문지
• /
• 제30권11A호
• /
• pp.1098-1103
• /
• 2005

에지나 윤곽 정보는 인간의 시각 시스템에 의하여 가장 잘 인식되며 객체의 인식과 지각에 사용되는 중요한 정보이다. 그러므로 비디오내의 객체간의 상호작용, 객체기반 코딩과 표현과 같은 응용을 위하여, 비디오객체의 추출과정에 에지정보를 적용하면 인간의 시각 시스템과 근접한 객체 경계를 얻을 수 있다. 대부분의 객체추출 방식은 연산량이 많고 반복적인 연산을 수행하므로 실시간 처리가 어렵다. 본 논문에서는 비디오객체 분할 과정에 에지 정보를 적용하여 정확한 객체 경계를 추출하는 VLSI 구조를 제안한다. 제안된 하드웨어 구조는 연산방식이 간단하므로 하드웨어로 쉽게 구현될 수 있으며, 제안된 VLSI 하드웨어 구조를 이용하면 객체기반 멀티미디어 응용을 위하여 실시간으로 비디오객체를 분할할 수 있다.

• 한국컴퓨터정보학회논문지
• /
• 제5권2호
• /
• pp.37-42
• /
• 2000

CORBA에서 객체 기술은 분산 및 이기종 기계에 분산되어 있는 소프트웨어 컴포넌트의 통합된 구현을 용이하게 한다. CORBA와 유사한 객체 통합 기술들은 표준화된 컴포넌트 통합 및 상호 동작 모델을 정하고 호환 불가능한 컴포넌트 구현을 캡슐 화하기 위한 객체지향 원리를 발전시켰다 본 논문에서는 Java Bean에 기반을 둔 분산 시스템 환경에서 객체와 객체간에 관련성을 모델링하기 위해 컴포넌트, 연결자 및 컴포넌트 스키마로 구성된 구조화된 모델을 제시하고 구현한다. 특히 Java Bean 환경에서 객체간의 관련성을 모델링하기 위한 연결자의 구성에 중점을 둔다.

• 대한조선학회논문집
• /
• 제29권2호
• /
• pp.30-37
• /
• 1992

선체는 매우 복잡한 구조물이므로 설계 및 생산의 효율적인 수행을 위하여 선체구조의 모델링에 의한 작업이 필수적이다. 선체모델 구축에 있어서 구조면의 모델링은 초기선형정의에서부터, 선각정보처리, 구획배치, 의장설계, 배관설계, 구조해석 등 선체관련분야와 직접 연관된다. 본 연구에서는 구조면의 효과적인 모델링을 위하여 구조면들간의 위상학적인 자료구조를 구성하므로써 선체구조의 기하학적 정보를 설계의 단계에 따라 발전시켜 나가고 효율적으로 변경시킬 수 있도록 하였다. 본 연구에서 수행된 구조면의 모델링은 선체설계에서부터 생산에 이르기까지 일관된 정보처리를 위한 통합선체모델(Unified Hull Model)구축의 기초가 될 것이다. 구축된 모델의 가시화를 위하여 컴퓨터그래픽스를 이용하여 선체모델을 실물감 있게 표현하였다.

• 한국CDE학회논문집
• /
• 제12권3호
• /
• pp.220-232
• /
• 2007

Current geometric modeling and analysis are commonly based on B-Rep modeling and a finite elements method respectively. Furthermore, it is difficult to represent an object whose material property is heterogeneous using the B-Rep method because the B-Rep is basically used for homogeneous models. In addition, meshes are required to analyze a property of a model when the finite elements method is applied. However, the process of generating meshes from B-Rep is cumbersome and sometimes difficult especially when the model is deformed as time goes by because the topology of deforming meshes are changed. To overcome those problems in modeling and analysis including homogeneous and heterogeneous materials, we suggest a unified modeling and analysis method based on implicit representation of the model using R-function which is suggested by Rvachev. For implicit modeling of an object a distance field is approximated and blended for a complex object. Using the implicit function mesh-free analysis is possible where meshes are not necessary. Generally mesh-free analysis requires heavy computational cost compared to a finite elements method. To improve the computing time of function evaluation, we utilize GPU programming. Finally, we give an example of a simple pipe design problem and show modeling and analysis process using our unified modeling and analysis method.

• Journal of Computational Design and Engineering
• /
• 제2권3호
• /
• pp.183-194
• /
• 2015

This paper proposes a novel method for visualizing the thickness and clearance of 3D objects in a polyhedral representation. The proposed method uses the distance field of the objects in the visualization. A parallel algorithm is developed for constructing the distance field of polyhedral objects using the GPU. The distance between a voxel and the surface polygons of the model is computed many times in the distance field construction. Similar sets of polygons are usually selected as close polygons for close voxels. By using this spatial coherence, a parallel algorithm is designed to compute the distances between a cluster of close voxels and the polygons selected by the culling operation so that the fast shared memory mechanism of the GPU can be fully utilized. The thickness/clearance of the objects is visualized by distributing points on the visible surfaces of the objects and painting them with a unique color corresponding to the thickness/clearance values at those points. A modified ray casting method is developed for computing the thickness/clearance using the distance field of the objects. A system based on these algorithms can compute the distance field of complex objects within a few minutes for most cases. After the distance field construction, thickness/clearance visualization at a near interactive rate is achieved.

• 전자공학회논문지S
• /
• 제36S권9호
• /
• pp.67-75
• /
• 1999

본 논문에서는 패턴인식과 영상압축을 목적으로 2-D 영상내에 포함되어 있는 물체들의 복잡한 형상을 형태론적 연산을 이용하여 단순한 원시형상 요소들로 분해하는 방법에 관해 연구하였다. 기존의 형태론적 형상분해 알고리즘에서 가장 큰 문제점은 형상을 표현하고 기술하는데 필요한 원시형상 요소의 수가 너무 많이 생성된다는 것이다. 본 연구에서는 이러한 문제점을 개선하기 위하여 형상의 기하학적인 특징과 가장 유사한 원시형상 요소와 4개의 스캔모드를 사용하는 형상분해법을 새롭게 제안하였다. 제안된 알고리즘은 4개의 스캔모드를 사용해서 원판, 정사각형, 마름모 꼴 등으로 구성되는 원시형상 요소를 추출하는 방법이다. 이와 같은 알고리즘은 기술 오차를 줄이면서 원시형상 요소의 수를 줄여 기술효율을 높일 수 있는 방법으로 최소의 중복성을 보장할 수 있으며, 알고리즘이 단순하고 계산시간이 감소한다는 특징이 있다.

• Bulletin of the Korean Chemical Society
• /
• 제30권8호
• /
• pp.1783-1792
• /
• 2009

Although overlapping resonances have been studied extensively in conventional resonance theories, there have not been many studies on them in multichannel quantum defect theories (MQDT). In MQDT, overlapping resonances occur between the channels instead of states, which pose far greater difficulty. Their systematic treatment was obtained for cases involving degenerate closed channels by applying our previous theory, which decouples background scattering from the resonance scattering in the MQDT formulation. The use of mathematical theory on con-diagonalization and con-similarity was essential for handling the non-Hermitian symmetric complex matrix. Overlapping resonances in rare gas spectra of Ar, Kr and Xe were analyzed using this theory and the results were compared with the ones of the previous alternative parameterizations of MQDT which make the open-open part $K^{oo}$ and closed-closed part $K^{cc}$ of reactance submatrices zero. The comparison revealed that separation of background and resonance scatterings achieved in our formulation in a systematic way was not achieved in the representation of $K^{oo}\;=\;0\;and\;K^{cc}$ = 0 when overlapping resonances are present.

• 한국조경학회지
• /
• 제26권4호
• /
• pp.95-104
• /
• 1999

The Landscape Drawing is the main medium for constructing better environment which provides human beings with a place for dwelling well, even though it has some limitations. To provide places for people's daily lives, by what measure do we design\ulcorner, by what measure do we give dimension\ulcorner, and how can we appropriately measure the place with drawing\ulcorner On any given site, of course, much has already happened, i.e, there is history. Any adequate interpretation of such a site should poetically take into account the present effects of its absent past. The findings to reveal the meaning and methodology of landscape drawings in the environmental design with a case study are as follows: (1) The value of landscape drawings is the poetic, artistic interpretation which we sense and bodily experience through the simultaneous exposure of the site's temporal, spatial, and tactile conditions. (2) The landscape drawing should represent the invisible (imaginary) world as well as the visible (physical) world. (3) There can be classified as two methods in representing the rich experience of a place: expression and notation. Expression is the abstract representation and do not have a direct relationship to construction due to its ambiguous and connotative structure. But notation is constructional due to it's purely objective symbol system which is obviously denotative structure. (4) Collage is a medium of specific act that can represent the complex layered phenomena and the rich experience of given site. (5) The landscape drawings represent not only topographical space which can be measured objectively, but also topic which can be cognized culturally.