• ETRI Journal
• /
• v.40 no.4
• /
• pp.499-510
• /
• 2018

A vector space based augmented structural kinematic (VSASK) feature descriptor is proposed for human activity recognition. An action descriptor is built by integrating the structural and kinematic properties of the actor using vector space based augmented matrix representation. Using the local or global information separately may not provide sufficient action characteristics. The proposed action descriptor combines both the local (pose) and global (position and velocity) features using augmented matrix schema and thereby increases the robustness of the descriptor. A multiclass support vector machine (SVM) is used to learn each action descriptor for the corresponding activity classification and understanding. The performance of the proposed descriptor is experimentally analyzed using the Weizmann and KTH datasets. The average recognition rate for the Weizmann and KTH datasets is 100% and 99.89%, respectively. The computational time for the proposed descriptor learning is 0.003 seconds, which is an improvement of approximately 1.4% over the existing methods.

• ETRI Journal
• /
• v.34 no.4
• /
• pp.572-582
• /
• 2012

The discrete Gaussian-Hermite moment (DGHM) is a global feature representation method that can be applied to square images. We propose a modified DGHM (MDGHM) method and an MDGHM-based scale-invariant feature transform (MDGHM-SIFT) descriptor. In the MDGHM, we devise a movable mask to represent the local features of a non-square image. The complete set of non-square image features are then represented by the summation of all MDGHMs. We also propose to apply an accumulated MDGHM using multi-order derivatives to obtain distinguishable feature information in the third stage of the SIFT. Finally, we calculate an MDGHM-based magnitude and an MDGHM-based orientation using the accumulated MDGHM. We carry out experiments using the proposed method with six kinds of deformations. The results show that the proposed method can be applied to non-square images without any image truncation and that it significantly outperforms the matching accuracy of other SIFT algorithms.

• 제어로봇시스템학회:학술대회논문집
• /
• 2000.10a
• /
• pp.311-311
• /
• 2000

We propose a DP-based formulation for matching line patterns by defining a robust and stable geometric representation that is based on the conceptual organizations. Usually, the endpoint proximity and collinearity of image lines, as two main conceptual organization groups, are useful cues to match the model shape in the scene. As the endpoint proximity, we detect junctions from image lines. We then search for junction groups by using geometric constraint between the junctions. A junction chain similar to the model chain is searched in the scene, based on a local comparison. A Dynamic Programming-based search algorithm reduces the time complexity for the search of the model chain in the scene. Our system can find a reasonable matching, although there exist severely distorted objects in the scene. We demonstrate the feasibility of the DP-based matching method using both synthetic and real images.

• Annual Conference on Human and Language Technology
• /
• 1999.10e
• /
• pp.231-236
• /
• 1999

시간표현은 문장에서 다른 명사와 결합하여 복합어를 이루는 경우가 있고, 용언과 결합하여 시간 부사의 역할을 하는 경우가 있는데, 이는 구문 분석에 있어서 중의적 해석이 두드러지며, 그 결과 구문 분석의 오류를 빈번히 야기하기도 한다. 본 논문에서는 이러한 시간 관계의 표현을 대량의 말뭉치로부터 획득하고 이들을 부분문법(local grammar)으로 표현한 후, 이것을 FST(Finite State Transducer)를 이용하여 부분 구문분석을 하고자 한다. 이를 위해 5천만 어절의 말뭉치에서 259개의 시간 단어를 추출하였고, 시간 단어들의 의미적 또는 기능적 사용에 의해서 26개의 어휘 범주로 분류하고 각 범주들의 결합관계를 일반화하였다. 실험을 통하여 인식을 위한 시간표현의 결합관계는 최고 97.2%의 정확률을 보였고, 품사태깅에 있어서는 평균 96.8%의 정확률을 보였다. 이는 시간 표현의 결합관계가 부분 구문분석에 있어서 유용한 정보임을 보여준다.

• Proceedings of the KOSOMBE Conference
• /
• v.1996 no.11
• /
• pp.293-296
• /
• 1996

In this paper, we studied two image characteristics, similarity and smoothness, which give rise to local and global redundancy in image representation. The similarty means that any patterns in the image repeat itself anywhere in the rest of image. The smoothness means that the gray level values within a given block vary gradually rather than abruptly. In this sense, we propose a lossless medical image compression scheme which exploits both types of redundancy. This method segments the image into variable size blocks and encodes them depending on characteristics of the block. The proposed compression schemes works better than other compression schemes such as the huffman, the arithmetic, the Lempel-Ziv and the lossless scheme of JPEG.

• Journal of Korean Institute of Industrial Engineers
• /
• v.18 no.2
• /
• pp.11-22
• /
• 1992

Up to now, most researches on production automation have concentrated on local automation, e. g. CAD, CAM, robotics, etc. However, to achieve total automation it is required to link each local modules such as CAD, CAM into a unified and integrated system. One such missing link is between CAD and computer vision system. This thesis is an attempt to link the gap between CAD and computer vision system. In this paper, we propose algorithms that carry out edge detection, thinning and pruning from the image data of manufactured parts, which are obtained from video camera and then transmitted to computer. We also propose a feature extraction and surface determination algorithm which extract informations from the image data. The informations are compatible to IGES CAD data. In addition, we suggest a methodology to reduce search efforts for CAD data bases. The methodology is based on graph submatching algorithm in GEFG(Generalized Edge Face Graph) representation for each part.

• Journal of the Korean housing association
• /
• v.24 no.6
• /
• pp.199-208
• /
• 2013

Developing a modern Han-style design and providing support for the commercialization development model in recent years has been propelled by the Han-style Support Strategies of the central government in conjunction with Han-style revitalization related projects that reflect the efforts of local governments. Han-style revitalization, the rekindling and revaluing of human behavior and interest in local governments following the social and cultural changes of the past decades, has emerged as an increasingly traditional area of concern in Han-style design. The purpose of the study was to provide a method which clarifies the design rules of the Han-style bathroom based on an evaluation of sensibilities and a rough set theory, and to give the components meaning and to systematize the method. Essentially, the Han-style bathroom design evaluation is a complex multi-criteria decision making process that seeks to improve the effectiveness and objectively of the Han-style bathroom design. Han-style bathroom design can be displayed in a graphical representation in response to input from the evaluation concerning sensibilities. Because the graphical representation is composed of 3D data, it is possible to display the Han-style bathroom design form in any desired perspective and also to perform shading and other operations. With the proposed method, it is possible to obtain a combination of several contributory components which can be referred to as Reducts, Covering Index and Column Score. Han-Style/Non Han-Style Bathroom Designs were identified by the combination of several components.

• The Journal of the Korea Contents Association
• /
• v.5 no.3
• /
• pp.73-81
• /
• 2005

CTOC(Classes To Optimized Classes) is a Java bytecode framework for optimization and analysis. Although Java bytecode has a significant amount of type information embedded in it, there are no explicit types for local variables. However, knowing types for local variables is very useful for both program optimization and analysis. This paper is a first part of CTOC framework. In this paper, we present methods for inferring static types for local variables in a 3-address, stackless, representation of Java bytecode. We use SSA Form(Single Static Assignment Form) for spliting local variables. Splited local variables will use to prepare for static type inference.

• Architectural research
• /
• v.13 no.4
• /
• pp.31-42
• /
• 2011

This study notices that all religions in Central Asia from the fourth century through seventh century C.E. provided considerable hands in keeping a uniform unity through a process of assimilation, although art and architecture were greatly stimulated by the creative genius of the many people. The study thus intends to argue that the common ideas of rituals and primitive forms of religious shrines lead the square-based layout of Buddhist shrines the unity and universality in the architectural products of particular regions or epochs: i.e. the "square-based plan" in Buddhist temples of Central Asia was a significant prototype in the synthesis with pre-Buddhist architectural models and Buddhist universal ideas. Thus, this thesis notes that they did not lose the universal principles of the Buddhist shrine plans due to ritual functions, and even there have been never differences from pre-Buddhist building models remarked by the periods and the venues in which they were produced, although there had been continuous evolutions and adaptive transformations in the local tastes of religious architecture. Accordingly, this study discusses how such plans in Buddhist architecture had been consistently produced within that regional style also representative of the local idioms of architecture, and how they were adopted in the sites, founded on the composition of ritual functions. The foreign architectural cultures were selectively chosen getting along with local building types of each site according to each taste for architecture as a result.

• Journal of the Korea Computer Graphics Society
• /
• v.24 no.3
• /
• pp.43-48
• /
• 2018

In this paper, we present a method for removing shadows from scanned documents. Compared to the existing methods such as one based on image pyramid representation or adaptive thresholding, our method produces more robust and higher quality results. The basic idea of the approach is to use the local image statistics and to separate interesting regions from the image such as the regions around letters and figures. For the separated regions, we adaptively adjust the local brightness and contrast, and apply the sigmoid function to the intensity values as well to enhance the clarity of the image. For separated the other empty regions, we apply the gradient-base image hole filling method to fill the region with smooth color change.