XML을 이용한 웹 상의 멀티미디어 내용 기반 검색 시스템

林海喆,金宣炅,金連姬 弘益大學校 科學技術硏究所 2001 科學技術硏究論文集 Vol.12 No.-

Recently the increase of multimedia data on the web and the user's various requirements lead many multimedia retrieval systems to develop usefully. This paper shows new multimedia content-based retrieval system that manages database for images independently and can retrieve most similar image like query image chosen by user. The proposed system represents text information using XML and exchanges information among modules. So, this system supports retrieval through structure of XML document and integrates easily in heterogeneous environments.

디렉토리 정보 베이스를 위한 관계형 데이타베이스 시스템의 구축

林海喆 弘益大學校 科學技術硏究所 1994 科學技術硏究論文集 Vol.5 No.-

주기억장치 기반 XML 문서 전용 저장 기법 연구

林海喆,金蓮姬,成元鏞 弘益大學校 科學技術硏究所 2002 科學技術硏究論文集 Vol.13 No.-

XML has recently considered as a new standard for data presentation and exchange on the web, many researches are on going to develop applications to store and retrieve XML documents efficiently. Most of these researches are based on the traditional database systems such as relational/object-oriented systems. But these repository systems have many problems of representing the semantics and structures completely, extra efforts to reconstruct XML documents and complicated updates and query processing. So the necessity of native XML repository systems are on the rise to overcome these weak points. The previous native XML repository systems are based on disk, these are inefficient to support real-time and mobile applications efficiently. Hence, in this paper, we design a main memory based XML repository system to support real-time applications efficiently. To do this, we design a logical data model for especially document-centric XML. The proposed data model is designed to represent the semantics and structural features of XML document itself fully. And we implement the proposed repository structure.

확정 정보와 불확정 정보를 위한 데이타 모델

林海喆 弘益大學校 1988 弘大論叢 Vol.20 No.2

We endeavour to classify various kinds of indefinite information occurred in real world, to expand the concepts of logical operators representing the semantics of indefiniteness. Also we define tables to represent indefinite information including definite information based on traditional relation, and to define relational algebraic operators applicable to these tables expanding the concepts of the traditional relational algebraic operators. Consequently, we show that when queries are transformed into relational algebraic expressions, and applied to these tables, many kinds of answers can be generated to the users maintaining the semantics of indefiniteness.

동적 환경을 위한 XML 문서에서의 구조 정보 인덱싱 및 구조적 질의 처리 기법 설계

林海喆,金蓮姬,金宣炅 弘益大學校 科學技術硏究所 2003 科學技術硏究論文集 Vol.14 No.-

In information retrieval(IR) community, recently, researches are in progress to apply its traditional indexing and retrieval schemes to the retrieval of XML documents. When, however, we use index structures in a dynamic environment where documents need frequent updates, there are much overheads for maintaining index. This thesis proposes a fast and efficient indexing and retrieval scheme for structured XML documents in a dynamic environment. Especially, this thesis focuses on supporting both structure and content updates and on supporting efficient retrievals of content and structure. when using the proposed index structure, it is possible to support fast and efficient updates on content and logical structure. For content retrieval, bottom-up-based approach is proposed. And in order to avoid tree traversal manner in structural retrievals, we propose a new approach utilizing the logical structure information namely, element type, included in the user query. To do this, experiments and analytical comparisons for the proposed indexing and the previous schemes were executed.

확장형 베이지안 정리를 이용한 추론 기법

임해철,이석호 한국정보과학회 1986 한국정보과학회 학술발표논문집 Vol.13 No.1

지금까지 많은 논문에서 추론 방법에 대하여 발표되었고 그 하나의 방법으로 베이지안 정리를 적용한 추론 기법이 있다. 그러나 베이지안 정리를 실제로 적용하기에는 가정이나 증거의 구조가 너무 단순하다는 문제가 있다. 본 논문은 그동안 발전되어온 베이지안 정리를 이용한 추론 방법들을 열거하고 문제점을 지적하였으며 이를 해결하기 위하여 각 사건들을 클래스로 나누고 각 사건 상호간에 미치는 영향을 수치로 주어 베이지안 정리를 적용하는 확장형 베이지안 추론 기법을 제안하였다.

멀티미디어 데이터베이스에서의 질의처리를 위한 필터링 방법 연구

임해철,김병곤,한정운 弘益大學校 科學技術硏究所 2000 科學技術硏究論文集 Vol.11 No.-

With the prevalence of multimedia data such as images content-based retrieval of data is becoming increasingly important. To handle multimedia data, multi-dimensional indexing methods such as the R-tree, R*-tree, TV-tree, MVP-tree and Pyramid-technique have been proposed. Numerous research results on how to effectively manipulate these structures have been presented during the last decade. Query processing strategies, which is important for reducing the processing time, is one such area of research. Much of the research on query processing of multimedia data have focused on two aspects, namely, range query and k-nearest-neighbor query (k-NNQ). This paper proposes algorithms for range query processing for R-tree based structures. The novel aspect of these algorithms is that they make use of the notion of VP filtering, a concept borrowed from the MVP-tree. By so doing, we attain considerable performance benefits while paying insignificant overhead during the construction of the index structure. Also, we retain the dynamic feature of R-tree based structures. In the remainder of this section, we briefly review the basics of relevant tree structures. Based on these discussions, we present the motivation behind the introduction of the notion of VP filtering. Despite success of proxy cache in the internet, proxy cache is not efficient for multimedia stream such as video and audio. In this paper, we propose a new proxy cache policy and a efficient cache replacement algorithm for multimedia stream. We expect that a new proxy cache policy is efficient for multimedia stream.

요구언어를 이용한 인사관리 검색 시스템에 관한 연구

林海喆 弘益大學校 1982 弘大論叢 Vol.14 No.2

In a Personnel Information Retrieval System, generally, when anyone wants personnel information from a computer, he must write an application program each time. This is very inefficient and cumbersome. This thesis designs the query language about personnel information which commands required conditions and reporting format and sorting form. Thus, anyone, who wants to retrieve personnel Information from a Data Base, only builds the guery commands.

불확정 정보를 위한 객체 지향 데이타 모델의 확장

林海喆 弘益大學校 1990 弘大論叢 Vol.22 No.2

This paper presents an approach to integrate the concept of indefinite information with certainty factors in an object-oriented framework. The concept of traditional indefinite information is extended to include certainty factors so that indefiniteness can be measurred. Indefinition in formation with certainty factors is classified into two categories: certainty factors of association and certainty factors of objects. Each of these categories is applied to each type of objects: scalar value, rules, operations, etc. Instances with indefinite information with certainty factors including definite object is shown to be stored in G-relation.

불확정 데이타베이스를 위한 관계 대수적 접근

林海喆 弘益大學校 1987 弘大論叢 Vol.19 No.2

A table to represent various kinds of indefinite information is proven to be well-defined. Relational algebraic operations are newly defined which should be applicable to this table, and algorithm to transform queries into relational algebraic expressions is devised. Consequently many kinds of answer types can be generated to the users.