• The Journal of the Korea Contents Association
• /
• v.14 no.11
• /
• pp.541-548
• /
• 2014

In this paper, we introduce a second version of modified binary particle swarm optimization using a concept of genotype-phenotype in genetic algorithms. Particle swarm optimization uses an information of difference between a position of the best solution and one's own position in the process of searching optimum. To obtain this difference of positions, the first version of modified binary particle swarm optimization uses a phenotype but the proposed second version uses a genotype. We can represent the solution space in large search space by using a genotype which provides continuous whole space as search space compared to a phenotype which provides only binary information. Experimental results in 10 De Jong benchmark function show that the second version outperforms the first version in six functions which has a broad search space and many local optima.

• Cartoon and Animation Studies
• /
• s.9
• /
• pp.143-160
• /
• 2005

This study was performed on purpose to consider the charade function of props in directing animation. The prop is gradually expanded its role and scope in modern animation. Its scope covers dramatic factors, creating and changing characters, the means of psychological description or emotional expression, mise-en-scene, montage, creating products etc. Like this, the prop has an important function emotionally or indispensably in human life. Thus, the prop should be a subject that can speak to audiences or viewers, not worthless and silent subject. Hence, the prop should be recognized as a charade having productional function. Nevertheless, the prop has been unnoticed up to now. The function of props should be valued hereafter. The prop itself is meaningless, but it always speaks to audiences on the screen.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2011.07a
• /
• pp.258-260
• /
• 2011

이 논문에서는 쿼리-바이-싱잉/허밍 (Query-by-singing/humming, QbSH) 시스템 또는 커버 노래 인식 (cover song identification) 시스템에서 사용 가능한 멜로디 유사도를 제안한다. QbSH 또는 커버 노래 인식은 디지털 음악의 사용이 보편화되면서 음악 검색의 방법으로 많은 연구가 진행되어 오고 있다. 멜로디 유사도는 이런 시스템을 구현하는데 필수적인 요소이며, 두 개의 음악에서 멜로디가 추출되었다고 가정하고, 추출된 멜로디 사이의 유사한 정도를 수치로 표현한다. QbSh 시스템이나 커버 노래 인식 시스템은 멜로디 유사도에 기반하여 입력 노래와 유사한 노래를 데이터베이스에서 검색하는 작업을 수행한다. 이 논문에서 제안하는 멜로디 유사도 방식은 기존의 많이 연구되던 동적 시간 왜곡 (dynamic time warping, DTW) 방법과 크로마 표현 방법 (chroma representation)을 사용하였다. DTW방법은 비대칭적으로 사용하고 미디 노트 영역에서 표현된 멜로디 특징은 0이상 12 미만의 크로마 레벨로 표현하였다. 기존의 방법에서는 정수값을 많이 사용하였으나 이 논문에서는 실수값을 사용한다. DTW 에 사용하는 거리 함수를 기존에 사용하던 차이의 절대값 대신 꺾인 함수 형태를 사용함으로써 성능을 높였다. QbSH 시스템에서의 실험을 통해서 성능을 검증하였다. 본 논문에서는 10-12초 길이의 1000번의 쿼리(Query)에 대해서 28시간 정도의 데이터베이스에서 실험한 결과, 순위 역의 평균 (Mean reciprocal rank, MRR) 값이 0.713을 보였다.

• Korean Journal of Cognitive Science
• /
• v.19 no.3
• /
• pp.257-281
• /
• 2008

The expression 'kes' is one of the most widely used ones in the language whose uses are highly dependent upon the context. These highly-context dependent uses make it hard to determine its grammatical properties. As a way of examining the properties in a rather controlled context, this paper collects a series of speeches made by government officials and examines the grammatical properties of the expression in the corpus. In particular, the paper, based on the 539 instances of 'kes' uses extracted from the corpus, focuses on the 7 types of 'kes' constructions most widely used in the collected speech corpus.

• Proceedings of the Korean Information Science Society Conference
• /
• 2011.06a
• /
• pp.282-285
• /
• 2011

서술논리를 기반으로 하는 OWL 온톨로지는 표준화된 형식적 언어로써 실세계의 도메인 지식을 표현하는데 적합하다. 따라서 논리를 바탕으로 명시적으로 정의된 지식 속에 내재되어 있는 새로운 지식의 추론이 가능하다. 그러나 OWL이 가지는 Open World Assumption(OWA)의 특성은 근거가 불완전하거나 완전한 정보획득이 불가능한 상황에서의 추론을 제한한다. 더불어 OWL이 가지는 또 다른 특성으로 Unique Name Assumption(UNA)의 비지원은 실제적 지식표현을 지원하는 반면, 표현의 불충분으로 인해 결과 도출의 불능을 야기한다. 이러한 특징을 고려하여, 본 논문에서는 지능형 에이전트 구성을 위한 서술논리 기반 지식 표현 방법을 제안한다. 이는 논리적 정당성을 유지하고 올바른 결과를 이끌어 낼 수 있도록 하며, 항상 논리적 결론 도출이 가능한 지식모델을 구성할 수 있도록 돕는다. 이를 통해, 지식모델에 정의된 불완전한 개념에 있어서 OWL이 가지는 특징으로 인하여 발생할 수 있는 문제점에 대한 해결방안을 제시한다. 이에 있어서, 모바일 온톨로지의 예를 통하여 OWA와 UNA에 따른 추론의 제약을 보이며, 이를 해결할 수 있는 방안을 논리적으로 표현함으로써 본 제안의 정당성을 증명한다.

• Cartoon and Animation Studies
• /
• s.34
• /
• pp.1-22
• /
• 2014

The purpose of this research is to study a case of allegory and experimental presentation technique in Film by Zbigniew Rybczynski. It is to analyse this study is to inquire into the fact that experimental animation carries diverse interpretation of comtemporary art and the purpose of this study is to identify certain style through the interpretation of allegory focused on film Tango which has expanded the field of experimental animation with the unique form of the film structure and technique. Therefore, in this research, I would highlight the characteristic of allegory in Film by Zbigniew Rybczynski and research the experimental presentation technique that appears in experimental animation focused on "fragmentation" which are the expressive style of allegory.

• Korean Journal of Cognitive Science
• /
• v.10 no.1
• /
• pp.45-55
• /
• 1999

This paper argues that the basic modes of spatial cognition can be best identified in terms of argument/participant location, and shows that natural language uses‘simple’types of semantic denotations to encode spatial cognition, and further notes that spatial expressions should be interpreted not as locating an event/state as a whole but as locating arguments/participants of the event. The ways of locating events/states are identified in terms of argument orientation(AO), Which indicates semantic patterns of linkiarticipant location. and shows that natural langrage uses ng locatives to specific arguments. Four patterns of argument orientation described here reveal substantial modes of spatial cognition. and the AO patterns are mostly determined by the semantic classes of English verbs combining with locative expressions, i.e., by the event type of the predicate. As for the denotational constraint of locatives, the paper concludes that semantic denotations of locative PPs are restricted to the intersecting functions mapping relations to relations.

• The Journal of the Korea Contents Association
• /
• v.7 no.10
• /
• pp.105-114
• /
• 2007

The purpose of this study is to investigate how to express Surrealism techniques in Chris Landreth who is very famous for 3D short such as Ryan. The case works are focused on 'the end', Chris Landreth's early work. He set up situation of the absurd and describes human being's alienation in animation. A phase of social life is not only revealed with vivid colors and his own surrealism techniques but communication capacity of 3D animation is also shown. Digital Technology, reproducing impossible images is not all about 3D animation. As well as this, it is not far away the way of creating arts. 3D animation will be shown the new possibilities with extension and harmony of all kinds of arts expression.

• Journal of KIISE:Computing Practices and Letters
• /
• v.12 no.6
• /
• pp.379-385
• /
• 2006

To improve reusability and adaptation of variable components, rule-based component development has been used. Rule engines usually need additional script languages for rule expression and have difficulty in expressing complex business rules. In this paper, we propose a compiler-based rule engine for rich rule expression and improving performance. This rule engine uses Java programming language to express conditions and action parts of rules and that it can easily express complex business rules. It creates and executes condition and action objects at run time. In view of Performance, the rule engine is better than a script based rule engine. According to our experiments, our compiler-based nile engine shows 2.5 times better performance that script-based JSR 94 rule engine.

• Journal of KIISE:Databases
• /
• v.31 no.1
• /
• pp.72-81
• /
• 2004

Theses days, a number of data over the Internet are represented using XML because of a virtue of XML. In proportion to the increase of XML data, query processing techniques are required that support quickly and efficiently the diverse queries to search the useful information on XML documents. But, up to now, the researches handling queries for XML data are methodologies focusing on how to process regular path expressions. Therefore, we have developed a new genealogy-based indexing technique to solve various queries such as not only regular path expression but also simple path expression, path expression referencing other elements, and so on. Also, we have applied this technique on object-relational model and evaluated the performance for many documents and various query types. The result shows improved performance in comparison with the other storage techniques.