애니메이션 시나리오와 캐릭터 파라미터 기반의 제스처 제안 시스템 = A Gesture Suggestion System Based on the Animation Scenario and Character Parameters

다국어 초록 (Multilingual Abstract)

This dissertation explores a gesture suggestion system which suggests appropriate gestures for the user input such as the meaning of gesture, the situation in animation scenario and character parameters. A gesture is a kind of the body movements for communication. It functions as a very important factor in order to express the scenario and the characteristics of characters in animation. However, a gesture is considered less importantly than the basic motion and facial expression in animation research field. Particularly, the study on gesture as a communication method is rare. Moreover, the engineering approach to a gesture take no notice of the aspect of gesture types including the overall scenario and objective data.
Therefore, this study analyzes character gestures in animation on the basis of Ekman & Friesen’s functional methodology. The analysis reaches the result that character gestures are classified into 7types; 1)emblems, 2)word-descriptive illustrators, 3)speech intension illustrators, 4)mental state adaptors, 5)physical state adaptors, 6)regulators, and 7)affect display. The meaning of gestures can be analyzed according to the context in the scenario. In particular, the meaning of word-descriptive illustrators is determined by the dialogue, and the meaning of speech intension illustrators is analyzed on the basis of the speech act.
As the result of the above analysis, total 909 gesture video clips of 25 characters are selected from 10 animations. And the meanings of gestures are determined. In this process, the objective measures of the gesture meaning are obtained from the survey. These gesture video clips are stored as the character gesture database that includes the metadata structured with 3 attributes of the gesture information, 8 attributes of the character, 4 attributes of the source animation, and 2 attributes of the camera work.
The gesture suggestion system is designed for the practical use of this gesture database. This system provides the searching methods for gesture video data according to the gesture type and meaning. In addition, it can suggest particular gesture video data by filtering character parameters and camera work type, and also show various detailed data informations based on metadata.
Effectiveness of this system is verified by the user survey for quantitative and qualitative analysis. Compared to existing searching method, this system has highly efficient performance in searching gesture data.
Accordingly, this gesture suggestion system will much contribute to the reference which helps animators to produce character gestures effectively and the fundamental database for auto gesture synthesis system in the future study.

목차 (Table of Contents)

• 1. 서론 = 1
• 1.1 연구의 목적 = 1
• 1.2 연구의 배경 및 필요성 = 4
• 1.3 연구의 범위 및 구성 = 7
• 2. 이론적 배경 = 10
• 2.1 제스처 개념과 연구 동향 = 11
• 2.2 제스처 애니메이션 생성 연구 = 16
• 2.2.1 대화 기반 모델링 = 16
• 2.2.2 내적 사고 기반 모델링 = 19
• 2.2.3 캐릭터 파라미터 기반 모델링 = 23
• 2.3 제스처 데이터베이스 관련 연구 = 24
• 2.4 기능적(functional) 접근 = 26
• 2.5 화용론적(pragmatic) 접근 = 31
• 3. 애니메이션 캐릭터 제스처 분석 = 34
• 3.1 캐릭터 제스처 유형 및 의미 분석 = 34
• 3.1.1 애니메이션의 의사소통 체계 = 34
• 3.1.2 기능적 유형에 기반한 캐릭터 제스처 분석 = 37
• 3.1.3 제스처 유형 및 의미 분석 결과 = 51
• 3.2 캐릭터 제스처 동작 표현 분석 = 54
• 3.3 시나리오에 따른 제스처 표현 분석 = 57
• 3.4 요약 = 60
• 4. 애니메이션 캐릭터 제스처 DB 구축 = 61
• 4.1 애니메이션 캐릭터 선정 및 데이터 추출 = 61
• 4.2 의미 목록화 작업 및 심리학적 의미 검증 = 63
• 4.2.1 의미 목록화 작업 = 63
• 4.2.2 심리학적 의미 검증 = 66
• 4.2.3 의미 검증 결과 = 69
• 4.3 데이터베이스 설계 = 75
• 5. 시스템 구축 및 검증 = 85
• 5.1 데이터 처리 과정 = 85
• 5.2 인터페이스 및 시스템 구축 = 87
• 5.3 시스템 검증 = 89
• 6. 결론 및 향후 과제 = 99
• 참고문헌 = 102
• 요약 = 109
• 부록 = 111