Knowledge-Based Representation of Cinematic Expression and its Application to Animation
Abstract
Camera control in virtual environments received growing attention recently, and some of the research turned to cinematography as a source for inspiration. Cinematic knowledge is a highly challenging domain for formalization. We have used a systematic knowledge-based approach for knowledge acquisition and formalized expert knowledge in the form of fine-grained rules. Given a screenplay, the rules interact to suggest a solution by symbolic constraint propagation and truth maintenance. We have implemented a prototype system that accepts screenplays in a formal language and generates editing decisions, based on a cinematic knowledge base. If the system is provided with 3D meshes and animations corresponding to the actions and objects mentioned in the screenplay, the system generates a 3D animation clip for the screenplay, based on its editing decisions. The system can be equipped with cinematic principles from various genres, and shows how complex cinematic phenomena can emerge from the interaction of simple rules.
BibTeX
@inproceedings {10.2312:egs.20021023,
booktitle = {Eurographics 2002 - Short Presentations},
editor = {},
title = {{Knowledge-Based Representation of Cinematic Expression and its Application to Animation}},
author = {Friedman, Doron and Feldman, Yishai},
year = {2002},
publisher = {Eurographics Association},
ISSN = {1017-4656},
DOI = {10.2312/egs.20021023}
}
booktitle = {Eurographics 2002 - Short Presentations},
editor = {},
title = {{Knowledge-Based Representation of Cinematic Expression and its Application to Animation}},
author = {Friedman, Doron and Feldman, Yishai},
year = {2002},
publisher = {Eurographics Association},
ISSN = {1017-4656},
DOI = {10.2312/egs.20021023}
}