CGI-FOR-FILM · Controllable global illumination for film production

The ultimate objective of the project is to start a new interdisciplinary computer graphics research group in the return host under the direction of the fellow. This leading position will allow him to fully exploit the skills acquired during the outgoing stage at the renowned Cornell University Program of Computer Graphics, and will be the foundation of his long-term independent research career. The central source of the new skills will be the execution of the proposed research, where our goal is to allow the artists in film production to create more compelling realistic computer-generated imagery through establishing global illumination as a common tool for image rendering. To reach this goal, we propose several novel scientific aspects, the most significant being the study of visual importance of various distance components of light transport on the perceived quality of synthetic images. Results of this study will drive the design choices in our development of new algorithms and data structures, such as: a) new illumination caching strategies for efficient global illumination computation in scenes with complex geometry and materials, b) memory-efficient representation of the light transport operator for real-time relighting in cinematic lighting design. In addition, a novel interface will be developed to provide artistic control over global illumination computation. The proposed research, encompassing several disciplines, such as psychophysics and cognitive sciences, statistical data modeling, and human-computer interaction, is designed specifically to deepen and widen fellow’s expertise and add to his multidisciplinary skills.