University of Cambridge > Talks.cam > Microsoft Research Machine Learning and Perception Seminars > Constraint Programming Techniques for Virtual Camera Control

Constraint Programming Techniques for Virtual Camera Control

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Oliver Williams.

In computer graphics, progress in modeling, animation and rendering means that rich, high fidelity interactive virtual worlds are now commonplace. But as photographers and cinematographers know, achievement of the intended informational and aesthetic goals is highly dependent on the position and motion of the camera in relation to the elements of the scene. In this talk, we explore a number of models and techniques that we have proposed in order to assist the control of a camera in a virtual 3D environment. In particular, we emphasize the use of interval constraint techniques for computing camera configurations and paths both in declarative and interactive frameworks. We will first concentrate on the problem of declarative camera planning. Camera planning in computer graphics can be considered as a specific complex instance of the classical PSPACE -hard path-planning problem. We will show how interval constraint techniques can be used to provide reliable and guaranteed solutions in a declarative framework and then propose the integration of semantics in the solving process that lays the groundwork for further high-level interactive approaches. We then will concentrate on the problem of interactive virtual cinematography. Virtual cinematography consists in positioning, in a 3D scene, the three major components of cinematography (characters, camera, lights) in order to achieve given intended objectives (informative, cognitive, aesthetic). Here we demonstrate how interval constraint techniques can be used to directly manipulate composition by relying on the Through-the-lens interaction metaphor extended to object and light manipulation. Complex issues pertaining to properties such as occlusion will be explored and discussed. We will finally see that enforcing higher level properties requires a better understanding of the users perception and cognitive models, which makes virtual camera control a key area for interdisciplinary research by both computer scientists and cognitive psychologists.

This talk is part of the Microsoft Research Machine Learning and Perception Seminars series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

 

© 2006-2020 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity