Modeling camera control with constrained hypertubes

Abstract

In this paper, we introduce a high-level modeling approach to camera control. The aim is to determine the path of a camera that verifies given declarative properties on the desired image, e.g., lo cationor orientation of objects on the screen at a given time. The path is composed of elementary movements called hyper tubes, based on established cinematographic techniques. Hyper tubes are connected by relations that guarantee smooth transitions. Interval consistency techniques and quantified constraint solving algorithms are used to compute and propagate solutions between consecutive hyper tubes. Preliminary experimental results from a prototype show a great improvement in time and quality of animations with respect to former approaches.