Aaron Bobick

Abstract

One of the tenets of machine vision is that choice of representation is critical in describing a scene; representations are selected that make important information explicit and that allow domain constraints to be brought to bear. In dynamic scene annotation, there may be no single representation appropriate for analysis, as the imagery is dynamic and choice of best representation may vary over time. The authors present two examples of where the selection of representational frame-the coordinate system in which the representation describes the scene-is determined by the dynamics of the scene.