Perception‐based animation rendering

Abstract

AbstractIn most currently produced computer generated animation there is significant visual information that is literally never seen because of the characteristics of the human visual system. By tailoring the rendering of animation images to match the characteristics of human visual perception, significant computational and image storage savings can be obtained while retaining perceived animation quality.Vision research indicates that the human visual system processes two information channels. The transient channel rapidly processes low spatial resolution full colour information. The sustained channel processes high spatial resolution luminance information at a much slower rate.This paper discusses how the rendering of animation images can be accomplished to match these perception characteristics and achieve significant savings while maintaining animation quality. Computation and storage savings of up to 80 per cent are possible. Several animation segments have been produced to demonstrate the viability of this approach.