Intelligent real-time rendering

Abstract

The perennial and increasing demand for fidelity in real-time computer graphics, coupled with hardware architecture advancement, pose many challenges to researchers as they endeavour to find the optimal solution to accommodate both speed of rendering and quality in interactive rendering applications.

Research done previously attempted to address the frame latency problem by providing mathematical models of the rendering process were either derived from some approximate experimentation or depended on specific application level data structures and they do not guarantee performance.

In this research, we focus on a systematic approach to realising a framework for modelling and control of the real time computer rendering process. The first part of this research centralizes on the modelling aspect of the real time rendering process. We propose a data driven modelling approach that provides accurate representation of the system behaviour of this process.

The second part of this research focuses on the design of the control strategy given the process nature investigated in the first part. The benefits of applying control theory in the context of a computer graphics system are explained and its relative advantage over the performance of existing heuristics and algorithms.