Rendering high quality images in real time

The Problem

The film and animation industries' demand for higher quality realistically rendered images is increasing dramatically. Rendering such high fidelity images of complex scenes takes a significant amount of computational resources and is very expensive.

The Research

• To develop new high-fidelity rendering software.

• To investigate visual perception techniques to reduce significantly rendering times while retaining high perceptual quality.

• To create a platform-independent parallel processing framework.

•  To be able to accurately predict rendering time from rapid profiling information.

The Achievements

• Saliency and Task Maps which increase efficiency by ensuring the quality of rendering is appropriate to the relative perceptual importance of any pixel.

• A Parallel Processing Engine to optimize the use of computing resources.

• The SharpEye rendering system, a physically based global illumination system capable of producing high-fidelity images.

• A Selective Rendering Framework to increase significantly the quality of images rendered for a given level of computing power or for a specified time constraint.

• A cost prediction system which provides accurate quotes for clients.

Key Applications of Use

• Archaeology

• Architectural design

• Games

• Film production