Optimal model of optical parameter extraction for material rendering

Abstract

Hyper-real rendering appropriate for movie-like gameplay is being actively studied in the graphics field. To express human skin or various materials realistically, shading technology that can reflect the physical features of a material is needed. The core factor in rendering is optical parameters, and the accuracy of these parameters based on the material's optical features determines the rendering quality. The parameters required for rendering are extracted with a spectrographic optical shooting device and by curve fitting. Existing processes for parameter extraction require significant time. To overcome this disadvantage, we produced an HDRI (high-dynamic range image) generating program with exposure fusion algorithm. The number of low-dynamic range images for an HDRI was optimised through an experiment. In addition, the amount of sample data, which mostly affect extraction time in automated systems, can be optimised to reduce time loss during the photo-shoot, fusion and extraction processes.

Keywords:

• HDRI
• BSSRDF
• Colour
• Optical parameter

Acknowledgements

This research was partially funded by Institute for Information and Communication Technology Promotion, IITP (2014 CiMR: Physically Based Cinematic Material Rendering Techniques optimised for Gameplay, No. 10043453).