Selective Rasterized Ray-traced Reflections on the GPU

Kastrati, Mattias Frid; Goswami, Prashant

Selective Rasterized Ray-traced Reflections on the GPU

Files

045-052.pdf (10.2 MB)

Date

2016

Authors

Kastrati, Mattias Frid
Goswami, Prashant

Publisher

The Eurographics Association

Abstract

Ray-tracing achieves impressive effects such as realistic reflections on complex surfaces but is also more computationally expensive than classic rasterization. Rasterized ray-tracing methods can accelerate ray-tracing by taking advantage of the massive parallelization available in the rasterization pipeline on the GPU. In this paper, we propose a selective rasterized raytracing method that optimizes the rasterized ray-tracing by selectively allocating computational resources to reflective regions in the image. Our experiments suggest that the method can speed-up the computation by up to 4 times and also reduce the memory footprint by almost 66% without affecting the image quality. We demonstrate the effectiveness of our method using complex scenes and animations.

        @inproceedings{10.2312:stag.20161363
,
booktitle = {Smart Tools and Apps for Graphics - Eurographics Italian Chapter Conference
},
editor = {Giovanni Pintore and Filippo Stanco
},
title = {{Selective Rasterized Ray-traced Reflections on the GPU
}},
author = {Kastrati, Mattias Frid and 
Goswami, Prashant
},
year = {2016
},
publisher = {The Eurographics Association
},
ISSN = {-
},
ISBN = {978-3-03868-026-0
},
DOI = {10.2312/stag.20161363
}
}