CUDA accelerated illumination preprocessing on GPUs

Author

Vandal, Nicholas A. ; Savvides, Marios

fYear

2011

fDate

6-8 July 2011

Firstpage

1

Lastpage

6

Abstract

In this paper we develop a parallelized implementation of the anisotropic diffusion image preprocessing algorithm for illumination invariant face recognition proposed by Gross and Brajovic. Our implementation employs Red-Black Gauss-Seidel relaxation running on inexpensive Graphics Processing Units (GPUs) programmed with Nvidia´s CUDA framework. We are able to achieve a 20X speedup over a multithreaded implementation running on a quadcore CPU. Additionally a comparison to an open-source implementation of anisotropic diffusion in the Torch3vision library is performed, demonstrating a GPU speedup of greater than 900X over this commonly used machine vision library.

Keywords

Gaussian processes; computer graphic equipment; coprocessors; image processing; multi-threading; CUDA accelerated illumination preprocessing; GPU; Torch3vision library; graphics processing units; illumination invariant face recognition; image preprocessing algorithm; machine vision library; multithreaded implementation; open source implementation; parallelized implementation; red black Gauss Seidel relaxation; Anisotropic magnetoresistance; Central Processing Unit; Face recognition; Graphics processing unit; Instruction sets; Libraries; Lighting; CUDA; GPGPU; GPU; Illumination Preprocessing; Real-time;

fLanguage

English

Publisher

ieee

Conference_Titel

Digital Signal Processing (DSP), 2011 17th International Conference on

Conference_Location

Corfu

ISSN

Pending

Print_ISBN

978-1-4577-0273-0

Type

conf

DOI

10.1109/ICDSP.2011.6004977

Filename

6004977