Publication detail

Image Processing Methods Optimization by Means of GPU Computing

ŠKORPIL, V. KOUBEK, T. ENDRLE, P.

Original Title

Image Processing Methods Optimization by Means of GPU Computing

English Title

Image Processing Methods Optimization by Means of GPU Computing

Type

conference paper

Language

en

Original Abstract

Optimization of computer resources allocation is crucial for real-time image processing applications. One of the optimization possibilities is the use of GPU (Graphics Processing Unit) instead of CPU for computing. The article explains augmented reality application (AuRel) and the possibility of speeding-up the image pre-processing phase. Further, two image processing algorithms – threshold and convolution are tested. These algorithms are tested in various implementations from naive code performed on CPU to optimized code running on GPU.

English abstract

Optimization of computer resources allocation is crucial for real-time image processing applications. One of the optimization possibilities is the use of GPU (Graphics Processing Unit) instead of CPU for computing. The article explains augmented reality application (AuRel) and the possibility of speeding-up the image pre-processing phase. Further, two image processing algorithms – threshold and convolution are tested. These algorithms are tested in various implementations from naive code performed on CPU to optimized code running on GPU.

Keywords

Image processing, GPU (Graphics Processing Unit), GPGPU, CUDA, OpenCL, Augmented Reality

RIV year

2012

Released

14.07.2012

Publisher

WSEAS

Location

Kos Island, Greece

ISBN

978-1-61804-109-8

Book

Recent Researches in Communications and Computers

Edition number

1

Pages from

95

Pages to

100

Pages count

6

BibTex


@inproceedings{BUT93517,
  author="Vladislav {Škorpil} and Tomáš {Koubek} and Pavel {Endrle}",
  title="Image Processing Methods Optimization by Means of GPU Computing",
  annote="Optimization of computer resources allocation is crucial for real-time image processing applications. One of the optimization possibilities is the use of GPU (Graphics Processing Unit) instead of CPU for computing. The article explains augmented reality application (AuRel) and the possibility of speeding-up the image pre-processing phase. Further, two image processing algorithms – threshold and convolution are tested. These algorithms are tested in various implementations from naive code performed on CPU to optimized code running on GPU.",
  address="WSEAS",
  booktitle="Recent Researches in Communications and Computers",
  chapter="93517",
  howpublished="print",
  institution="WSEAS",
  year="2012",
  month="july",
  pages="95--100",
  publisher="WSEAS",
  type="conference paper"
}