skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Implementing wide baseline matching algorithms on a graphics processing unit.

Abstract

Wide baseline matching is the state of the art for object recognition and image registration problems in computer vision. Though effective, the computational expense of these algorithms limits their application to many real-world problems. The performance of wide baseline matching algorithms may be improved by using a graphical processing unit as a fast multithreaded co-processor. In this paper, we present an implementation of the difference of Gaussian feature extractor, based on the CUDA system of GPU programming developed by NVIDIA, and implemented on their hardware. For a 2000x2000 pixel image, the GPU-based method executes nearly thirteen times faster than a comparable CPU-based method, with no significant loss of accuracy.

Authors:
; ; ;
Publication Date:
Research Org.:
Sandia National Laboratories
Sponsoring Org.:
USDOE
OSTI Identifier:
921737
Report Number(s):
SAND2007-6301
TRN: US200806%%21
DOE Contract Number:
AC04-94AL85000
Resource Type:
Technical Report
Country of Publication:
United States
Language:
English
Subject:
99 GENERAL AND MISCELLANEOUS//MATHEMATICS, COMPUTING, AND INFORMATION SCIENCE; ACCURACY; ALGORITHMS; COMPUTER GRAPHICS; IMPLEMENTATION; PERFORMANCE; PATTERN RECOGNITION; Gaussian processes.; Computer vision.; Visual texture recognition.; Pattern recognition.

Citation Formats

Rothganger, Fredrick H., Larson, Kurt W., Gonzales, Antonio Ignacio, and Myers, Daniel S. Implementing wide baseline matching algorithms on a graphics processing unit.. United States: N. p., 2007. Web. doi:10.2172/921737.
Rothganger, Fredrick H., Larson, Kurt W., Gonzales, Antonio Ignacio, & Myers, Daniel S. Implementing wide baseline matching algorithms on a graphics processing unit.. United States. doi:10.2172/921737.
Rothganger, Fredrick H., Larson, Kurt W., Gonzales, Antonio Ignacio, and Myers, Daniel S. Mon . "Implementing wide baseline matching algorithms on a graphics processing unit.". United States. doi:10.2172/921737. https://www.osti.gov/servlets/purl/921737.
@article{osti_921737,
title = {Implementing wide baseline matching algorithms on a graphics processing unit.},
author = {Rothganger, Fredrick H. and Larson, Kurt W. and Gonzales, Antonio Ignacio and Myers, Daniel S.},
abstractNote = {Wide baseline matching is the state of the art for object recognition and image registration problems in computer vision. Though effective, the computational expense of these algorithms limits their application to many real-world problems. The performance of wide baseline matching algorithms may be improved by using a graphical processing unit as a fast multithreaded co-processor. In this paper, we present an implementation of the difference of Gaussian feature extractor, based on the CUDA system of GPU programming developed by NVIDIA, and implemented on their hardware. For a 2000x2000 pixel image, the GPU-based method executes nearly thirteen times faster than a comparable CPU-based method, with no significant loss of accuracy.},
doi = {10.2172/921737},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Mon Oct 01 00:00:00 EDT 2007},
month = {Mon Oct 01 00:00:00 EDT 2007}
}

Technical Report:

Save / Share: