Reducing IO bandwidth for GPU based moment invariant classifier systems

C. H. Messom, A. L C Barczak

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review

1 Citation (Scopus)

Abstract

This paper introduces an IO bandwidth reduction technique for real-time moment invariant classifier systems running on both CPUs and GPUs. This system can run in real time on commodity general purpose graphics processor unit (GPGPU) systems. The output IO is reduced by calculating the locations of objects of interest using a projection of the 2D classified outputs onto the two axes of the image. The two projections are then used to calculate the positions of a large proportion of the hits in the original image. For a system with a low number of hits there is no loss during this compression, while a system with a large number of hits only suffer losses in a small number of degenerate cases that have a low probability of occurrence in real classifier systems. Lower compression rate approaches can reduce the probability of losses at the expense of higher bandwidth and potentially lower frame rates.

Original languageEnglish
Title of host publication2009 IEEE Intrumentation and Measurement Technology Conference, I2MTC 2009
Pages1200-1204
Number of pages5
DOIs
Publication statusPublished - 2009
Externally publishedYes
EventIEEE International Instrumentation and Measurement Technology Conference 2009 - Singapore, Singapore
Duration: 5 May 20097 May 2009
Conference number: 26th
https://ieeexplore.ieee.org/xpl/conhome/5159258/proceeding (Proceedings)

Conference

ConferenceIEEE International Instrumentation and Measurement Technology Conference 2009
Abbreviated titleI2MTC 2009
Country/TerritorySingapore
CitySingapore
Period5/05/097/05/09
Internet address

Keywords

  • Classifiers
  • GPGPU
  • Image processing
  • Moment invariant
  • Summed area tables

Cite this