An effective implementation of vector quantization encoder on commodity graphics hardware

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper presents an effective implementation of vector quantization (VQ) encoder on a commodity personal computer (PC). A programmable graphics processing unit (GPU) is used for finding the nearest codeword to each data vector, so-called nearest neighbor search, which dominates the most execution time of the VQ encoding process. In our implementation, the nearest neighbor search is bound to GPU's programmable rendering pipeline so that the GPU can well exhibit its parallel-processing ability. Experimental results show that the single-instruction multiple-data nature of the nearest neighbor search allows the GPU to efficiently execute the vector quantization encoding process. Comparing with vector quantization encoder fully implemented on a general-purpose microprocessor, our implementation can achieve better computational efficiency for the VQ encoding process with a large codebook.

Original languageEnglish
Title of host publicationProceedings of the Second International Conference on Information Technology and Applications (ICITA 2004)
Pages269-274
Number of pages6
Publication statusPublished - 2004 Dec 27
EventProceedings of the Second International Conference on Information Technology and Applications (ICITA 2004) - Harbin, China
Duration: 2004 Jan 92004 Jan 11

Publication series

NameProceedings of the Second International Conference on Information Technology and Applications (ICITA 2004)

Other

OtherProceedings of the Second International Conference on Information Technology and Applications (ICITA 2004)
CountryChina
CityHarbin
Period04/1/904/1/11

Keywords

  • Graphics processing unit
  • Image data compression
  • Programmable rendering pipeline
  • Vector quantization

ASJC Scopus subject areas

  • Engineering(all)

Fingerprint Dive into the research topics of 'An effective implementation of vector quantization encoder on commodity graphics hardware'. Together they form a unique fingerprint.

  • Cite this

    Takizawa, H., & Kobayashi, H. (2004). An effective implementation of vector quantization encoder on commodity graphics hardware. In Proceedings of the Second International Conference on Information Technology and Applications (ICITA 2004) (pp. 269-274). (Proceedings of the Second International Conference on Information Technology and Applications (ICITA 2004)).