GPU-assisted evolutive image predictor generation (画像工学) GPU-Assisted Evolutive Image Predictor Generation

Search this Article

Author(s)

Abstract

Evolutive Image Coding has shown promising results in efficiency compared to other lossless coding methods, but until now, the processing power required for the fitness evaluation has limited its usefulness outside of large computer clusters. Using the CUDA programming language on comparatively inexpensive NVIDIA graphics cards, we have obtained speed increases of up to 150 times for the fitness evaluation. Some of the techniques we have used to improve performance include utilizing the GPU's fast shared memory whenever possible as well as performing some calculations for which the GPU is not as well suited, such as a histogram-based calculation, on the CPU while the GPU simultaneously calculates the fitness evaluation in order to minimize idle time.

Journal

  • IEICE technical report

    IEICE technical report 110(275), 25-28, 2010-11-04

    The Institute of Electronics, Information and Communication Engineers

References:  7

Cited by:  1

Codes

  • NII Article ID (NAID)
    110008152437
  • NII NACSIS-CAT ID (NCID)
    AN10013006
  • Text Lang
    ENG
  • Article Type
    Journal Article
  • ISSN
    09135685
  • NDL Article ID
    10914895
  • NDL Source Classification
    ZN33(科学技術--電気工学・電気機械工業--電子工学・電気通信)
  • NDL Call No.
    Z16-940
  • Data Source
    CJP  CJPref  NDL  NII-ELS 
Page Top