• DocumentCode
    545908
  • Title

    Real-Time Scan Conversion for Ultrasound Imaging Based on CUDA with Direct3D Display

  • Author

    Wang, Li ; Shi, Dan ; Zhao, Anyuan ; Tan, Chaowei ; Liu, Dong C.

  • Author_Institution
    Comput. Sci. Coll., Sichuan Univ., Chengdu, China
  • fYear
    2011
  • fDate
    10-12 May 2011
  • Firstpage
    1
  • Lastpage
    4
  • Abstract
    Ultrasonic scan conversion (SC) is one of the most important and widely-used technology in Medical Ultrasound Imaging. However, real-time processing of SC is computationally intensive, involving in calculation of inverse tangents, square roots to determine the scanning angle and depth of a beam followed by interpolation both in axial and lateral directions. In this paper, a very efficient implementation of ultrasonic scan conversion on the Compute Unified Device Architecture (CUDA™) platform developed by NVIDIA® and using the interoperability of Direct3D (D3D) and CUDA to display is presented. By exploiting the implicit parallelism and reasonable performance optimization strategies based on the graphics hardware, this parallel scan conversion algorithm obtains a high speed up. It achieves a frame rate of about more than 746 fps with the picture size of 3121* 936, which is about 300 times faster than the CPU implementation.
  • Keywords
    biomedical ultrasonics; computer graphics; coprocessors; graphical user interfaces; medical image processing; optimisation; CUDA based medical ultrasound imaging; Compute Unified Device Architecture platform; NVIDIA platform; direct3D display; graphics hardware; implicit parallelism; interpolation method; inverse tangents; optimization; real-time processing; real-time scan conversion; square roots; Graphics processing unit; Imaging; Instruction sets; Interpolation; Memory management; Pixel; Ultrasonic imaging;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Bioinformatics and Biomedical Engineering, (iCBBE) 2011 5th International Conference on
  • Conference_Location
    Wuhan
  • ISSN
    2151-7614
  • Print_ISBN
    978-1-4244-5088-6
  • Type

    conf

  • DOI
    10.1109/icbbe.2011.5780361
  • Filename
    5780361