• DocumentCode
    602681
  • Title

    CRISP-II: Coarse-grained reconfigurable image stream processor for image-processing and intelligent operations in QFHD video cameras

  • Author

    Teng-Yuan Cheng ; Liang-Gee Chen ; Shao-Yi Chien

  • Author_Institution
    Dept. of Electr. Eng., Nat. Taiwan Univ., Taipei, Taiwan
  • fYear
    2012
  • fDate
    12-14 Nov. 2012
  • Firstpage
    209
  • Lastpage
    212
  • Abstract
    A 372.3 mW coarse-grained reconfigurable image stream processor, CRISP-II, for image-processing and intelligent operations is implemented in TSMC 90 nm low-power technology with a core size of 15.21 mm2. With the proposed multi-stream mode unified protocol, hierarchical ring architecture, and adaptive computing engine, CRISP-II is able to solve the emerging scalability and flexibility problems. It could execute several advanced operations efficiently, like high-dynamic-imaging and face detection. Compared with other state-of-the-art processors, CRISP-II achieves 8.42 times power efficiency than the highly parallel SIMD processor, and meets the real-time requirement of video cameras with QFHD (3840 × 2160) resolution.
  • Keywords
    face recognition; parallel processing; protocols; reliability; video cameras; CRISP-II; QFHD video cameras; adaptive computing engine; coarse-grained reconfigurable image stream processor; face detection; flexibility problems; hierarchical ring architecture; high parallel SIMD processor; high-dynamic-imaging; image-processing; intelligent operations; multistream mode unified protocol; power 372.3 mW; scalability problems; size 90 nm;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Solid State Circuits Conference (A-SSCC), 2012 IEEE Asian
  • Conference_Location
    Kobe
  • Type

    conf

  • DOI
    10.1109/IPEC.2012.6522662
  • Filename
    6522662