Invention Grant
US08736467B2 Adaptive binarization for arithmetic coding 有权
算术编码的自适应二值化

  • Patent Title: Adaptive binarization for arithmetic coding
  • Patent Title (中): 算术编码的自适应二值化
  • Application No.: US13638803
    Application Date: 2011-04-08
  • Publication No.: US08736467B2
    Publication Date: 2014-05-27
  • Inventor: Frank Jan Bossen
  • Applicant: Frank Jan Bossen
  • Applicant Address: JP Tokyo
  • Assignee: NTT DoCoMo, Inc.
  • Current Assignee: NTT DoCoMo, Inc.
  • Current Assignee Address: JP Tokyo
  • Agency: Brinks Gilson & Lione
  • International Application: PCT/US2011/031782 WO 20110408
  • International Announcement: WO2011/127403 WO 20111013
  • Main IPC: H03M7/00
  • IPC: H03M7/00
Adaptive binarization for arithmetic coding
Abstract:
The present invention first provides adaptive binarization in which a binarizer outputs binary symbol in length, which is variable adaptively to the probability of the source. When the probability is low, it is desirable to decrease the length of the binary symbols to improve the efficiency of arithmetic coding and reduce the complexity of coding calculation. On the other hand, when the probability is high, it is desirable to increase the length of the binary symbols to improve the overall process speed of a decoder. Specifically, a binarizer, according to the present invention, binarizes mapping unit values from a non-binary symbol into binary symbols. The number of binary symbols is inversely proportional to the size of the mapping unit value. In the present invention, the mapping unit value is made variable adaptively to the probability. Thus, the number of binary symbols from the binarizer is also variable adaptively to the probability parameter.
Public/Granted literature
Information query
Patent Agency Ranking
0/0