• DocumentCode
    3129278
  • Title

    The Use of k-Means Algorithm to Compute the Line Spectrum Pair Frequencies with Tschirnhaus Transform

  • Author

    Chen, Shi-Huang ; Hsu, Ming-Lung

  • fYear
    2010
  • fDate
    15-17 Oct. 2010
  • Firstpage
    288
  • Lastpage
    291
  • Abstract
    This paper proposes a fast 10-order line spectrum pair (LSP) frequencies calculation method using the k-means algorithm and the Tschirnhaus transforms. The first step of the proposed method is to derive a quartic equation from the 1st derivative of the given 5-degree LSP polynomial. Then the extremes of the 5-degree LSP polynomial can be found by applying the Tschirnhaus transform to the above quartic equation. By the use of k-means algorithm, the proposed method can build up precise cosine lookup tables with the minimum memory size for Tschirnhaus transform. Finally these extremes could be used as the initial approximations to solve the roots of the 5-degree LSP polynomial via the Newton method and get the accurate LSP frequencies. One of the main advantages of the proposed method is the rapid root determination of a quartic equation without complex number operations and resulting in considerable computational saving. Compared to other methods, the proposed algorithm can determine the precise LSP frequencies with the lowest computational complexity as well as memory usage.
  • Keywords
    computational complexity; signal processing; wavelet transforms; LSP; Newton method; Tschirnhaus transform; Tschirnhaus transforms; computational complexity; frequencies calculation method; k-means algorithm; line spectrum pair frequencies; memory usage; quartic equation; Erbium; Newton method; Polynomials; Signal processing algorithms; Speech; Table lookup; Transforms;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Intelligent Information Hiding and Multimedia Signal Processing (IIH-MSP), 2010 Sixth International Conference on
  • Conference_Location
    Darmstadt
  • Print_ISBN
    978-1-4244-8378-5
  • Electronic_ISBN
    978-0-7695-4222-5
  • Type

    conf

  • DOI
    10.1109/IIHMSP.2010.79
  • Filename
    5638032