• DocumentCode
    1306802
  • Title

    Paraunitary filter banks over finite fields

  • Author

    Phoong, See-May ; Vaidyanathan, P.P.

  • Author_Institution
    Dept. of Electr. Eng., California Inst. of Technol., Pasadena, CA, USA
  • Volume
    45
  • Issue
    6
  • fYear
    1997
  • fDate
    6/1/1997 12:00:00 AM
  • Firstpage
    1443
  • Lastpage
    1457
  • Abstract
    In real and complex fields, unitary and paraunitary (PU) matrices have found many applications in signal processing. There has been interest in extending these ideas to the case of finite fields. We study the theory of PU filter banks (FBs) in GF(q) with q prime. Various properties of unitary and PU matrices in finite fields are studied. In particular, a number of factorization theorems are given. We show that (i) all unitary matrices in GF(q) are factorizable in terms of Householder-like matrices and permutation matrices, and (ii) the class of first-order PU matrices (the lapped orthogonal transform in finite fields) can always be expressed as a product of degree-one or degree-two building blocks. If q>2, we do not need degree-two building blocks. While many properties of PU matrices in finite fields are similar to those of PU matrices in complex field, there are a number of differences. For example, unlike the conventional PU systems, in finite fields, there are PU systems that are unfactorizable in terms of smaller building blocks. In fact, in the special case of 2×2 systems, all PU matrices that are factorizable in terms of degree-one building blocks are diagonal matrices. We derive results for both the cases of GF(2) and GF(Q) with q>2. Even though they share some similarities, there are many differences between these two cases
  • Keywords
    Galois fields; band-pass filters; filtering theory; matrix algebra; signal processing; transforms; Galois fields; Householder like matrices; complex fields; degree-one building blocks; degree-two building blocks; diagonal matrices; factorization theorems; finite fields; first-order PU matrices; lapped orthogonal transform; matrix factorization; paraunitary filter banks; paraunitary matrices; permutation matrices; real fields; signal processing; unitary matrices; Channel bank filters; Energy conservation; Filter bank; Finite impulse response filter; Galois fields; Image coding; Image reconstruction; Quantization; Signal processing; Video sharing;
  • fLanguage
    English
  • Journal_Title
    Signal Processing, IEEE Transactions on
  • Publisher
    ieee
  • ISSN
    1053-587X
  • Type

    jour

  • DOI
    10.1109/78.599956
  • Filename
    599956