A new class of reciprocal-orthogonal parametric transforms

  • Authors:

  • Saad Bouguezel;M. Omair Ahmad;M. N. S. Swamy

  • Affiliations:

  • Department of Electronics, Faculty of Engineering, Setif University, Setif, Algeria;Center for Signal Processing and Communications, Department of Electrical and Computer Engineering, Concordia University, Montreal, QC, Canada;Center for Signal Processing and Communications, Department of Electrical and Computer Engineering, Concordia University, Montreal, QC, Canada

  • Venue:
  • IEEE Transactions on Circuits and Systems Part I: Regular Papers
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper, a new class of reciprocal-orthogonal parametric (ROP) transforms having 3N/2 independent parameters for a sequence length N that is a power of two is proposed. The basic idea behind the proposed transforms is to appropriately combine a new parametric kernel with that of the well-known Walsh-Hadamard transform that results in a square parametric matrix operator of order N with some very interesting properties. It is shown that the inverse matrix operator of the proposed class of transforms can be easily obtained by taking the reciprocal of each of the entries of the forward matrix and then transposing the resulting matrix. In addition, a simple method is introduced in order to facilitate the generation of the matrix operator of the proposed ROP transforms. This method is then used specifically to construct new classes of unitary and multiplication-free transforms. Many other new transforms, as well some of the existing ones, can be derived from these proposed ROP transforms. An efficient algorithm is developed for a fast computation of the proposed transforms. In view of the availability of this fast algorithm and the property of easily computable inverse transform, the proposed ROP transforms can be used in many transform-based applications, with their independent parameters providing more degrees of freedom such as affording an additional secret key in watermarking and encryption applications.