Applications of the linear canonical transform to digital image processing

Abstract

In this paper, an existing approximation of discrete linear canonical transform (DLCT) is analyzed, and constraints are derived to fulfill some paramount properties as inversibility and additivity or the possibility to perform classical image operations in the frequency domain as image filtering. Giving some special values to the DLCT parameters and taking advantage of the division of the image spectrum in four zones of different significance, an application of image feature classifications is successfully investigated. Also, the required constraints are obtained to determine the suitability of the selected approximation when working with digital images.