Representation of images through group-delay functions

In this correspondence we propose a new representation for images through the use of group-delay functions. We provide algorithms for the computation of the group-delay functions and for the recovery of images from them. We give several examples to show that most of the perceptually significant information of an image is retained in this representation. We derive the minimum phase equivalent images from the Fourier transform (FT) magnitude, as well as from the phase through their respective group-delay functions. We compare these images to those obtained through iterative reconstruction from the FT magnitude and phase, respectively.