Abstract
As evidenced by many physiological and
physiological reports, the receptive fields of the first-stage set of mechanisms
of the visual process fit to 2D compactly supported harmonic functions.
The application of this set of band-pass filter functions to the input
signal implies that the human visual system (HVS) performs some kind of
conjoint space/spatial frequency transform. In this paper we present a
procedure to obtain the linear characterization of the HVS performance
in any conjoint domain from the classical linear characterization in the
Fourier domain. We have called this weighting function (in the particular
case of the Gabor transform) the Gabor stimuli Sensitivity Function
(GSF) by analogy with the usually employed weighting function in the Fourier
domain: the Contrast Sensitivity Function (CSF). The accuracy of
the procedure is proved showing the equivalence between some experimental
2D CSFs and the corresponding GSFs. The main adventage of this new characterization
is that a weighting function in a space/spatial frequency domain can account
for spatially variant behaviour, which cannot be included in a unique CSF,
so a single GSF would be needed to include extra-foveal and large eccentricity
behaviour. One example is given of how non-homogeneous systems can be easily
characterized in this way. |