Neural basis for the design of fractional discriminant functions

Fractional discrimination is a process by which an aggregated continuous differentiation of any signal can be carried out. In this paper the author presents the basis for the design of fractional discriminant functions for image processing based on the visual receptive fields. It is hypothesised in this paper that the biological visual system employs some orders of fractional discrimination to carry out visual perception. It is shown that perceptual information of good quality is extracted despite poor PSNR and NMSE in images. It is the author´s observation that merely improving the SNR and the statistical measures of an image does not necessarily improve the perceptual quality of the image on the contrary the author has shown by employing fractional discriminant functions that perceptual quality is better in spite of poor statistical measurements