Human recognition of patterns and objects relies on previously acquired categories that guide perceptual classification. Our psychophysical approach to categorization employs sets of so-called compound Gabor patterns, grey-level patterns that are defined within a two-dimensional evenness-oddness Fourier feature space. As illustrated in the figure, this space provides a low-dimensional continuum of form, in which each point uniquely determines the structure of a pattern, and clusters of points may be used to define classes to be learned by an observer.
Above paradigm has been used to investigate
Ingo Rentschler (Institute for Medical Psychology, University of Munich, Germany)
Terry Caelli (National Information and Communications Technology Centre of Excellence (NICTA), Canberra, Australia)
Theodor Landis (Service de Neurologie, University Hospital Geneva, Switzerland)