Om visuel genkendelse af form

Abstract

Our capacity for visual recognition on the basis of shape but regardless of size or orientation has been investigated by reaction time methods in numerous studies in the past several years. The general finding is that if two visual patterns differ with respect to size or
orientation but not otherwise, then the time taken to decide that the patterns are identical increases approximately linearly as a function of the linear size ratio or the angular orientation between the pattems. According to the transformational account of these results it is assumed that recognition is achieved by first encoding one of the patterns as a visual image and then mentally transforming the image to fit the impression of the other pattem. Mentally transforming the image, i.e. processes of mental rotations and mental size transformations may be considered as basic visual routines which may be assembled to programs
that will allow us to identify patterns that differ with respect to both size and orientation. The transformational account of simple visual pattem recognition is introduced by way of short and necessarily highly selective reviews of the classical studies of Mach, Høffding and others.