VISUAL SPACE PERCEPTION MODEL IDENTIFICATION BY EVOLUTIONARY SEARCH
DS 36: Proceedings DESIGN 2006, the 9th International Design Conference, Dubrovnik, Croatia
Year: 2006
Editor: Marjanovic, D.
Author: Bittermann, M.; Sariyildiz, S.; Ciftcioglu, Ö.
Section: METHODS AND TOOLS IN DESIGN PRACTICE
Page(s): 185-192
Abstract
Visual perception of Spaces is relevant for design. Designs, which satisfy perceptual requirements are found based on assessments of perceptual implications. For this purpose a probabilistic model of human visual space perception is used. Focus of this paper is the identification of optimal model parameters, so that the perception model matches the perception of human experimenters. This is accomplished by genetic algorithm, which is an evolutionary optimization method from the domain of computational intelligence, which is able to deal with the probabilistic and discrete nature of the perception model to be identified.
Keywords: modelling visual space perception, measurement of perceptual space properties, genetic algorithm