Image Analysis and Processing, International Conference on (2001)
Sept. 26, 2001 to Sept. 28, 2001
Giuseppe Boccignone , Universit? di Salerno
Mario Ferraro , Universit? di Torino
Terry Caelli , The University of Alberta
Abstract: An approach to active vision based on information theory and statistical mechanics is presented. Density of entropy production measured along a spatio-chromatic diffusion of a color image is used to build a conspicuity map of the image. The map is successively given as input to a dynamical neural network in order to drive a focus-of-attention scanpath.
G. Boccignone, T. Caelli and M. Ferraro, "An Information-Theoretic Approach to Active Vision," Image Analysis and Processing, International Conference on(ICIAP), Palermo, Italy, 2001, pp. 0340.