Acquisition of Chromatic Adaption Using Neural Networks
Autor: | Ichiro Kimura, Jun Yasuoka, Makoto Kawamura, Fumitaka Nishikawa, Yasuaki Kuroe |
---|---|
Rok vydání: | 1996 |
Předmět: | |
Zdroj: | Journal of the Visualization Society of Japan. 16:277-280 |
ISSN: | 1884-037X 0916-4731 |
DOI: | 10.3154/jvs.16.1supplement_277 |
Popis: | Human color perception is so complicated that a conventional image processing system is unable to recognize color as well as we do. If such a complicated artificial mechanism for color perception is technically realized, it is applied to intelligent robots, “Kansei” engineering for design, and various systems which need human recognition for color.This paper presents an artificial system for chromatic adaption which is a most important function for human color perception. The system is constructed using a neural network in which neurons are arbitrary connected. The network structure is based on a color perception model made from its physiological and psychological knowledge. After learning, the network outputs show almost the same response as we do even if a light source changes its color temperature. Furthermore, the network after learning proves “Helson-Judd” effect that under a red light, light gray looks reddish while dark gray looks slightly blue-green. |
Databáze: | OpenAIRE |
Externí odkaz: |