A new selection system for exterior tiles using 2 step FNN models

Abstract

A new selection method for exterior tiles was investigated, in which three colors (standard, light and dark colors) could be decided from the desired images of consumers. FNN (fuzzy neural network) models deciding the standard color and the light and dark colors from sensory evaluation were constructed. Accuracy estimated by the FNN model for the standard color was about 90%, and for that the light and dark colors was about 88%. This selection method is be a promising approach for the direct decision of exterior tiles matching with the image of consumer.