Derivation and Evaluation of Fundamental Hue Scales

Abstract

ABSTRACT Hue can be described with two types of scales for hue discrimination and hue appearance purposes. A fundamental and physiologically plausible model, called FHS for Fundamental Hue Scales, with both types of hue scales, built directly from cone fundamentals, was introduced in the previous study. In the current study, the FHS model was improved. The first step of the FHS model is a simple conversion of the LMS cone responses into the opponent axes. The opponent axes can be converted into a hue discrimination scale using a rotation matrix that is generated based on five principal hues, or they can be converted into a hue appearance scale using a rotation matrix based on four unique hues. The hue discrimination scale is called , and the hue appearance scale is called . The proposed hue model was evaluated in terms of linearity, spacing, and hue quadrature using traditional datasets including the Munsell, NCS, Hung–Berns, and Ebner–Fairchild. In addition to the available datasets, three experiments were completed to evaluate the FHS model for stimuli with higher saturation and higher luminance levels. In the first experiment, constant hue surfaces of eight hue angles were found to assess the linearity of the hue scales. In the second experiment, the spacing between 10 pairs of constant hue surfaces was examined. The third experiment was conducted to evaluate the hue quadrature of intermediate hues. Results showed that the FHS is successful in terms of linearity, spacing, and hue quadrature prediction. The FHS hue scales are mathematically very simple, physiologically and perceptually plausible, and directly based on cone fundamentals to allow easy generalization to individual observers and perform at least as well as more complicated models.

Abstract

Related Papers