Color Scales For Image Data: Design And Evaluation

Abstract

In this chapter¹ we discuss the use of color to represent a parameter in a two-dimensional image. We state desirable properties of color scales. We introduce the notion of an optimal color scale, and describe the development of a particular optimal color scale; we state restrictions on the order of colors in an optimal color scale, and we present an algorithm to search for scales that obey those constraints. We briefly discuss the linearization of color scales; Chapter 8 goes into the details. We present the result of such optimization-linearization process in the form of the Linearized Optimal Color Scale (LOCS). We describe observer performance experiments to evaluate the merits of color scales for image data. The evaluations show that observers perform somewhat better with the developed LOCS than with a previously advocated scale, the heated-object color scale, but they perform significantly better with a linearized gray scale than with either of the color scales. We discuss possible reasons for this result.