Hue is a fundamental dimension of colour appearance that is usually described in terms of unique hues. Specifically, hue is quantified in terms of the relative amounts of unique hues (called chromatic components). Hue scaling and hue cancellation techniques are traditionally used to measure the chromatic components. The hue-cancellation technique is essentially based on the assumption that unique hues are produced by linear opponent colour channels. However, there is strong evidence that these channels are not linear. Furthermore, there is no satisfactory definition of unique hues. Reducing it to verbal categories, as it is usually done, leaves actually to observers to decide which hue is, say, unique red. We have developed a hue scaling technique to evaluate the chromatic components which neither involves hue naming, nor requires observers to estimate directly the amount of unique hues in compound colours. It is based on the partial colour matching method (Logvinenko, 2006), in which observers class two colours as partially matching if they consider them to share a component hue in common (e.g., lilac and orange as they share red). Using a set of 20 Munsell chips we have conducted an experiment on partial colour matching with five normal trichromats. The whole set of chips can be partitioned into chromaticity classes each of which is defined as the largest sets of chips which partially match each other. We have proven that a chromaticity class comprises all chips having the same unique hue. Hence, there is one-to-one correspondence between chromaticity classes and unique hues. Chromaticity classes for each observer have been derived from the partial colour matching data.
展开▼
