Colour constancy is the ability of the human visual system to comprehend the constant colour of an object with no regard to changes in the illumination spectrum. Each year witnesses the appearance of new scientific works in various countries. The works touch upon different aspects of the phenomenon. Although scientific research has been conducted for many years, there is no recognised theory explaining the mechanisms that are decisive for colour constancy. The aim of the present work was to establish the influence of contrast, adaptation and background structure upon colour constancy.
All experiments were performed with the application of asymmetric matching. Five subjects took part in the experiments. Colour constancy was examined by performing experiments in three cycles. The test illuminants were two Planckian illuminants, standard illuminant A (u’= 0.2559, v’ = 0.5243), illuminant S (u’= 0.1744, v’= 0.3923) and illuminants r (u’/v’ = 0.2618/0.4533), g (u’/v’ = 0.1517/0.4667), v (u’/v’ = 0.2116/0.3766), y (u’/v’ = 0.1939/0.5180. The first set of 7/4 (value – 7/, chroma – /4) of 10 Munsell samples served as test stimuli in the experiments of the second cycle. Being 2° in size, they were generated in the centre of the monitor. The samples were surrounded by a neutral background, sized N7 20° (N meaning the colour being neutral, with the value of 7). Around the neutral background, subject of the experiment saw the non-radiating part of the screen, which we called the „black surround“ (Fig. 1). In the second set of the experiment, the subjects observed the stimuli through
a cardboard tube (Fig. 2). Neutral N7 Munsell samples and a set of 7/4 (value – 7/, chroma – /4) of 10 Munsell samples served as test stimuli in the experiments of the third cycle.
Colour constancy is quantitively evaluated on the basis of the Brunswik Ratio (BR) (Troost and de Weert, 1991). Higher values of BR were associated with longer adaptation periods, but only when a larger background was used (second cycle of experiments (Fig. 7)). Supplementary experiments showed that the changes in colour appearance were related to a slight shift in the perceived colour of the background. The timing of the colour shifts was modelled in terms of cone opponent responses.
Our conclusions are:
1. When stimuli are exposed for a short period of time (1 s), partial colour constancy takes place. It does not change (or changes insignificantly) when the stimulus is observed for a longer time (5 and 30 s). Such colour constancy is determined by fast processes which confirm the influence of simultaneous contrast on colour constancy.
2. When the background is small, the influence of adaptation on colour constancy is little, even it the observation time is long (5 or 30 s).
3. When the background fills the entire field of vision, adaptation results in a high colour constancy perception.