Abstract
The perception of an objects' colour remains constant despite large variations in the chromaticity of the illumination; colour constancy. Hering (1920) suggested that memory colours, the typical colours of objects, could help in estimating the illuminant's colour and therefore be an important factor in establishing colour constancy. Gegenfurtner and colleagues (Hansen et al., 2006) have shown that the colour appearance of natural objects can be biased towards the typical colour and that this effect is robust under illuminant changes (Olkkonen et al., 2008). Here we test whether the presence of objects with diagnostical colours (e.g., fruits and vegetables) within a scene influence colour constancy for other unknown objects in the scene. Subjects matched one of four Munsell papers (2.5p 5/4; 2.5y 5/4; 2.5b 5/4 and 2.5gy 5/4) placed in a scene illuminated under either a reddish or a greenish lamp with the Munsell Book of colour illuminated by a neutral lamp. The Munsell papers were embedded in four different scenes; one scene containing diagnostically coloured objects, one scene containing incongruent coloured objects, a third scene with geometrical objects of the same colour as the diagnostically coloured objects and one scene containing non-diagnostically coloured objects (e.g., a yellow coffee mug). All objects were placed against a black background. Colour constancy was on average significantly higher (8%) for the scene containing the diagnostically coloured objects compared to the scene containing the incongruent coloured objects and the scene with the geometrical objects of the same colour and significantly higher (4%) compared to the scene containing the non-diagnostically coloured objects. Placing one of the objects in front of the Munsell paper to guarantee local colour contrast, did not affect colour constancy. We conclude that the colours of familiar objects help in obtaining colour constancy for unknown objects.