Abstract
A new phenomenon of colour ambiguity became recently known as #TheShoe, i.e., the photo of a shoe of ambiguous colours, seen in front of a dark background. The shoe is perceived as either grey leather with turquoise laces or pink leather with white laces. Spectrometric measurements revealed that the colour loci of the photo are distributed along the L/M axis. Presenting the photo of #TheShoe to 92 subjects, we found that 53 % of the observers perceived leather and laces as grey and turquoise, 34 % as pink and white and 11 % as pink and turquoise, respectively; this was independent of age and gender, and did not correlate with the perceived colours of #TheDress. Answering a questionaire, all subjects reported the scene being illuminated from the front, the colour of the light being neutral (pink/white viewers) or blue-greenish (grey/turquoise viewers). We investigated the phenomenon in more detail by having a subset of the subjects (n=20) match the colours of the shoe (presented on an 7.9" iPad display) with a colour picker program running on a separate display. The results showed a distribution of the settings along the L/M cardinal axis, whereby the settings of grey/turquoise and pink/white viewers formed distinct clusters, correlating with the individual colour names. Thus, the data show that colour ambiguities are not as rare as previously thought; importantly, such ambiguities do not seem to be specific for the daylight locus or S/(L+M) axis, as it has been suggested for the dress phenomenon. An explanation considering specific characteristics of daylight compensation is therefore not applicable in the case of #TheShoe; instead we suggest that the #TheShoe ambiguity can be explained by conflicting cues for anchoring the shoes chromaticities. Analysing the data from #TheDress we will show that similar processes can also explain the dress ambiguity.
Meeting abstract presented at VSS 2018