Purchase this article with an account.
Marina Danilova, John Mollon; The anisotropy of color space. Journal of Vision 2013;13(9):294. doi: https://doi.org/10.1167/13.9.294.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Color discrimination is conventionally thought to depend on two neural signals. One signal, carried by midget ganglion cells, represents the ratio of the long- (L) and middle-wave (M) cone excitations. This signal corresponds to the horizontal axis of the MacLeod-Boynton chromaticity diagram. The second signal, carried by small bistratified ganglion cells, represents the ratio of short-wave (S) cone excitation to some combination of L and M excitations, and corresponds to the vertical axis of the MacLeod-Boynton diagram. Phenomenologically, however, the MacLeod-Boynton space is not symmetrical about a vertical line through the white point: It is divided into reddish and greenish hues by a line that runs obliquely, from unique yellow to unique blue. We scale the MacLeod-Boynton diagram so that the yellow-blue line lies at an angle of ̴9;45° (Danilova & Mollon, 2012, Vision Research, 62, 162-172). Measuring discrimination along lines orthogonal to the yellow-blue line (i.e. lines at +45° in the space), we previously found that forced-choice thresholds were minimal at the category boundary between reddish and greenish hues – and not, say, at a constant value of L/M or of S/(L+M). However, what would happen if we reflect our experiment about the vertical axis of MacLeod-Boynton space, measuring chromatic discriminations along lines that are at -45°? We measured such discriminations for 150-ms, foveal targets divided into four quadrants, requiring observers to report the quadrant that differed in chromaticity from the other three. Observers were adapted to a field metameric to Illuminant D65. Two main results emerge: (i) discrimination along ̴9;45° lines does show a shallow minimum and (ii) for lines not passing through D65, the minimum does not coincide with the L/M value of the background. We conclude that a phenomenological boundary is not necessary for a minimum in discrimination measurements.
Meeting abstract presented at VSS 2013
This PDF is available to Subscribers Only