Purchase this article with an account.
Michael E. Rudd, Iris K. Zemach; A quantitative model of achromatic color induction based on separate lightness and darkness filling-in processes. Journal of Vision 2002;2(10):23. doi: 10.1167/2.10.23.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The idea that the brain computes achromatic color by a neural filling-in mechanism that derives its input from border contrast information is a longstanding hypothesis in visual perception research. Here, we present psychophysical evidence that supports a quantitative model of achromatic color induction based on such a mechanism (Rudd, 2001; Rudd & Arrington, 2001; Rudd & Zemach, 2002). Using a lightness matching paradigm and stimuli that consist of side-by-side pairs of disks and surround rings, we demonstrate the validity of the following postulates: 1) Achromatic color appearance results from the linear combination of separate underlying lightness and darkness induction signals. 2) The lightness or darkness polarity of these induction signals is determined by the contrast polarity of their inducing borders. 3) Induction signal magnitude is proportional to the log luminance ratio associated with the inducing border. 4) Induction signal magnitude falls off with distance from the inducing border. Together, these postulates form the basis of the quantitative induction model. In a particularly strong test of this model we demonstrate that, by first measuring the magnitudes of the induction effects produced by individual borders for particular psychophysical observers, we can predict the achromatic color matches made by these same observers when multiple borders are combined to create a more complex visual display. Finally, we consider ways in which the model could be instantiated in a neural system. We show that our data is inconsistent with previously proposed neural network models of the filling-in process and we present a new model of achromatic color computation based on the assumption that achromatic color is computed by combining the outputs of separate lightness and darkness filling-in networks.
Rudd, M. E. (2001). Lightness computation by a neural filling-in mechanism. SPIE Proceedings, 4299, 400–413.
Rudd, M. E., & Arrington, K. F. (2001). Darkness filling-in: A neural model of darkness induction. Vision Research, 41, 3649–3662.
Rudd, M. E., & Zemach, I. K. (2002). Contrast, assimilation, and neural edge integration. Vision Sciences Society Absracts.
This PDF is available to Subscribers Only