A Compound Computational Model for Filling-In Processes Triggered by Edges: Watercolor Illusions

Abstract

The goal of our research was to develop a compound computational model with the ability to predict different variations of the “watercolor effects” and additional filling-in effects that are triggered by edges. The model is based on a filling-in mechanism solved by a Poisson equation, which considers the different gradients as “heat sources” after the gradients modification. The biased (modified) contours (edges) are ranked and determined according to their dominancy across the different chromatic and achromatic channels. The color and intensity of the perceived surface are calculated through a diffusive filling-in process of color triggered by the enhanced and biased edges of stimulus formed as a result of oriented double-opponent receptive fields. The model can successfully predict both the assimilative and non-assimilative watercolor effects, as well as a number of “conflicting” visual effects. Furthermore, the model can also predict the classic Craik–O'Brien–Cornsweet (COC) effect. In summary, our proposed computational model is able to predict most of the “conflicting” filling-in effects that derive from edges that have been recently described in the literature, and thus supports the theory that a shared visual mechanism is responsible for the vast variety of the “conflicting” filling-in effects that derive from edges.

Computational modeling of color perception with biologically plausible spiking neural networks

“…Furthermore, we demonstrate individual differences in color perception using the #theDress and #theShoe images. Our proposed model can further explain both the watercolor [ 14 ] and the Cornsweet illusion [ 16 ] through the reconstruction of images from adapted gradients, as we recently demonstrated [ 32 ].…”

“…In V1, visual data is represented as Spatio-temporal edges, constituting the image’s gradients. The perception of filled surfaces from image gradients can be described using the diffusion/heat equation: where is the gradient operator, is the Laplacian operator, div is the divergence ( ), I is the perceived image (i.e., the reconstructed image), and I input is the input image (stimulus) [ 31 ], [ 32 ]. In the diffusion process, the inactive center of the V1-represented stimulus is gradually filled-in with neuronal activity, supporting the perception of light intensity at the center of the outlined stimulus.…”

“…), I is the perceived image (i.e., the reconstructed image), and I input is the input image (stimulus) [31], [32]. In the diffusion process, the inactive center of the V1-represented stimulus is gradually filled-in with neuronal activity, supporting the perception of light intensity at the center of the outlined stimulus.…”

Central mechanisms of perceptual filling-in

“…This suggests that neural responses are integrated across double-opponent cells along the path from V1 to V3A. A recent computational model, in fact, takes into account oriented double-opponent cells to compute the perceived surface from a filling-in process (Cohen-Duwek & Spitzer, 2019). This model is able to predict the filling-in phenomenon depending on distant contours such as the COCE and the WCE.…”

Signals from Single-Opponent Cortical Cells in the Human cVEP

“…In this way, the Double-Opponent neurons provide the signals that allow people to perceive the colors of colored-surfaces (Livingstone and Hubel, 1984;Friedman et al, 2003;Johnson et al, 2008;Nunez et al, 2018). Double-Opponent neurons have also been proposed as a possible source of the watercolor effect (Devinck et al, 2014;Cohen-Duwek and Spitzer, 2019; for review, see Devinck and Knoblauch, 2019). Single-Opponent cells may be important in judging the color of the illuminant of a scene and in detecting shallow spatial gradients of illumination color (Johnson et al, 2008;Nunez et al, 2018;Shapiro et al, 2018).…”