Preattentive texture segmentation and grouping by the boundary contour system

An improved boundary contour system (BCS) neural network model of preattentive vision is applied to two images that produce strong `pop-out´ of emergent groupings in humans. The improved BCS model is able to segment multi-element textures in a manner analogous to texture segmentation by human observers. In humans, these images generate groupings that are collinear with or perpendicular to image contrasts. Analogous groupings occur in computer simulations of the model. Long-range cooperative and short-range competitive processes of the BCS dynamically form the stable groupings of texture regions in response to the images. The groupings of the model shown attempt to emulate one of the most difficult and fundamental of human visual competencies: the dynamic, automatic synthesis of the perceptually salient units in a visual scene