Visual feature integration and categorization:: A Galois lattice-based model.

According to analytic models of visual perception, visual features are coded independently before integration by attentional focalization. The parallel or sequential detections are explained by different ways of dividing attention in the visual field. The model based on the Galois lattice explains these phenomena in terms of processing similarity and differentiation of the objects' properties. These processes occur either during the preattentive stage or the attentive stage according to the distribution of the properties across all the objects. The main predictions of the model were tested by an immediate memorization task concerning the objects' properties. The results show an effect of similarity and differenciation on the memorization of the properties. This effect is independent of the integration of the objects' properties. It seems necessary to consider also the nature of the visual properties in order to explain the effects of similarity and differenciation on attentional integration.