Visual Saliency for Smell Impulses and Application to Selective Rendering

Abstract

A major challenge in generating high-fidelity virtual environments is to be able to provide interactive rates of realism. However this is very computationally demanding and only recently visual perception has been used in high-fidelity rendering to improve performance considerably by a series of novel exploitations; to render parts of the scene that are not currently being attended by the viewer at a much lower quality without the difference being perceived. This paper investigates the effect various smells have on the visual attention of the user when free viewing a set of engineered images. We verify the worth of investigating these saccade shifts (fast movements of the eyes) due to attention distraction to a congruent smell object. By analysing the gaze points, we identify time spent attending a particular area of a scene. We also present a technique from measured data to remodulate traditional saliency maps of image features to account for the observed results. We show that smell provides an impulse on attention to affect perception in such a way that this can be used to guide selective rendering of scenes through use of the remodulated saliency maps.