A Functional and Statistical Bottom-Up Saliency Model to Reveal the Relative Contributions of Low-Level Visual Guiding Factors

Abstract

Springer New York. ISSN : 1866-9956International audienceWhen looking at a scene, we frequently move our eyes to place consecutive interesting regions on the fovea, the retina centre. At each fixation, only this specific foveal region is analyzed in detail by the visual system. The visual attention mechanisms control eye movements and depend on two types of factor: bottom-up and top-down factors. Bottom-up factors include different visual features such as colour, luminance, edges, orientations, etc. In this paper, we evaluate quantitatively the relative contribution of basic low-level features as candidate guiding factors to visual attention and hence to eye movements. We also study how these visual features can be combined in a bottom-up saliency model. Our work consists of three interactive parts: a functional saliency model, a statistical model and eye movement data recorded during free viewing of natural scenes. The functional saliency model, inspired by the primate visual system, decomposes a visual scene into different feature maps. The statistical model indicates which features best explain the recorded eye movements. We show an essential role of high frequency luminance and an important contribution of central fixation bias. The relative contribution of features, calculated by the statistical model, is then used to combine the different feature maps into a saliency map. Finally, the comparison between the saliency model and experimental data confirmed the influence of these contributions

Efficient human face detection in infancy

“…We found that, by 6 months of age, human faces, compared with primate and mammal faces, were more likely to be detected, were detected more quickly, and held attention longer, supporting the proposal that own‐species faces are biologically important in infancy (e.g., Sanefuji et al, ). Critically, our stimulus analysis (see Supplemental Materials) revealed that these results are unlikely to be due to low‐level salience (Gluckman & Johnson, 2013; Hershler & Hochstein, 2005; Ho‐Phuoc, Guyader, & Guérin‐Dugué, ; Walther & Koch, 2006).…”

Preference for human eyes in human infants

“…Luminance values were estimated for each face on the basis of spectrophotometric measurements of screen emittance (SpectraScan PR650, PhotoResearch). With respect to contrast, the effect of inserting human eyes into a non-human primate face is illustrated here by the Chimpanzee (b) and Barbary macaque (d) saliency maps, expressed in arbitrary units normalized for the maximum saliency found over the set of stimuli (Ho Phuoc, Guyader, & Guérin-dugué, 2010). As shown, inserting human eyes into non-human primate faces resulted in a significant increase of contrast in the eye region where saliency for each pair was, on average, multiplied by 2.8 compared to stimuli with original eyes ( t [5] = 6.81, p < .005).…”

Face detection in 2‐ to 6‐month‐old infants is influenced by gaze direction and species

“…For all arrays, we chose heterogeneous images instead of matching images on low‐level features (Hershler & Hochstein, ). A Saliency toolbox analysis ensured that there were not significant differences in low‐level salience, including features such as color or shape (Gluckman & Johnson, ; Hershler & Hochstein, ; Ho‐Phuoc, Guyader, & Guérin‐Dugué, ; Walther & Koch, ).…”

“…A Saliency toolbox analysis ensured that there were not significant differences in low-level salience, including features such as color or shape (Gluckman & Johnson, 2013;Hershler & Hochstein, 2005;Ho-Phuoc, Guyader, & Guérin-Dugué, 2010;Walther & Koch, 2006).…”