A new approach to modeling the influence of image features on fixation selection in scenes.
Bottom Line: Which image characteristics predict where people fixate when memorizing natural images?To answer this question, we introduce a new analysis approach that combines a novel scene-patch analysis with generalized linear mixed models (GLMMs).Importantly, neither luminance nor contrast had an independent effect above and beyond what could be accounted for by the other predictors.
Affiliation: Psychology Department, School of Philosophy, Psychology and Language Sciences, University of Edinburgh, United Kingdom.Show MeSH
Related in: MedlinePlus
Mentions: This paper introduces a new analysis approach to assess quantitatively the extent to which image features predict the probability with which scene regions are selected for fixation. Specifically, we combine a scene-patch analysis with a statistical modeling approach that allows for directly describing the relationship between continuous feature values and fixation probability. The approach unites four desirable properties by explicitly accounting for (1) generic biases (e.g., central bias), (2) inter-item and inter-subject variability, (3) nonmonotonic (e.g., quadratic) effects of feature values on fixation probability, and (4) dependencies between features (see Fig.5 for a visual summary).
Affiliation: Psychology Department, School of Philosophy, Psychology and Language Sciences, University of Edinburgh, United Kingdom.