Memorability of line drawings of scenes: the role of contour properties

Seohee Han,Morteza Rezanejad,Dirk B. Walther
DOI: https://doi.org/10.3758/s13421-023-01478-4
2023-11-01
Memory & Cognition
Abstract:Why are some images more likely to be remembered than others? Previous work focused on the influence of global, low-level visual features as well as image content on memorability. To better understand the role of local, shape-based contours, we here investigate the memorability of photographs and line drawings of scenes. We find that the memorability of photographs and line drawings of the same scenes is correlated. We quantitatively measure the role of contour properties and their spatial relationships for scene memorability using a Random Forest analysis. To determine whether this relationship is merely correlational or if manipulating these contour properties causes images to be remembered better or worse, we split each line drawing into two half-images, one with high and the other with low predicted memorability according to the trained Random Forest model. In a new memorability experiment, we find that the half-images predicted to be more memorable were indeed remembered better, confirming a causal role of shape-based contour features, and, in particular, T junctions in scene memorability. We performed a categorization experiment on half-images to test for differential access to scene content. We found that half-images predicted to be more memorable were categorized more accurately. However, categorization accuracy for individual images was not correlated with their memorability. These results demonstrate that we can measure the contributions of individual contour properties to scene memorability and verify their causal involvement with targeted image manipulations, thereby bridging the gap between low-level features and scene semantics in our understanding of memorability.
psychology, experimental
What problem does this paper attempt to address?