Can we accurately predict where we look at paintings?

Abstract

The objective of this study is to investigate and to simulate the gaze deployment of observers on paintings. For that purpose, we built a large eye tracking dataset composed of 150 paintings belonging to 5 art movements. We observed that the gaze deployment over the proposed paintings was very similar to the gaze deployment over natural scenes. Therefore, we evaluate existing saliency models and propose a new one which significantly outperforms the most recent deep-based saliency models. Thanks to this new saliency model, we can predict very accurately what are the salient areas of a painting. This opens new avenues for many image-based applications such as animation of paintings or transformation of a still painting into a video clip.

How do people distribute their attention while observing The Night Watch?

“…The heatmaps and corresponding analyses of AOIs showed that bright and textured elements in The Night Watch attract the observer's eyes. These findings are in line with modern research into scene exploration and painting viewing, which has shown that salient features (Le Meur et al, 2020;Quian Quiroga & Pedreira, 2011;Xu et al, 2010) including faces (Bindemann et al, 2005;Massaro et al, 2012;Theeuwes & Van der Stigchel, 2006) attract attention. What is worth noting is that Rembrandt mastered these techniques and managed to bring out the prominent figures (captain and lieutenant, and the girl, the company mascot) long before the onset of eye-tracking research or experimental psychology as a discipline.…”

Exploring Visitors’ Visual Behavior Using Eye-Tracking: The Case of the “Studiolo Del Duca”

“…According to literature, without a specific interest or task at hand, visitors' attention is stimulus-driven, attracted by low-level visual features such as an intense colour or a high contrast [10,11,14,15]. Some portraits of the illustrious men attracted first the visual attention of participants with respect to the lower part of the studiolo, where the wooden panels are substantially "homogeneous" in colour.…”

“…Data collected with the mobile eye-tracker were aggregated and mapped onto static images of the particular stimuli encountered during the recording and relevant metrics were measured by Imotions ® Attention Tool ® . A synopsis of eye-tracking metrics relevant for analysing artworks is reported in Table 3 [10,19,20]. Heatmaps and gaze plots are data visualisations that allow for identifying relevant aspects of a subject's visual behaviour.…”

“…In the context of art, the use of eye-tracking is not new [1,[8][9][10][11][12][13][14][15][16][17][18][19][20]. Quiroga et al [13] studied participants' differences in visual behaviour by comparing data collected viewing an original painting in the Tate Britain gallery and the digital image of the same artwork.…”

“…according to the instructions provided and to the context [14,15,19,21]. Studies showed that images physical features, such as colour, luminance, or orientation, define visual saliency, which is a physical property of a visual item that makes it stand out from other items and immediately grab subjects' attention [1,10,11,14,15,22]. Moreover, salient regions that are determined using specific algorithms [23,24] are inspected first by participants [14,25].…”