English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Bottom-up and top-down, as well as low-level and high-level factors influencewhere we fixate when viewing natural scenes. However, the importance of each ofthese factors and how they interact remains a matter of debate. Here, wedisentangle these factors by analysing their influence over time. For thispurpose we develop a saliency model which is based on the internalrepresentation of a recent early spatial vision model to measure the low-levelbottom-up factor. To measure the influence of high-level bottom-up features, weuse a recent DNN-based saliency model. To account for top-down influences, weevaluate the models on two large datasets with different tasks: first, amemorisation task and, second, a search task. Our results lend support to aseparation of visual scene exploration into three phases: The first saccade, aninitial guided exploration characterised by a gradual broadening of thefixation density, and an steady state which is reached after roughly 10fixations. Saccade target selection during the initial exploration and in thesteady state are related to similar areas of interest, which are betterpredicted when including high-level features. In the search dataset, fixationlocations are determined predominantly by top-down processes. In contrast, thefirst fixation follows a different fixation density and contains a strongcentral fixation bias. Nonetheless, first fixations are guided strongly byimage properties and as early as 200 ms after image onset, fixations are betterpredicted by high-level information. We conclude that any low-level bottom-upfactors are mainly limited to the generation of the first saccade. All saccadesare better explained when high-level features are considered, and later thishigh-level bottom-up control can be overruled by top-down influences.

Disentangling top-down vs. bottom-up and low-level vs. high-level influences on eye movements over time