EyeCatchingMaps, a Dataset to Assess Saliency Models on Maps

EyeCatchingMaps, a Dataset to Assess Saliency Models on Maps

This dataset is made up of a total of 322 maps. It is made up of maps from google maps, ign and osm.

It also includes maps from school textbooks.

In the folder map we have all the map of the datatset

in the folder coord_on_map, for each map, a file containing all the fixation points of each participant, named ’coord_fixation_name_map.png.csv’ in CSV format with the following elements and during a fixation, a person looks at an area and can move slightly. The fixation will then be represented by a few points:

- worl_index : the id of fixation

- id_fixation : the id of fixation point

- time : the time of the fixation point in relation to the time of the survey

- x : x coordinate of the point of fixation on the image

- y : y coordinate of the point of fixation on the image

- dispersion : distance between all gaze locations during a fixation, in degree

- accuracy : in degree

- precision :in degree

- participant : Participant ID

- distance : distance between eye-tracker and screen incm

- time_to_map : the time of the fixation point in relation to the map display

- height : height of the map

- width : width of the map

In the folder heatmap we have heatmaps for each map named ’heatmap_name_map.png’.

in the  folder 'script' there are the different scripts used to produce the data