. Author manuscript; available in PMC: 2021 Jun 22.

Published in final edited form as: Proc IEEE Comput Soc Conf Comput Vis Pattern Recognit. 2020 Aug 5;2020:190–199. doi: 10.1109/cvpr42600.2020.00027

Table 1: Comparison of fixation datasets.

Previous datasets either did not use a search task, or had very few target-object classes, subjects, or fixations.

Dataset	Search	Image	Class	Subj/img	Fixation
SALICON [27]	✗	10000	-	60	4600K^*
POET [42]	✗	6270	10	5	178K
People900 [15]	✓	912	1	14	55K
MCS [56]	✓	2183	2	1–4	16K
PET [18]	✓	4135	6	4	30K
COCO-Search18	✓	6202	18	10	300K

^*:

Fixations are approximated by mouse clicks.