Using the
E-Prime 2.0 software (Psychology Software Tools, Pittsburgh, PA), the 475 chosen 3D images were randomly divided into five sections. Every picture was displayed for 4 s, and a transition picture with a fixed white cross on the center was displayed for 2 s as illustrated in
Figure 3. Users were asked to take a 10-min break between two viewing sessions to reduce fatigue. Our presentation time for every stimulus was set to 4 s, which is similar to several other experiments reported: 6 s for the NUS-3DSaliency data set (Lang et al.,
2012), 2 s for the FIFA data set (Cerf, Frady, & Koch,
2009), 4 s for the Toronto database (Bruce & Tsotsos,
2009), 3 s for the Massachusetts Institute of Technology (MIT) database (Judd et al.,
2009), and 5 s for the NUSEF database (Ramanathan, Katti, Sebe, Kankanhalli, & Chua,
2010). The only exception is the 3DGaze eye-tracking database (Wang et al.,
2013), in which the presentation time is set to 15 s, which is much longer than the others. Additionally, the literature shows that compared with the 2D experiments, the additional depth sensation leads to an increased number of fixations, shorter and faster saccades, and an increased spatial extent of exploration (Jansen et al.,
2009; Häkkinen et al.,
2010). Therefore, we adopted the 4-s presentation time, which seems to be adequate for 3D visual attention research (further discussed later).