Abstract
To investigate whether vanishing point (VP) plays a role in gaze guidance during visual search, we asked 14 subjects (4 female, mean age=23.07, SD=1.26) to search for a target character (T or L) placed randomly on a 3x3 imaginary grid within an image (dva 37.6x24). Subjects reported their answers by pressing one of two arrow keys (left arrow for T). The scene remained visible until keypress for 10 seconds maximum. We measured subjects' reaction times (RT) and accuracies. Stimuli consisted of 270 color images (180 with a single VP, 90 without). The target happened with equal probability inside each cell (15 times L, 15 times T). Analyzing images with VP, we found that all subjects were significantly faster when the target happened inside the cell containing the VP, compared to cells without VP (median across 14 subjects 1.34 secs vs. 1.96; Wilcoxon rank-sum test; p=0.0014). Reaction time at VP cells was significantly lower than RT on images without VP (median 2.37; p=4.77e-05). Reaction time at off-VP cells (over images with VP) was lower than RT on images without VP (p=0.0072). This result was unexpected since we anticipated that VP would act as a distractor hence raising RT over off-VP cells. We attribute the higher RT for off-VP cells to lower complexity of scenes with VP as those scenes usually had relatively less clutter (Further, images without VP contained more face and text). Median (and mean) accuracies over subjects were above 95%. Accuracies for the target at VP cells were significantly higher than off-VP cells (medians 100% vs. 97%; p=0.02) and images without VP (median 95%; p=0.01). These findings support our previous results (Borji et al., arXiv 2015) that vanishing point, similar to face and text (Cerf et al., JOV 2009), attracts attention and gaze in free-viewing and visual search tasks.
Meeting abstract presented at VSS 2016