Using Gaze-contingent Depth of Field to Facilitate Depth Perception

The lens system of the human eye has a limited depth of field that creates blur patterns that depend on its focus and the relative distances of the observed objects. Using an eye tracker it is possible to simulate this effect in virtual scenes by generating appropriate blur based on the gaze position and scene geometry. We implemented gaze-contingent depth of field (GCDOF) and conducted two experiments to investigate how it affects perceived realism and the perception of depth (Mauderer et al., 2014). In experiment (1) we asked participants to assess their impression of realism and depth watching gaze-contingent and conventional presentation of images using a questionnaire. With experiment (2) we tried to assess the accuracy of depth perception by showing pairs of stimuli to the participants and asking for depth judgements using manual input. The results from (1) show an increase in reported realism and depth for the GCDOF condition. Experiment (2) showed a small increase in ordinal depth discrimination and an increase in perceived depth for the gaze-contingent condition. The findings suggest that GCDOF can be used to increases subjective perceived realism and depth and contains information about depth order and relative spatial position. These findings agree with previous investigations of GCDOF (e.g., Hillaire et al. 2008; Mantiuk, et al. 2011) and might give additional insight about the general role of DOF blur in perception.