Abstract
Unlike conjugate eye movement, pursuing a moving object in a 3D space (i.e., disconugate eye movement) might produce eye movement-induced disparity for any dynamic object that is present around a pursuit object. Li (2004) showed that the perceived slant of the target surface defined by the sequential display of partial contour information is systematically distorted during disconjugate eye movement, suggesting that eye movement-induced disparity is not compensated for with eye position information. The target surface employed in the research of Li does not exist in real life, and this can explain why we usually do not experience perceptual distortion of 3D slant during disconjugate eye movement in a natural environment. To examine this possibility, we simulated a series of stereoscopic movies containing a luminance-defined object that is moving vertically in a frontoparallel plane or in a 3D space and asked subjects to estimate the 3D slant of the path of the moving object. It should be noted that we usually encounter this type of stimulus in a real life environment. When the eye movement of the observers was not disconjugate, their estimates of the 3D path of the moving target were nearly correct. On the contrary, when their eye movement was disconjugate, their perception of the 3D path was systematically distorted. This finding is in complete agreement with Li's finding and implies that eye movement-induced disparity is not compensated for with eye position information.
Supported by the Realistic 3D-IT Research Program of Kwangwoon University under the National Fund from the Ministry of Education and Human Resources Development (2005)