Abstract
Axes of bilateral symmetry in an image are a potential cue to the 3D orientation of objects. Published results are mixed with regards to our ability to detect symmetry rotated in depth. We investigated the role of depth cues on the detection of symmetry for patterns rotated in depth. METHODS: Dot patterns (24 light grey dots on a black background) 1° in diameter were presented at the center of a CRT screen. Vertical, Horizontal, 45°, and 135° orientations of symmetry were presented in separate blocks of trials. Symmetry detection was tested for 7 different rotations of the CRT screen spanning ±75° relative to the observer. Each depth rotation was tested in a separate block of trials. Subjects were tested monocularly using their dominant eye. In FULL VIEWING the room was illuminated, and the subject could see the CRT screen and all other surfaces that could provide monocular depth cues. In the NO DEPTH CUES condition, the room lights were off, and the central region of the CRT was viewed through a tube which prevented subjects from seeing anything but the dot patterns. The accuracy and reaction time for detecting symmetry and asymmetry were recorded. RESULTS: Asymmetric patterns viewed through the tube appeared to float in space at an indeterminate distance from the subject. Subjects were unable to correctly determine the screen rotation when looking through the tube. For FULL VIEWING, the speed and accuracy (>90%) of symmetry detection were similar for depth rotations between ±60°. At the ±75° rotations, symmetry detection was slower and less accurate for all symmetry orientations, with a larger effect for oblique symmetry. Testing in the NO DEPTH CUES condition extended the disruption of symmetry detection to the ±60° depth rotations. CONCLUSIONS: We found that symmetry could be detected despite large rotations in depth. Although restricting depth cues made symmetry detection more difficult, the effect was not to the degree reported in other studies.
Research support provided by UNT