Projective invariance and the kinetic depth effect

Area of Science:

• Visual Perception
• Cognitive Psychology
• Computational Neuroscience

Background:

• The kinetic depth effect (KDE) relies on the assumption of direct perceptual access to shape equivalence in projective geometry.
• This assumption is fundamental to 'inverse optics' theories of visual form perception.
• Understanding how observers perceive 3D shape from motion is crucial for modeling visual processing.

Purpose of the Study:

• To empirically test the assumption of direct access to projective shape equivalence in the KDE.
• To investigate how rotation in depth affects the perception of shape equivalence.
• To evaluate the robustness of projective geometry assumptions in visual form perception models.

Main Methods:

• Seven experiments were conducted using computer graphics to present rotating 2D quadrilaterals and 3D prisms.
• Observers adjusted a comparison shape to match a standard shape under varying conditions of rotation and texture.
• Projective equivalence was quantified using planar and spatial cross-ratio measurements.

Main Results:

• Perception of projective equivalence was not uniform, particularly under rotation in depth.
• Orientation changes up to 180 degrees around a single axis had no significant effect on shape matching.
• Changes in orientation around multiple axes significantly altered shape perception, contradicting uniform projective equivalence.
• Adding texture or altering rotation speed did not rectify these departures from projective equivalence.

Conclusions:

• The assumption of direct and reliable access to projective shape equivalence is not consistently supported by empirical evidence in the KDE.
• Visual form perception is sensitive to multi-axis rotations, suggesting limitations in current inverse optics models.
• Future models of visual form perception need to account for the complex interplay between motion, orientation, and perceived shape.