Related Experiment Videos
Pooling of vertical disparities by the human visual system
W Adams1, J P Frisby, D Buckley
1Al Vision Research Unit, University of Sheffield, UK.
Perception
|January 1, 1996
Summary
Scaling vertical disparities affects perceived depth of 3D surfaces. Observed depth changes were smaller than predicted, indicating oculomotor cues influence perception, especially when disparity information is mixed.
Area of Science:
- * Visual perception
- * Stereopsis
- * Computational vision
Background:
- * The perception of depth from stereoscopic images relies on binocular disparity.
- * Theories by Mayhew & Longuet-Higgins and Garding et al. predict how scaling vertical disparities should alter perceived depth based on simulated viewing distances.
- * Understanding these scaling effects is crucial for accurate 3D scene reconstruction.
Purpose of the Study:
- * To experimentally investigate the impact of scaling vertical disparities on the perceived amplitude of dome-shaped surfaces.
- * To compare observed depth scaling effects with predictions from existing computational models.
- * To explore how mixed disparity scaling conditions influence perceived depth.
Main Methods:
- * Two experiments were conducted using dome-shaped surfaces rendered with horizontal disparities.
- * A nulling task was employed, where observers adjusted horizontal disparities to perceive the surface as flat, effectively canceling scaling effects.
- * Conditions included uniform vertical disparity scaling and mixed scaling (50% scaled, 50% unscaled).
Main Results:
- * Significant changes in perceived depth were observed due to vertical disparity scaling.
- * The magnitude of these scaling effects was consistently smaller than predicted by current theories.
- * In mixed scaling conditions, perceived depth was influenced by equally weighted pooling of disparity information, but the lower scaling factor dominated when dissimilar scalings were presented.
Conclusions:
- * Oculomotor cues likely play a significant role in determining perceived fixation distance, moderating the effects of scaled vertical disparities.
- * The findings suggest a more complex pooling process for vertical disparity information than previously modeled, potentially involving Hough parameter estimation.
- * Current models may need refinement to fully account for the interplay of different visual cues in stereoscopic depth perception.