Related Experiment Videos
The representation of uniform motion in vision
M T Swanston1, N J Wade, R H Day
1Dundee College of Technology, UK.
Perception
|January 1, 1987
Summary
This study introduces a four-level model for how moving systems perceive object motion, from eye-based (retinocentric) to world-based (geocentric) representations. This framework helps explain various motion perception phenomena.
Area of Science:
- Perception and Cognition
- Computational Neuroscience
- Robotics
Background:
- Accurate object motion detection requires systems to differentiate self-motion from object-motion.
- Existing models often lack a hierarchical framework for motion allocation.
Purpose of the Study:
- To develop and present a novel four-level model for allocating motion between a detector and objects in space.
- To explain veridical and non-veridical motion perception phenomena using this model.
Main Methods:
- Development of a computational model for motion allocation in simplified scenarios (e.g., points of light).
- Proposal of four successive representational levels: retinocentric, orbitocentric, egocentric, and geocentric.
- Analysis of how angular and 3D motion signals are integrated across levels.
Main Results:
- The model successfully integrates retinal signals with eye, head, and body movements.
- It distinguishes between angular (retinocentric to egocentric) and 3D (geocentric) motion representations.
- The framework accounts for various observed motion perception phenomena.
Conclusions:
- Motion perception is fundamentally geocentric but relies on integrated signals from earlier representational levels.
- The proposed hierarchical model provides a unified explanation for diverse motion detection and perception processes.