Visual Psychophysics of Egomotion.

Abstract

Two psychophysical studies investigated an observer's ability to perceive self motion. In both studies, the stimuli were computer-generated images simulating an observer moving relative to a volume of randomly-positioned dots. The first study investigated an observer's ability to determine whether s/ he was moving forward along a straight or curved path as forward speed was varied. The results showed that with eye movements, the deviation required to detect a departure from a straight path increased with forward speed. When eye movements were restricted, the required deviation remained constant across forward speeds. A second study investigated the effectiveness of various sizes and retinal locations of the stimulus in determining the direction of a curved path. The results showed an increasing linear relationship between optimal size and retinal eccentricity. Given optimally-scaled stimuli, the central and peripheral retinal locations yielded equivalent performance. Finally, a computational model has been developed to emulate a human observer's ability to detect a curved path of motion. Computer simulations of the model have been run on a task to discriminate between a curved and straight path of motion. The simulation results closely match psychophysical data. Egomotion, Motion perception, Curvilinear motion, Self motion perception, Eccentricity, Eye movements.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Feb 24, 1992
Accession Number
ADA248349

Entities

People

  • Kathleen Turano

Organizations

  • Purdue University

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Abstracts
  • Computational Science
  • Computer Programs
  • Computer Simulations
  • Computers
  • Eccentricity
  • Eye
  • Eye Movements
  • Information Systems
  • Neural Networks
  • Observers
  • Perception
  • Simulations
  • Stratified Fluids
  • Two Dimensional
  • Visual Perception

Fields of Study

  • Psychology

Readers

  • Computer Vision.
  • Human-Computer Interaction (HCI).
  • Robotics and Automation.