Related Experiment Videos
Motion contrast: a new metric for direction discrimination
M A Georgeson1, N E Scott-Samuel
1School of Psychology, University of Birmingham, UK. m.a.georgeson@bham.ac.uk
Vision Research
|May 2, 2000
Summary
A new motion contrast metric accurately predicts human direction discrimination performance, outperforming the standard opponent energy model. This finding suggests a refined model for visual motion perception.
Area of Science:
- Visual neuroscience
- Computational vision
- Perception psychology
Background:
- The Adelson-Bergen energy model is a standard framework for first-order motion processing.
- Opponent energy (EL - ER) indicates motion direction but poorly predicts performance.
- Divisive inhibition in striate cells may require contrast gain control (Heeger, 1992).
Purpose of the Study:
- To evaluate the predictive power of opponent energy for direction discrimination.
- To introduce and validate a new motion perception metric: motion contrast.
- To test if a modified opponent energy model can predict observed performance.
Main Methods:
- Human observers viewed dynamic grating sequences with varying contrast ratios.
- Direction discrimination performance was measured.
- Opponent energy and a new motion contrast metric were calculated and compared to performance.
Main Results:
- Opponent energy was a poor predictor of direction discrimination.
- Motion contrast (EL - ER)/(EL + ER) strongly predicted performance across contrast levels.
- Discrimination thresholds were approximately 0.5 in motion contrast units.
Conclusions:
- Motion contrast is a superior metric for predicting direction discrimination compared to opponent energy.
- A modified opponent energy model incorporating response variance and functions of cortical cells can predict motion contrast dependence and thresholds.