Related Experiment Videos
Directional filling-in
1Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, Cambridge, MA 02139, USA.
Neural Computation
|February 15, 1996
Summary
This study advances the filling-in theory of brightness perception, improving how vision models handle brightness relations. The enhanced theory now successfully explains all Arend
Area of Science:
- Neuroscience
- Computational Vision
- Psychophysics
Background:
- The filling-in theory of brightness perception is a prominent framework in vision science.
- Existing models based on this theory struggle with transitive brightness relations.
Purpose of the Study:
- To address the limitations of the current filling-in theory regarding transitive brightness.
- To present an advanced version of the filling-in theory.
Main Methods:
- Incorporation of the advanced filling-in theory into the BCS/FCS neural network model.
- Testing the model's ability to account for Arend's comprehensive set of brightness perception stimuli.
Main Results:
- The revised BCS/FCS model successfully accounts for all of Arend's test stimuli.
- This represents a significant improvement over previous instantiations of the filling-in theory.
Conclusions:
- The enhanced filling-in theory provides a more robust explanation for brightness perception.
- The revised model offers a potential new teleology for parallel ON- and OFF-channels in visual processing.