Related Experiment Videos
Adaptive filtering in spatial vision: evidence from feature marking in plaids.
1Cognitive Science Research Centre, School of Psychology, University of Birmingham, Edgbaston, UK. m.a.georgeson@bham.ac.uk
Perception
|February 9, 2000
Summary
Early vision uses spatial filters for grouping and segmentation. Adding a third harmonic to a plaid stimulus causes perceptual segmentation, suggesting flexible filter combination in the Fourier domain.
Area of Science:
- Visual perception
- Computational neuroscience
- Image processing
Background:
- Early visual processing utilizes spatial filters tuned to frequency and orientation.
- These filters are crucial for grouping and segmenting visual information.
Purpose of the Study:
- To investigate how the human visual system segments complex stimuli.
- To determine if perceptual segmentation can be explained by static linear filters.
- To propose a model for flexible spatial filter combination in early vision.
Main Methods:
- Psychophysical experiments using a feature-marking technique.
- Stimuli included stationary plaids composed of sinusoidal gratings.
- Manipulation involved adding a third harmonic in square-wave phase to the plaid.
Main Results:
- Perceptual segmentation of a plaid into its constituent gratings was observed when a third harmonic was added.
- This segmentation could not be explained by zero crossings from static linear filters.
- Results indicate a dynamic process of combining and segmenting spatial filters.
Conclusions:
- Perceptual segmentation relies on a flexible process of combining and segmenting spatial filters in the patchwise Fourier domain.
- This process operates at the level of spatial filters before feature extraction.
- The findings challenge models relying solely on static feature primitives like zero crossings.