Equiluminant Border Ownership Cells as a Missing Link in Color Form Perception
Mehrani, P.; Tsotsos, J. K.
Show abstract
We propose the existence of a sub-population of border ownership neurons that signal figure side along equiluminant borders, those defined purely by color contrast. We postulate that these neurons, similar to primate equiluminant V1 [26] and V4 [5] cells, are strongly activated by equiluminant contrast on their preferred side, with diminished response for strictly luminance contrast. To test this hypothesis, we propose a hierarchical and mechanistic model that explains how equiluminant border ownership signals can be achieved in the ventral stream. Our simulation results suggest that model equiluminant neurons are highly color-selective while exhibiting similar orientation and border ownership selectivity as their luminance counterparts. Additionally, our model equiluminant cells indeed exhibited the above postulated pattern of responses. We suggest that these neurons are the missing links in the equiluminance channels in the ventral stream that transform equiluminant oriented edge signals in V1 to equiluminant object-centered shape representations in V4. Our findings point to the importance of re-examining the border ownership neuron population to further detail their role in equiluminant channels in the ventral stream.
Matching journals
The top 3 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Predictive coding as a unifying principle for explaining a broad range of brightness phenomena 96%
- Asymmetries around the visual field: From retina to cortex to behavior 96%
- Increasing neural network robustness improves match to macaque V1 eigenspectrum, spatial frequency preference and predictivity 96%
Similar papers in this journal
Similar papers in this journal
- Predictive coding with spiking neurons and feedforward gist signalling 96%
- Dynamic Control of Eye-Head Gaze Shifts by a Spiking Neural Network Model of the Superior Colliculus 94%
- Leveraging prior concept learning improves ability to generalize from few examples in computational models of human object recognition 94%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.