A Coutour Detection Model Based on Surround Inhibition with Multiple Cues
Sufficient physiological studies have revealed that surround inhibition substantially occurs when difference exists between the classical receptive field (CRF) and its surrounding (i.e. non-CRF) of most neurons in primary visual cortex (V1) for any local visual features. In this paper, we propose an improved contour detection model based on the biologically-plausible computational steps with non-CRF inhibition (also called surround inhibition) in V1. Through principal component analysis (PCA) we combine multiple local cues, including orientation, luminance and contrast, to improve contour detection in natural images. The results on a commonly used large image dataset demonstrate that surround inhibition combining multiple local cues can remarkably improve contour detection in complex scenes.
Keywordscontour detection surround inhibition receptive filed multiple local cues V1
Unable to display preview. Download preview PDF.
- 1.Hubel, D.H., Wiesel, T.N.: Receptive fields, binocular interaction and functional architecture in the cat’s visual cortex. The Journal of Physiology 160(1), 106–154 (1962)Google Scholar
- 5.Jones, H., Grieve, K., Wang, W., Sillito, A.: Surround suppression in primate V1. Journal of Neurophysiology 86(4), 2011–2028 (2001)Google Scholar
- 8.Canny, J.: A computational approach to edge detection. Readings in Computer Vision: Issues, Problems, Principles, and Paradigms 184, 87–116 (1986)Google Scholar