Contour detection based on the property of orientation selective inhibition of non-classical receptive field

The majority of neurons in primary visual cortex (V1) of brain are orientation-selective. Both the classical receptive field (CRF) and the non-classical receptive field (NCRF), which modulates the CRF and mainly yields inhibition, could be orientation-selective and may obtain different tune. For a single neuron, the response is determined by the interaction of its CRF and NCRF. And the horizontal connections play an important role when forming inhibition. Inspired by those visual cortical mechanisms, a modified inhibition model, called orientation selectivity of NCRF, is introduced to improve the performance of contour detectors. The orientation saliency is determined by the energy, the output of a Gabor Energy filter, in each direction. And the inhibition term is taken based on the saliency of the orientation. This proposed method could selectively retain the object contours and suppress texture edges effectively, which is demonstrated by the processing of several natural images.