Visual orientation and recognition of an image

Abstract

An image structure provides the visual information of an image which plays an important role in visual perception. Visual Perception is considered as the end product of visual system. It can be described as the way the brain interprets what the eye sees. Human Visual System can easily extract the structure information and structure character is widely used in perception-oriented image processing tasks, such as image retrieval, face detection, texture classification, and quality assessment. The existing structure descriptor describes about the intensity character and luminance contrast of a local region. By the orientation selectivity mechanism, the correlations among pixels in a local region based on the similarities of their preferred orientation are computed. On imitating the arrangement of the excitatory/inhibitory cells, the correlations between the central pixel and its neighbors i.e local neighbors got binarized, and the orientation selectivity based pattern is used. Then, taking both the gradient magnitude and the orientation selectivity based pattern into account, a rotation invariant structure descriptor is introduced. According to the Orientation selectivity mechanism on primary visual cortex,a novel structure descriptor is introduced which represents the spatial correlation of an image which is applied in texture classification and image quality assessment. With the novel structure descriptor, image structure was extracted,an orientation selectivity based structural histogram and recognition of an image was derived to represent the image content.