Image edge grading-detection method based on visual pathway orientation sensitivity

Abstract

The invention discloses an image edge grading-detection method based on visual pathway orientation sensitivity. A first neural network sensitive to multiple specific directions is established under the action of neuron synaptic connection on receptive field optimal orientation centripetal distribution, an image pixel is taken as network input, a pulse firing sequence of neuron in a certain time window is recorded, and a discharge frequency is calculated to serve as network output; network output in multiple directions is integrated and mapped to a gray scale, and an edge sensitive image is formed; as for the edge sensitive image, the inner lateral inhibition range and the inhibition quantity of a receptive field are determined, a second neural network is formed, and the laterally inhibited image is output; finally, an edge detection result is acquired after threshold processing. According to the method, important vision mechanisms such as directional receptive field, lateral inhibition and the like are considered, the grading processing effect of different levels of structures on a visual pathway during profile sensing is simulated, and the edge detection performance of a long-scale contrast image is improved.

Description

technical field [0001] The invention belongs to the field of visual neural computing, and relates to an image edge classification detection method based on visual path orientation sensitivity. Background technique [0002] Contour feature extraction will provide important dimensionality reduction information for image understanding or moving target behavior analysis. The extraction process usually needs to meet: (1) under the premise of accurately locating edges, no missed detection occurs; (2) avoid false edges. However, in practical applications, due to the influence of unfavorable factors such as lighting and shadows, the contrast of image edges is reduced, and it is difficult for traditional detection methods to meet the above two requirements at the same time; and the current edge detection method based on the visual neural mechanism simplifies the process of real neurons. The electrophysiological characteristics in signal processing ignore the hierarchical processing e...

AI Technical Summary

Problems solved by technology

However, in practical applications, due to the influence of unfavorable factors such as lighting and shadows, the contrast of image edges is reduced, and it is difficult for traditional detection methods to meet the above two requirements at the same time; and the current edge detection method based on the visual neural mechanism simplifies the process of real neurons. The electrophysiological characteristics in signal processing ignore the hierarchical processing effect of different hierarchical structures on the visual pathway in contour perception, and essentially use a black-box mathematical model to simulate the visual mechanism

Images