A method for contour detection base on that synergistic effect of multi-level inhibition regions of visual path

Abstract

The invention relates to a contour detection method based on the synergistic effect of multi-level inhibition regions of a visual path. The invention firstly constructs a direction-sensitive lateral geniculate body neuron array, records the contour direction selected by the lateral geniculate body, and calculates the response of the lateral geniculate body neurons. The visual input of the classical receptive field and inhibitory area of the external geniculate neurons is quantified with distance as a factor, and the difference between them is regarded as the effective response of the inhibitory area, and the parameters of synergistic action were obtained by dynamic rectification and exponential normalization. The visual input of primary visual cortex is obtained by fusing the response of lateral geniculate body neurons with local window weighting, and the precise contour response is detected. The response of primary visual cortex neuron inhibitory region is calculated and the final contour response is obtained. The invention considers the gradual refinement detection mechanism when the neurons of different levels perceive the direction in the visual path, simulates the synergisticeffect of the inhibition areas of different levels, and can effectively improve the contour detection performance of the natural image.

Description

technical field [0001] The invention belongs to the field of machine vision, and mainly relates to a contour detection method based on the synergistic effect of multi-level inhibition zones of visual pathways. Background technique [0002] Contour detection is one of the important links in the early stage of image understanding or visual analysis. The obtained contour features will effectively express the key visual details after de-redundancy. The difficulty of the contour detection task is mainly manifested in two aspects: over-detection and under-detection. The former is due to the interference of false contours such as texture noise, and the latter is due to the difference in the contour contrast distribution of images. The traditional contour detection method is mainly based on the spatial jump of image information, so mathematical means based on differential or morphological operations are adopted, which can usually achieve good performance when the image quality is go...

AI Technical Summary

Problems solved by technology

However, the traditional method completely ignores the biological neural mechanism in visual perception, so it is difficult to meet the detection performance requirements when faced with complex tasks such as multi-level contour detection.

Although the current contour detection methods based on biological vision mechanism simulate the ability of biological vision to extract image contour features to a certain extent, they pay more attention to a certain level of neurons in the visual pathway when simulating the visual information flow processing process of the visual pathway. The classic receptive field or inhibitory area itself, but ignores the synergistic effect of inhibitory areas between different levels in the visual pathway. This synergistic effect changes the isolation of each level of the visual perception model, and can fully utilize the characteristics of overall synergy. Exploiting the important role played by each level in perception

Images