A multi-channel biomimetic vision method for complex scene image recognition

Abstract

The invention discloses a multi-channel bionic vision method for complex scene image recognition, the steps of which mainly include three transformation channels. For the first channel, by simulating the multi-channel visual processing mechanism of the human brain, the forward channel calculation is performed on the original image to obtain the candidate target set, and the memory information of two stages is generated at the same time. For the second channel, for each candidate target in turn, the memory information is used to perform reverse channel calculations to obtain the hit map in the original image to form the target edge map. For the third channel, the second forward channel algorithm is used to transform the target edge map twice to obtain the feature transformation map and verify it in the candidate set to complete image recognition. By simulating the multi-channel visual processing mechanism of the human brain, visual memory information is extracted, and the reverse channel calculation is used to effectively simulate the visual information processing process of the human brain on the target, and the recognition image obtained has high accuracy.

Description

technical field [0001] The invention belongs to the intersecting field of biological information and machine vision technology, and in particular relates to a multi-channel bionic vision method for complex scene image recognition. Background technique [0002] Image recognition in complex scenes is a difficult and hot issue in the field of computer vision. It is well known that the human visual system can effectively block distracting information and preserve target images in complex scenes. However, using traditional computer vision algorithms to achieve target recognition in complex scene images is a very challenging task. With the continuous revelation of the response mechanism of the human visual cerebral cortex, Hubel reported in Nature that biological visual cortical cells respond very strongly to lines of certain lengths or directions. Inspired by this biological visual response mechanism, if machine vision can extract line features of different lengths and directio...

AI Technical Summary

Problems solved by technology

However, there are two deficiencies. First, the model uses a box filter, which is implemented by using the weighted average value of pixels around the image, which is not consistent with the human visual perception mechanism. Therefore, it is particularly sensitive to noise.

Secondly, this method can only be applied to binary image recognition without background, and a little background interference will greatly reduce the recognition effect

Images