Image recognition method, equipment, storage medium and device

Abstract

The invention discloses a picture recognition method, equipment, storage medium and device. The method includes: performing feature extraction on the picture to be recognized, obtaining the feature value of each pixel in the picture to be recognized, and using a first two-dimensional matrix Characterizing the eigenvalues ​​of each pixel in the image to be identified, the first two-dimensional matrix is ​​established based on Cartesian coordinates, and determining the corresponding polar coordinates through the Cartesian coordinates of each pixel in the image to be identified , establishing a second two-dimensional matrix based on the polar coordinates of each pixel in the picture to be identified, and analyzing the second two-dimensional matrix through a preset image convolutional neural network model, so as to realize the identification of the picture to be identified identify. In the present invention, the characteristics of the polar coordinates are used to convert the change of the rotation of the picture into the change of the translation, so as to improve the ability of the graph convolutional neural network to extract the features that are invariant to the rotation of the picture.

Description

technical field [0001] The present invention relates to the technical field of picture recognition and classification, in particular to a picture recognition method, equipment, storage medium and device. Background technique [0002] Convolutional networks (CNNs) are a class of neural networks that are particularly well-suited for computer vision applications because they enable hierarchical abstraction of representations using local operations. There are two key design ideas that drive the success of convolutional architectures in computer vision. First, CNN exploits the 2D structure of images, and pixels in adjacent regions are usually highly correlated. Therefore, instead of using one-to-one connections between all pixel units (as most neural networks do), CNNs can use grouped local connections. Second, the CNN architecture relies on feature sharing, thus, each channel (i.e. output feature map) is generated by convolution with the same filter at all locations. [0003]...

AI Technical Summary

Problems solved by technology

[0005] The main purpose of the present invention is to provide a picture recognition method, equipment, storage medium and device, aiming to solve the technical problem in the prior art that the image convolutional neural network has weak ability to extract rotation-invariant features

Images