An image feature segmentation method based on a graph convolutional network

Abstract

The invention discloses an image feature segmentation method based on a graph convolutional network. The method comprises steps of segmenting the preprocessed image by using a uniform grid; Constructing a directed unweighted graph taking the central image block as a vertex, and writing an adjacent matrix, a feature matrix and a degree matrix of each node corresponding to the graph by utilizing therelationship of the image blocks; Setting a weight matrix according to priori knowledge, and using a formula f (X, A) = D-1*A*X*W to carry out first-layer graph convolution on the graph; Updating thenode information by using a convolution result and taking the node information as an initial value of the next layer of convolution; And constructing a new image again, carrying out convolution, andcarrying out layer-by-layer iteration until the feature segmentation of the whole image is completed. According to the method, before a graph convolution network is made, an image is segmented by using uniform grids, the calculation amount of convolution operation is reduced to a great extent, and the accuracy of feature segmentation is improved by adopting a layer-by-layer iteration method. According to the method, image feature segmentation is carried out by using the graph convolutional network, so that the problem that the convolutional neural network cannot process irregular images is solved, the segmentation effect is greatly improved, and an optimization effect on a feature segmentation result is achieved.

Description

technical field [0001] The invention relates to the field of image feature segmentation, is applicable to the feature segmentation of regular and irregular images, and relates to an image feature segmentation method based on a graph convolution network. Background technique [0002] In image research, people are often interested in some parts of the image, and these interested parts generally correspond to specific areas with special properties in the image, which are called targets; while other parts are called the background of the image. In order to identify the target and the background, it is necessary to isolate the target from an image, which is the problem to be studied in image segmentation. Image segmentation is the first step in image analysis. The next tasks of image segmentation, such as feature extraction and target recognition, all depend on the quality of image segmentation. The current technology, in terms of feature extraction, mostly uses the method of co...

AI Technical Summary

Problems solved by technology

[0005] With the increase of image types, there are more options for image presentation shapes. People have always wanted to use computers to actively judge features in images in order to reduce human workload. In the usual convolutional neural network segmentation In the image method, due to the fixed value of the convolution kernel, data with non-European structure cannot be processed, and the accuracy of feature extraction is low. In order to solve the above problems, the present invention provides an image feature segmentation based on graph convolutional network method, the purpose is to solve the problem that the convolutional neural network cannot handle non-European structure data, and greatly improve the effect of feature extraction. It constructs a graph through the relationship between image blocks to obtain the adjacency matrix, feature matrix and degree matrix, and define the weight matrix by manually setting the weights, thereby layering the graph convolution to obtain the feature segmentation image. To achieve this purpose, the present invention provides an image feature segmentation method based on the graph convolution network, including the following steps :

Images