A method and device for extracting image features

Abstract

An embodiment of the present invention provides a method for extracting image features. The method includes: calculating HOG feature information corresponding to each block structure in the target image, the block structure including a pixel matrix composed of a preset number of pixel values; The block structure, according to the pixel matrix in the block structure, generates the input set of the PCA algorithm; uses the PCA algorithm to calculate the PCA feature information corresponding to each block structure; cascades the HOG feature information corresponding to each block structure and the block structure. PCA feature information to generate new HOGP feature information corresponding to each block structure; concatenate new HOGP feature information corresponding to all block structures to generate new HOGP feature information of the target image. By applying the embodiments of the present invention, rich image features including detail features and contour features can be extracted.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a method and device for extracting image features. Background technique [0002] With the rapid development of image recognition technology, the accuracy of image recognition in various application fields is getting higher and higher. In order to improve the accuracy of machine recognition of images, people use image feature visualization methods to improve the image recognition algorithms used by machines in image recognition. [0003] Image feature visualization specifically refers to restoring the target image according to the image features of the target image extracted by the machine. In this way, the image recognition algorithm can be improved according to the difference between the original target image and the restored target image. If the difference between the original target image and the restored target image is small and the similarity is high, it means tha...

AI Technical Summary

Problems solved by technology

[0004] At present, in the process of image feature visualization, image recognition algorithms for extracting image features include principal component analysis (Principal Component Analysis, PCA) algorithm and feature detection (Histogram of OrientedGridients, HOG) algorithm; among them, PCA algorithm is good at extracting image However, for images with less detailed features, such as sky, sea, wall, etc., the image information contained in the extracted PCA feature information is incomplete; while the HOG algorithm is good at extracting the contour features of the image, but the extracted HOG feature information ignores the detailed information of the image

In this way, according to the image features extracted using the PCA algorithm or the HOG algorithm, the restored image has a large difference and a low similarity with the original target image

Images