Image recognition method based on scale co-occurrence local binary pattern

Abstract

The invention discloses an image recognition method based on a scale co-occurrence local binary pattern, and relates to the technical fields of digital image processing, computer vision and the like. The method comprises the following specific steps: 1) constructing an LBP symbiotic space of an image through simulation scale transformation, and extracting stable structure information in the LBP symbiotic space; and 2) representing stable structure information by using a cross-scale symbiotic pair in an LBP symbiotic space. The cross-scale symbiosis pair can keep good description capability even under the condition that the image is scaled; and 3) proposing a rotation consistency adjustment strategy, and uniformly adjusting the cross-scale symbiotic pairs to enhance the rotation invariance of the cross-scale symbiotic pairs. According to the method, a cross-scale symbiotic pair and a rotation consistency adjustment strategy are fully utilized, so that the method has ideal geometric invariance, and an image description capability which is the same as that of a symbiotic-based LBP method is obtained under the condition that the feature dimension is relatively low.

Description

technical field [0001] The invention relates to an image recognition method based on a scale co-occurrence local binary pattern, belonging to the technical fields of digital image processing, computer vision and the like. Background technique [0002] Image classification is one of the most fundamental tasks in computer vision. It is an image processing method that uses the intrinsic characteristics reflected by the image to classify different images into corresponding categories. Since an image contains a lot of information, it is difficult to easily extract useful information from the image. Therefore, the extraction of image visual features has become an increasingly important research topic. According to the different properties of images, image feature extraction can be divided into five categories: texture feature extraction, color feature extraction, shape feature extraction and spatial relationship feature extraction. Compared with other image features, texture fe...

AI Technical Summary

Problems solved by technology

However, the existing LBP-based features still have the following problems: (1) In the face of scale transformation, most features sacrifice the feature dimension to improve the scale robustness of the feature, resulting in a multiplied feature dimension; (2) For images that have undergone rotation transformation, the feature's ability to describe the image will be greatly reduced, and it does not have good rotation robustness

Images