A local image feature extraction method, device, system and readable storage medium

Abstract

The application discloses a method for extracting local image features, including: modifying the convolution parameters in the convolutional neural network to obtain the modified convolution parameters; obtaining a local image; Extract the feature information of the local image by parameters to obtain the feature vector; wherein, the correction method of the convolution parameter includes: inputting the obtained local image information including different categories into the convolutional neural network, and obtaining the feature vector corresponding to each local image information; calculating The feature center point corresponding to the local image information of each category; calculate the similarity between each category and the local image in the same category according to the feature center point; modify the convolution parameters according to the similarity. The extraction method has strong local feature extraction ability and good detail feature extraction effect. The present application also discloses a local image feature extraction device, system and computer-readable storage medium, which have the above beneficial effects.

Description

technical field [0001] The present application relates to the field of domain self-adaptation, in particular to a local image feature extraction method, device, system and a computer-readable storage medium. Background technique [0002] Local image features are the local expression of image features, which reflect the local characteristics of the image. Compared with global image features such as line features, texture features, and structural features, local image features have a rich number of implication in the image, and the correlation between features The advantages of low degree of occlusion and the disappearance of some features will not affect the detection and matching of other features have been widely used in face recognition, 3D reconstruction, target recognition and tracking, film and television production, panoramic image stitching and other fields. The extraction of local image features is usually the first step in many problems in computer vision and digita...

AI Technical Summary

Problems solved by technology

[0003] However, images often undergo changes in scale, translation, rotation, illumination, perspective, and blur. Especially in practical application scenarios, images will inevitably have large noise interference, complex backgrounds, and large target pose changes. In this unfavorable situation, the extraction effect of local features is often unsatisfactory.

Images