Image matching method and device

Abstract

The embodiment of the invention provides an image matching method and device, and relates to the field of computer vision, and the method comprises the steps: determining image feature points of two to-be-matched images, carrying out the coordinate clustering according to the image feature points, and dividing the to-be-matched images into a plurality of image regions according to a coordinate clustering result; carrying out importance degree sorting on the plurality of image areas of each to-be-matched image, carrying out feature comparison on the two image areas of the two to-be-matched images in sequence according to the importance degree sequence, and determining the similarity of the two image areas; determining whether the two to-be-matched images are matched or not according to the similarity of the two image areas; according to the method, the image retrieval precision and the matching efficiency can be effectively improved.

Description

technical field [0001] The present application relates to the field of computer vision and can also be used in the field of finance, in particular to an image matching method and apparatus. Background technique [0002] With the popularization of smart phones and digital products, users begin to use digital products to take pictures and share them on some Internet social networking sites with some geographic tags, which makes the Internet's image data information explode. At the same time, the demand for image retrieval is also growing. [0003] Image retrieval refers to the key information of the input image, so as to retrieve related images on the Internet. The key information here includes the label of the image or the image itself. For users, the most convenient way is to upload the image directly to the Internet to query the relevant images of the local image. For example, if a user wants to know about a building, he can directly The building is photographed, and then...

AI Technical Summary

Problems solved by technology

The key information here includes the label of the image or the image itself. For users, the most convenient thing is to directly upload the image to the Internet to query related images of the local image. For example, if the user wants to know about a building, he can directly Take a picture of this building, and then compare it with the large database of the Internet by extracting image features, and finally get the information that the user wants. However, as the user's demand for image retrieval accuracy and speed continues to rise, some traditional image retrieval methods Can no longer meet the needs of users

Images