Shape representing and matching method for trademark image retrieval

Abstract

The invention discloses a shape representing and matching method for trademark image retrieval and relates to image search. The method comprises characteristic representation of trademark object shape and match of extracted characteristic, specifically comprising the following steps: 1) normalizing an object; 2) calculating the minimum circumscribed circle of the target object; 3) concentrically dividing a target region in equal area; 4) dividing the concentric circular region to blocks; 5) calculating and extracting characteristic; 6) generating corresponding binary shape characteristic string for each trademark object in a trademark library to match; and 7) judging the matching result of the binary shape characteristic string is a similarity numerical value in an interval of [0, 1]. The method can be directly applied to trademark image and further can be applied after dividing the trademark image to obtain the target object, so that the image is precisely matched in the object level, thereby effectively solving the problems of invariance of translation, rotation and rotation transformation of the target object in the shape matching process, more quickly and accurately searching for the target picture and reducing the trademark searching time and efficiently outputting the required target picture.

Description

technical field [0001] The invention relates to image search, in particular to a shape representation and matching method for trademark image retrieval. Background technique [0002] A trademark is one of the important signs of a commodity. The registration of a trademark is very important to protect the legal rights of the trademark holder. Therefore, before a new trademark is registered, it must be searched in the trademark database to ensure that it is clearly different from other registered trademarks. . [0003] The management of word mark is relatively easy, and it can be searched by means of text annotation. As for image trademarks or combined trademarks containing images, it is not easy to describe through text. If the method of classification and coding is adopted, there will be disadvantages such as time-consuming classification and incomplete description. Therefore, content-based image retrieval technology is gradually being favored by people. Pay attention to. ...

AI Technical Summary

Problems solved by technology

Existing retrieval methods still have deficiencies in terms of scaling and rotation invariance, retrieval ability for geometric deformation, retrieval accuracy, and the consistency between images and human visual perception. It is necessary to continue to study more effective retrieval methods and integrate multiple algorithms. Meet search needs

Images