Trademark labeling method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a trademark labeling method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining a plurality of preset trademark data sets, carrying out the training of a first labeling model, carrying out the pre-labeling of newly-added trademark data, and obtaining a saturation labeling result, and training the first labeling model according to a saturation labeling result to obtain a second labeling model, and finally labeling the to-be-detected image based on the second labeling model. Therefore, the labeling problem caused by inconsistent cognition of the trademarks is avoided, the cognition of the trademarks is more uniform and normative, the cognition and recognition range of the second labeling model for each trademark is expanded, the category and style covered by trademark detection are improved, and the trademark labeling range is expanded; the recall rate of the second identification model can be improved; the second recognition model is wider in cognition and recognition range and high in reusability, can be flexibly transplanted to other trademark labeling scenes, and is high in universality.

Description

technical field [0001] The present invention relates to the technical field of target detection, and in particular, to a trademark marking method, device, electronic device and storage medium. Background technique [0002] With the rapid development of artificial intelligence, various applications based on artificial intelligence technology have increasingly penetrated into various production and life, greatly effectively improving social productivity. As an important research branch of artificial intelligence, computer vision has achieved great development in recent years and has become a golden flower with relatively mature applications in the industrial field. In particular, target detection technology has gradually developed into an important part of visual tasks. [0003] At present, the accuracy of target detection technology is mainly limited by the quality of manually labeled data. How to effectively improve the quality of manual labeling is related to the accuracy i...

AI Technical Summary

Problems solved by technology

[0005] Embodiments of the present invention provide a trademark labeling method, device, electronic equipment, and storage medium to solve the problems in the prior art that the scope of trademark labeling is limited to fixed categories and the universality of the trademark labeling model is not high

Images