Watermark recognition method and recognition system based on artificial intelligence

Abstract

The invention provides a watermark recognition method and system based on artificial intelligence, and relates to the field of document visual recognition, and the method comprises the steps: obtaining a to-be-recognized target image; performing feature extraction on the to-be-recognized target image to obtain multiple pieces of target feature information; and based on a data stream neural network classifier, identifying the multiple pieces of target feature information, and judging whether the target image to be identified is an image with a watermark, wherein the data stream neural network classifier is obtained by training a plurality of images with watermarks and a plurality of images without watermarks. A network neural model is adopted to detect a document or an image, a part of or all pages in the document are detected, each page is converted into an image (if the image format is not needed, the step is not needed) by taking the page as a unit, the image is input into a trained neural network classifier, whether the watermark is contained or not is identified, and the user can be prompted whether the watermark needs to be edited after identification is finished.

Description

technical field [0001] The invention relates to the field of document visual recognition, in particular to an artificial intelligence-based watermark recognition method and recognition system. Background technique [0002] When documents or images spread on the Internet, more and more individuals and organizations choose to embed watermark information on their images or documents due to their speed and portability. Trademarks or "draft" watermarks to protect ownership of images or documents. Therefore, before using or spreading an image or document, it is necessary to review the image or document to identify whether the image or document contains watermark information, so as to avoid misuse and infringement. With the rapid development of the Internet, image or document providers obtain a large amount of image or document information through user uploads or crawler downloads every day, and the number has far exceeded the limit of manual review. Therefore, it has become an u...

AI Technical Summary

Problems solved by technology

[0003] The visual significance of watermark information in images or documents is very low, and it has the characteristics of small area, light color, and high transparency. The difference between images or documents with watermarks and images or documents without watermarks is often small, and the degree of discrimination is low.

At this stage, there are few effective watermark recognition technologies to realize accurate recognition of watermarked images or documents

Images