Content identification method and device, electronic equipment and readable storage medium

Abstract

The embodiment of the invention provides a content identification method and device, electronic equipment and a readable storage medium. The invention relates to the field of content identification, and the method comprises the following steps: obtaining source text data, source image data and video data in webpage information, converting the video data into corresponding image data, and identifying and labeling the source text data, the source image data and the image data converted from the video data through a first learning model and a second learning model; therefore, characters, images and video data in the webpage can be identified at the same time. The requirement for identifying the multi-modal webpage content is met, and meanwhile, the identification efficiency can be greatly improved.

Description

Technical field [0001] The present invention relates to the field of content identification, and in particular, to one content identification method, apparatus, electronic device, and readable storage medium. Background technique [0002] With the development of Internet technology, the Internet platform has become the main platform for people's information. On various Internet platforms, massive texts, images and video data are generated, while in these massive data, very May cause unsafe content that hazards network health. [0003] For unsafe content, if you rely on traditional manual audit methods to address the audit needs of Internet mass content, there is a problem of high audit cost, low audit efficiency, uncontrollable audit accuracy. [0004] The prior art is often detected by a single content identification technique, which can only be identified for a single text or image, and it is impossible to meet the requirements for content recognition of the multi-modal web pag...

AI Technical Summary

Problems solved by technology

[0002] With the development of Internet technology, the Internet platform has become the main platform for people to obtain information. On various Internet platforms, there are massive text, image and video data generated all the time. Among these massive data, there are extremely May produce unsafe content that is harmful to online health

[0003] If the review of insecure content relies on traditional manual review methods to meet the review needs of massive Internet content, there will be problems such as high review costs, low review efficiency, and uncontrollable review accuracy.

[0004] Existing technologies often use a single content recognition technology to detect text or images in websites, which can only recognize a single text or image, and cannot meet the requirements for content recognition of multi-modal web content

Images