Content review method based on sensitive marks

Abstract

The invention discloses a content review method based on a sensitive mark, which comprises the following steps of: 1, acquiring a content file of the sensitive mark through a data acquisition module,and transmitting the content file of the sensitive mark to an integration unit for classification and integration operation to obtain an audio file, an image file and a text document; 2, transmittingthe audio file data, the image file data and the text document data to an identification unit by the integration unit, identifying the file types by the identification unit according to the YPi, the TXa and the WBb marked in the step 1; respectively transmitting the signals to corresponding units in the analysis module; according to the method, the audio analysis unit and the image analysis unit are arranged; the text analysis unit performs violation identification on audios and images, performs layer-by-layer screening on text document data through the number of pinyin, pinyin initials, pinyin tails and tones, performs sensitive occurrence frequency calculation according to the tones and the pinyin tail data, and increases accurate judgment on sensitive characters.

Description

technical field [0001] The invention relates to the technical field of content review, in particular to a content review method based on sensitive marks. Background technique [0002] The censorship system refers to the inspection and screening of literature, art, network resources and other fields facing the public by official and more authoritative institutions to achieve their intended purpose. These are to remove some unhealthy content, or some key words or content. [0003] The announcement number is CN100456830C, which is a user terminal device that realizes streaming media content review. The user terminal device that realizes streaming media content review solves the problem of how to realize in-depth content review of video streams on user terminal devices in the prior art , but the user terminal equipment that implements streaming media content review cannot quickly identify files during file transmission, and cannot recognize the same pronunciation of some sensiti...

AI Technical Summary

Problems solved by technology

[0006] (1) How to classify and mark sensitively marked content files according to different types by setting the integration unit and the identification unit, so as to solve the problem that it is difficult to quickly query different types of files in the prior art;

[0007] (2) How to identify violations of audio and images through the settings of the audio analysis unit and the image analysis unit. At the same time, the text analysis unit screens the text document data layer by layer through the number of pinyin, the first letter of pinyin, the tail of pinyin and the tone. To solve the problem that it is difficult to accurately identify sensitive words in the prior art;