High-duration video pornographic content detection method based on content

Abstract

The invention discloses a high-duration video pornographic content detection method based on content. The method comprises the following steps: constructing an S3D-G network which is subjected to convolution by separating a time domain from a space domain; training to obtain a video classification network model; inputting the processed thumbnail video into a video classification network model for video classification; constructing a text sentiment classification network, and training the text sentiment classification network by using the data set to obtain a comment pornographic content classification model; classifying comment texts attached to the original video, and recording results; constructing an image classification network, and training the image classification network by using the data set to obtain a cover pornographic content classification model; classifying cover images attached to the original video, and recording a result; and comprehensively evaluating the video, cover and comment results of the original video. According to the invention, the defects of low efficiency, high hardware requirement and high delay in processing the long video with high resolution change in the prior art are overcome.

Description

technical field [0001] The invention relates to a video content review method, in particular to a content-based high-duration video pornographic content detection method. Background technique [0002] The rapid development of Internet technology and self-media has brought a lot of convenience to people's work and life. However, while the sharing of online video resources provides convenience for people, some harmful videos such as terrorist violence and pornography seriously endanger the physical and mental health of young people. my country's "Constitution" and "Minor Protection Law" expressly prohibit harmful video content. This makes the video review technology an indispensable and important technology for major video websites, which has attracted widespread attention. [0003] Existing video review methods mostly use manual methods, and video reviewers are often inefficient when reviewing. However, some intelligent identification methods that do not require human inter...

AI Technical Summary

Problems solved by technology

[0005] In order to overcome the shortcomings of the existing million-level resolution and video review method with a duration of more than 30 minutes, which are highly dependent on manual work, take a long time, and consume huge computing resources, the present invention proposes a content-based high-duration video pornographic content detection method to efficiently and intelligently to review

Images