Violence and terrorism picture safety detection system based on deep learning

Abstract

A violence and terrorism picture safety detection system based on deep learning comprises a violence and terrorism picture detection and violence and terrorism degree analysis module, a violence and terrorism scene classification and gun identification module, an image-text analysis module in a natural scene, a violence and terrorism political head sample generation module, a violence and terrorism political head identification module and a violence and terrorism flag identification module, wherein the violence and terrorism picture detection and violence and terrorism degree analysis module, the image-text analysis module in the natural scene, the violence and terrorism political head sample generation module and the violence and terrorism flag identification module are connected in parallel and respectively receive picture information. According to the method, the pictures under various scenes serve as input, whether the pictures are violent and terrorism pictures or not can be judged in a short time, violent and terrorism elements existing in the pictures can be detected and identified, the recognition range and recognition accuracy of a picture safety detection system are remarkably improved, the recognition efficiency is high, flexibility is high, model updating is convenient, and each module model can be regularly trained in a strengthened manner according to use conditions, so that the system performance is improved.

Description

technical field [0001] The present invention relates to a technology in the field of information security, in particular to a security detection system for violent and terrorist pictures based on deep learning. Background technique [0002] Image detection systems for specific elements of violence and terror are increasingly used in various fields. It can not only be used for the prevention, emergency response, evidence collection and filing of public security incidents, but also provide a data basis for retrospective and reconstruction after the event. As an emerging research hotspot in the field of information security, terrorist image detection technology is one of the most challenging issues in the field of content security. It describes the process of determining whether a picture to be detected contains violent and terrorist elements, and identify and classify specific violent and terrorist elements. Due to the variety of elements involved in violent and terrorist pic...

AI Technical Summary

Problems solved by technology

[0003] Aiming at the above-mentioned deficiencies in the prior art, the present invention proposes a security detection system for violent terrorism pictures based on deep learning. Aiming at the shortcomings that the current violent terrorism picture detection products cannot be detected across domains, the violent terrorism picture detection and the analysis of the degree of violence and terrorism, The technical modules of violent terror scene classification and gun recognition, graphic analysis in natural scenes, violent terrorist political leader sample generation, recognition, and violent terrorist flag recognition are organically combined

Images