Method and device for carrying out analysis on sensitive images

Abstract

The embodiment of the invention discloses a method and device for carrying out analysis on sensitive images, relates to the technical field of image identification, and can promote an automation levelof carrying out identification detection on advertising images and reduce manual review cost. The method comprises the steps of: carrying out clustering on sample images in a training sample set; then according to the clustered sample images, by a convolutional neural network, training an identification model corresponding to each category; and then by utilizing the identification model obtainedby training, identifying sensitive images corresponding to each category from a to-be-detected image library. The method and device disclosed by the invention are applicable to identification on the sensitive images on an online platform.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a method and device for analyzing sensitive images. Background technique [0002] With the development of Internet technology and the construction of various online trading platforms, online marketing platforms and other network platforms, major operators and shops of all sizes will place massive Internet advertisements on network platforms all the time. In order to regulate the release of Internet advertisements and protect the legitimate rights and interests of consumers, the new Advertising Law promulgated in 2015 clearly stipulates that Internet advertising activities must also abide by the provisions of the Advertising Law. [0003] In the current practical application, the means of monitoring Internet advertising activities by various network platforms is mainly to identify and warn of potentially illegal advertisements by detecting sensitive images. Most of the ...

AI Technical Summary

Problems solved by technology

[0004] However, the recognition accuracy of the existing methods is low, and advertisements and publicity images of poultry products, underwear, sporting goods, family planning products and other categories of goods under normal circumstances are often misreported as sensitive images. It is difficult to meet the current real-time monitoring needs of massive Internet advertisements placed on network platforms, especially the e-commerce platform’s requirements for filtering and monitoring massive commodity promotion images. It is necessary to develop detection methods with a high degree of automation in order to control labor costs

Images