Picture classification method and device

Abstract

An embodiment of the invention provides a picture classification method and a picture classification device, which subtly utilize the principle that splicing positions among pictures are edges, namely, edge intensity of corresponding rows or columns of the splicing positions among pictures is maximum. The picture classification method comprises the steps of: firstly, acquiring a horizontal edge graph and a vertical edge graph of a picture; acquiring edge intensity of each row and edge intensity of each column of the picture according to the horizontal edge graph and the vertical edge graph; calculating a repetition period of the edge intensities of the rows and a repetition period of the edge intensities of the columns of the picture according to the edge intensities of the rows and the edge intensities of the columns of the picture; and judging whether the picture is a multi-picture spliced picture according to the repetition period of the edge intensities of the rows and the repetition period of the edge intensities of the columns. Therefore, the manual screening of the multi-picture spliced picture from massive pictures is avoided, and the picture auditing efficiency is improved.

Description

technical field [0001] The present application relates to the technical field of image processing, and more specifically relates to an image classification method and device. Background technique [0002] Picture sharing is a distribution service for network picture storage. Network users can store their local pictures on the website server that provides picture sharing services, and share them by embedding them in blogs or forums through certain picture sharing codes. Due to the diversity of Internet users, the uploaded images are also mixed with illegal content, including pornographic images, political images and advertising images. As the staff of the website server, they are responsible for a huge task of reviewing pictures. [0003] At present, automatic image recognition algorithms have been adopted to automatically identify whether the content of pictures uploaded by network users is legal, so as to reduce the intensity of manual work. However, with the popularity o...

AI Technical Summary

Problems solved by technology

[0005] In view of this, the present invention provides a picture classification method and device to overcome the need for staff in the prior art to manually screen out multi-picture collage pictures from a large number of pictures, and manually review the multi-picture collage pictures, thereby reducing the number of pictures. The problem of audit efficiency

Images