Illegal picture identification method, medium and computing device

Abstract

The embodiment of the invention provides a violation picture identification method. The method comprises the following steps: inputting a to-be-recognized first picture into a recognition model, and determining the compliance of the first picture based on an output result of the recognition model; wherein the identification model comprises a plurality of mutually independent supervision models and a plurality of mutually independent semi-supervision models; the identification model is obtained by training the plurality of supervised models and the plurality of semi-supervised models. According to the method, the recognition module used for recognizing the illegal images can be trained through the multiple mutually independent supervision models and the multiple mutually independent semi-supervision models, the problem that the generalization ability of a single training model is weak is solved, the recognition accuracy of the recognition model on the illegal images is remarkably improved, and the recognition precision of the illegal images is high. In addition, the embodiment of the invention provides a medium and computing equipment.

Description

technical field [0001] The embodiments of the present disclosure relate to the field of image recognition, and more particularly, the embodiments of the present disclosure relate to a method, medium, and computing device for identifying illegal pictures. Background technique [0002] This section is intended to provide a background or context for the embodiments of the present disclosure that are recited in the claims. The descriptions herein are not admitted to be prior art by inclusion in this section. [0003] With the increasing requirements for network purification, it is necessary to intensify efforts to remove illegal pictures on the network. And the illegal pictures can be found by the recognition model. [0004] In an exemplary technique, the recognition model is learned by a supervised model. When conducting supervised learning, it is necessary to use illegal images as annotated samples for learning. Compared with normal pictures, the proportion of illegal pict...

Images