Image anti-theft system and method based on deep learning

Abstract

The invention discloses an image anti-theft system and method based on deep learning, the system comprises a generator, a discriminator and an optimizer, the optimizer configures a loss function, andthe generator and the discriminator form an adversarial network. The reciprocal third layer of the generator reduces the channel dimension to 3 channels, the reciprocal second layer of the generator lifts the channel to 12 dimensions, and the last layer of the generator reduces the channel dimension to 3 channels; the discriminator outputs a feature map by continuously down-sampling and expandinga channel so as to judge whether the graph is true or false; the optimizer optimizes the generator and the discriminator through a mode of combining MSE loss and GAN loss. The generator and the discriminator are subjected to cross training, the output of the last but one layer of the trained generator is used as an object accessible to a user, and after the user determines the accessed object, animage with the effect consistent with that of the original image is obtained through the last two layers of processing of the optimizer and displayed. The display effect of the image is not influenced, a user can only access the intermediate image, and the anti-theft effect is good.

Description

technical field [0001] The invention relates to the field of machine learning, in particular to an image anti-theft system based on deep learning, and an image anti-theft method that uses the network structure of the system for training and then operates. Background technique [0002] Most of the image resources that can be seen everywhere on the Internet can be downloaded and used at will, which is very unfriendly to some companies. For example, some course sellers need anti-theft protection for image resources. [0003] At present, most anti-theft technologies are realized through the following methods. One and the simplest is to add a watermark to the image and attach a set label. This method first affects the display of the image content, and the other is that the watermark can also be used to a certain extent. is removed, the security is weaker. Another anti-theft technology is to limit the image URLs that are exposed to the outside world. Anonymous visitors can only g...

AI Technical Summary

Problems solved by technology

[0003] At present, most anti-theft technologies are realized through the following methods. One and the simplest is to add a watermark to the image and attach a set label. This method first affects the display of the image content, and the other is that the watermark can also be used to a certain extent. is removed, the security is weaker

Another anti-theft technology is to limit the image URLs exposed to the outside world. Anonymous visitors can only get thumbnailed or watermarked images and cannot download the original images. However, this method has a great impact on the display effect.

Images