Watermark embedding-based copyright verification method for neural network model

Abstract

The invention relates to the field of digital watermark embedding, in particular to a watermark embedding-based copyright verification method and device for a neural network model, equipment and a storage medium. The method comprises the steps of obtaining a watermark image data set, extracting a plurality of watermark images in the watermark image data set as an original trigger set, and performing singular value decomposition processing on the original trigger set to obtain a trigger set sample, extracting classification label data corresponding to the watermark image in the original trigger set, and enabling the classification label data to correspond to the trigger set samples one by one to obtain the trigger set, inputting the watermark image data set and the trigger set into a neural network model in which the watermark is to be embedded, and training the neural network model for a plurality of times to obtain a target neural network model, and in response to a model copyright verification instruction of a user, inputting the trigger set into the target neural network model, obtaining a classification result, and obtaining a verification result output by the target neural network model according to the watermark image in the trigger set, the corresponding classification label data and the classification result.

Description

technical field [0001] The invention relates to the field of digital watermark embedding, in particular to a copyright verification method, device, equipment and storage medium based on a watermark embedding neural network model. Background technique [0002] With the rapid development of the field of deep learning, neural network models are more and more widely used in enterprises. However, because attackers tamper with other people's neural network models and earn profits by illegally authorizing them to others, the rights of the model creators are violated. Copyright brings huge economic losses to model creators. Therefore, it is necessary to protect the copyright of neural network models. [0003] With the widespread use of digital watermarking technology, the neural network model can be protected by embedding watermarks in the neural network model, and the current existing neural network watermark embedding methods will affect the function of the neural network model it...

AI Technical Summary

Problems solved by technology

[0002] With the rapid development of the field of deep learning, neural network models are more and more widely used in enterprises. However, because attackers tamper with other people's neural network models and earn profits by illegally authorizing them to others, the rights of the model creators are violated. Copyright brings huge economic losses to model creators, therefore, it is necessary to protect the copyright of neural network models

[0003] With the widespread use of digital watermarking technology, the neural network model can be protected by embedding watermarks in the neural network model, and the current existing neural network watermark embedding methods will affect the function of the neural network model itself, thus Problems that lead to a decrease in model task classification accuracy, insufficient watermark versatility, or a decrease in robustness

Images