Unlock instant, AI-driven research and patent intelligence for your innovation.

Adversarial network image steganography method based on novel U-Net structure generator

A network image and generator technology, applied in the field of steganography and digital image, can solve the problem that image steganography technology cannot meet the strong demand of information security

Active Publication Date: 2021-07-06
SHANGHAI UNIVERSITY OF ELECTRIC POWER
View PDF9 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0002] In recent years, with the development of technologies such as big data and artificial intelligence, traditional image steganography has gradually been unable to meet people's strong demand for information security.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Adversarial network image steganography method based on novel U-Net structure generator
  • Adversarial network image steganography method based on novel U-Net structure generator
  • Adversarial network image steganography method based on novel U-Net structure generator

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0041] refer to Figure 1~4 , as an embodiment of the present invention, provides a kind of anti-network image steganography method based on the novel U-Net structure generator, comprising:

[0042] S1: Use the convolution strategy and the deconvolution strategy to construct a symmetrical U-shaped network, and use the shortcut (direct connection) connection strategy to connect the U-shaped network layer by layer as the generator; it should be noted that,

[0043] The process of constructing a symmetrical U-shaped network includes,

[0044] Use a 3×3 or 1×1 convolution kernel to convolve the carrier image with a step size of 2;

[0045] Each convolution operation is followed by the Batch Normalization (batch normalization) layer and the Leaky Relu activation function;

[0046] The number of convolution kernels increases at a rate of 2 times until the size of the original carrier image becomes 1×1 or 2×2;

[0047] The same deconvolution operation is performed on the cover ima...

Embodiment 2

[0084] refer to Figure 5-7It is another embodiment of the present invention. In order to verify and illustrate the technical effect adopted in this method, this embodiment uses a traditional technical solution to carry out a comparative test with the method of the present invention, and compares the test results by means of scientific demonstration to verify that the method has real effect.

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses an adversarial network image steganography method based on a novel U-Net structure generator, and the method comprises the steps: constructing U-shaped networks which are symmetric in size through employing a convolution strategy and a deconvolution strategy, and enabling the U-shaped networks to be connected layer by layer through employing a short cut connection strategy to serve as a generator; constructing a steganalysis network as a discriminator and constructing a secret information embedding simulator; inputting a carrier image, constructing a loss function to carry out adversarial training on the generator and the discriminator, embedding secret information based on a secret information embedding simulator, and generating an embedding probability image; and constructing a steganography framework based on minimum distortion according to a rate distortion principle, and carrying out information embedding and extraction on the carrier image based on the embedding probability image to finish the steganography of the adversarial network image. According to the method, the embedding difficulty of the steganography algorithm is reduced, and the practicability of the algorithm is improved; higher steganography security is realized; the targeted steganalysis is avoided; and the method has relatively high expansibility and the possibility of continuously improving steganography security.

Description

technical field [0001] The invention relates to the technical field of digital images and steganography, in particular to a method for anti-network image steganography based on a novel U-Net structure generator. Background technique [0002] In recent years, with the development of big data, artificial intelligence and other technologies, traditional image steganography technology has gradually been unable to meet people's strong demand for information security. With the continuous maturity of the application of convolutional neural network in the field of steganalysis, how to apply deep learning technology to the field of steganography to realize the safe transmission of carriers has become an important research topic. [0003] The generative confrontation network was first applied to the generation and modification of images. Because there are still large differences between the images generated by GAN technology and natural images, it is fundamentally determined that the ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06F21/60G06N3/04G06N3/06G06N3/08
CPCG06F21/602G06N3/061G06N3/08G06N3/045
Inventor 栗风永于宗良
Owner SHANGHAI UNIVERSITY OF ELECTRIC POWER