Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Generative robust image steganography method

A generative, image-based technology, applied in image watermarking, image data processing, image data processing, etc., can solve the problems of not considering the consistency between generator and extractor, limited DCGAN network performance, and low accuracy of information extraction. Achieve the effect of strong statistical undetectability of secret-carrying images, overcome the poor quality and high concealment of secret-carrying images

Pending Publication Date: 2020-08-28
SUN YAT SEN UNIV
View PDF3 Cites 15 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, there are three problems in the existing DCGAN-based generative image steganography: first, the generated pseudo-image has low resolution and serious distortion, because the performance of the DCGAN network itself is limited; second, the embedding capacity is low (less than 0.01bpp), the load capacity is limited, which is caused by encoding information through hidden vectors; third, the consistency of the generator and extractor during training is not considered, and the accuracy of information extraction is relatively low

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
  • Generative robust image steganography method
  • Generative robust image steganography method
  • Generative robust image steganography method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0071] Such as figure 1 As shown, a generative robust image steganography method includes the following steps:

[0072] S1: Construct an image dataset and preprocess the image dataset;

[0073] S2: Build and initialize the deep learning network architecture;

[0074] S3: Use the combined-fine-tuning method to train the deep learning network architecture to obtain the network architecture model;

[0075] S4: Use the network architecture model to generate secret-carrying pseudo-images and perform secret communication to complete the process of generative robust image steganography.

[0076]In the specific implementation process, by using the Generative Adversarial Network StyleGAN, the embedding process of secret information is integrated into the image generation process, and a generative image steganography architecture that can bear a large amount of secret information and has certain robustness is constructed. , the resulting generative image steganography method has the ...

Embodiment 2

[0105] More specifically, the existing technology has the following technical defects:

[0106] (1) Traditional carrier-modified image steganography algorithms, also known as content-adaptive image steganography algorithms, are greatly challenged in the face of advanced steganalyzers based on convolutional neural networks. This is because the method needs to heuristically design the distortion cost function. Due to a series of factors such as the designer's own limitations and the complexity of the algorithm, it is difficult to achieve accurate modeling comprehensively and effectively, which leads to content-based adaptive Image steganography algorithms are easily detected by steganalyzers based on convolutional neural networks.

[0107](2) The carrier-modified image steganography scheme based on generative adversarial networks is still in-depth research. At present, only UT-GAN is slightly more secure than traditional image steganography algorithms, but it is still easy to be...

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 provides a generative robust image steganography method, which comprises the following steps of: constructing an image data set, and preprocessing the image data set; constructing and initializing a deep learning network architecture; training a deep learning network architecture by adopting a joint-fine adjustment method to obtain a network architecture model; and generating a secret-carrying pseudo graph by using the network architecture model and carrying out secret communication to complete an image steganography process. According tothe image steganography method, a generative adversarial network StyleGAN is utilized to fuse the embedding process of the secret information into the generation process of the image; a generative image steganography architecture which can bear secret information with relatively large capacity and has certain robustness is constructed; therefore, the obtained generative image steganography method has the advantages of being large in embedded capacity, good in generated image quality, high in secret-loaded image statistics undetectability, high in practicability and the like, and the problems that an existing generative image steganography method is poor in secret-loaded image quality, low in embedded capacity, low in information extraction accuracy and the like are solved.

Description

technical field [0001] The invention relates to the technical field of image steganography in multimedia security, and more specifically, to a generative robust image steganography method. Background technique [0002] Image steganography is to embed secret information into the carrier image through some algorithms to generate a secret image, and transmit it through an open channel. The receiver can obtain the secret information in the secret image through corresponding extraction methods, and generate The confidential image should not have obvious traces visually and statistically, so that a third party cannot know whether the image contains secret information. According to the way the carrier image is used, image steganography algorithms can be divided into carrier modification and carrier generation. The former is a method commonly used at present, which makes small modifications to the pixel values ​​of the carrier image or the coefficients of the transform domain to ac...

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): G06T1/00G06N3/04G06N3/08
CPCG06T1/005G06N3/088G06N3/045
Inventor 黄晓万林鸿倪江群
Owner SUN YAT SEN UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com