Supercharge Your Innovation With Domain-Expert AI Agents!

Double-branch network image steganography framework and method based on convolutional neural network

A convolutional neural network, network image technology, applied in the field of double-branch network image steganography framework, can solve the problem of low quality of secret images

Active Publication Date: 2021-06-08
HENAN UNIVERSITY
View PDF7 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] Aiming at the problem of low quality of generated secret images and restored secret images in traditional image steganography methods, the present invention improves the generated secret images and Aiming at the quality of the recovered secret image, a framework and method for image steganography based on convolutional neural network are provided

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
  • Double-branch network image steganography framework and method based on convolutional neural network
  • Double-branch network image steganography framework and method based on convolutional neural network
  • Double-branch network image steganography framework and method based on convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment approach

[0069] As an implementable manner, the loss function L of the encoding network enco der defined as:

[0070] L enco der =L(C,C')

[0071] The decoding network loss function L encoder defined as:

[0072] L encoder =L(S,S')

[0073] The overall loss function of the encoder-decoder network is defined as:

[0074] L total =L encoder +β Ldecoder

[0075] Among them, L(C,C')=αMSSIM(C,C')+(1-α)L1(C,C')+MSE(C,C')

[0076] L(S,S')=αMSSIM(S,S')+(1-α)L1(S,S')+MSE(S,S')

[0077] Among them, C represents the natural carrier image, C′ represents the encrypted carrier image, S represents the original secret image, S′ represents the recovered secret image, α is the coefficient for weighing MSSIM and L1 loss, β is the factor controlling the quality of the recovered secret image parameter.

[0078] Specifically, the loss of the encoding network includes two parts, one is the cover image reconstruction loss between the secret image and the cover image, which is realized by embeddin...

specific Embodiment

[0090] In order to further illustrate the image steganography method provided by the present invention, the present invention also provides another specific embodiment, specifically as follows:

[0091] Step S201: Construct a convolutional neural network for grayscale secret image steganography: including an encoding network and a decoding network; the parameter configuration of the network is as follows:

[0092] The encoding network includes a feature extraction network. Among them, the feature extraction network consists of seven layers. Except for the first layer, each layer contains two convolution modules, which are used to extract features from grayscale secret images and obtain feature maps of different sizes; JPEG grayscale images A matrix composed of pixel values ​​is used as input; the size of the convolutional kernel of the first layer is expressed in the format of "input channel number × output channel number × (convolution kernel height × convolution kernel width...

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 double-branch network image steganography framework and method based on a convolutional neural network. The framework comprises a coding network and a decoding network; the coding network comprises a feature extraction network and a steganography network; the feature extraction network is used for feature extraction of the gray secret image; the steganography network is used for embedding the grayscale image features extracted by the feature extraction network into a natural carrier image to obtain a secret-containing carrier image; and the decoding network is used for extracting a gray-scale secret image from the secret-containing carrier image. According to the invention, a double-branch steganography network in which the secret image and the carrier image are parallel is designed, a traditional single-branch image steganography framework based on a convolutional neural network is broken through, the security and transparency of image steganography are improved, the problems of gradient disappearance and the like during network training are avoided, and the training process of the network is accelerated.

Description

technical field [0001] The invention relates to the technical field of image information steganography, in particular to a convolutional neural network-based dual-branch network image steganography framework and method. Background technique [0002] Image steganography is to embed secret information into the natural carrier image without changing the semantic information and statistical characteristics of the carrier image, so as to realize the hiding of information. According to different steganographic methods, existing image steganographic algorithms can be divided into spatial domain methods and transform domain methods. [0003] The traditional spatial steganography method is to directly embed the secret information into the pixel value of the image through a certain algorithm, such as the Least Significant Bit (LSB) algorithm. The LSB algorithm has the least impact on the visual perception of the carrier image, and does not It is easy to be detected, but the steganogr...

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): G06K9/46G06N3/04G06N3/08G06T1/00G06F21/60
CPCG06T1/0021G06F21/602G06N3/084G06V10/44G06N3/048G06N3/045
Inventor 甘志华钟宇浩程晓辉张杰熊宝柱宋亚林庞子龙
Owner HENAN UNIVERSITY
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

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

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com