CN114928681A - Information hiding method and system based on generation countermeasure network - Google Patents

Information hiding method and system based on generation countermeasure network Download PDF

Info

Publication number
CN114928681A
CN114928681A CN202210263544.1A CN202210263544A CN114928681A CN 114928681 A CN114928681 A CN 114928681A CN 202210263544 A CN202210263544 A CN 202210263544A CN 114928681 A CN114928681 A CN 114928681A
Authority
CN
China
Prior art keywords
secret
image
texture
tensor
information
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202210263544.1A
Other languages
Chinese (zh)
Other versions
CN114928681B (en
Inventor
刘熙尧
马子平
张健
方辉
马隽星
张伟
贺建飙
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Central South University
Original Assignee
Central South University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Central South University filed Critical Central South University
Priority to CN202210263544.1A priority Critical patent/CN114928681B/en
Publication of CN114928681A publication Critical patent/CN114928681A/en
Application granted granted Critical
Publication of CN114928681B publication Critical patent/CN114928681B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
    • H04N1/32—Circuits or arrangements for control or supervision between transmitter and receiver or between image input and image output device, e.g. between a still-image camera and its memory or between a still-image camera and a printer device
    • H04N1/32101—Display, printing, storage or transmission of additional information, e.g. ID code, date and time or title
    • H04N1/32144—Display, printing, storage or transmission of additional information, e.g. ID code, date and time or title embedded in the image data, i.e. enclosed or integrated in the image, e.g. watermark, super-imposed logo or stamp
    • H04N1/32149—Methods relating to embedding, encoding, decoding, detection or retrieval operations
    • H04N1/32267—Methods relating to embedding, encoding, decoding, detection or retrieval operations combined with processing of the image
    • H04N1/32272—Encryption or ciphering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/047—Probabilistic or stochastic networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
    • H04N1/32—Circuits or arrangements for control or supervision between transmitter and receiver or between image input and image output device, e.g. between a still-image camera and its memory or between a still-image camera and a printer device
    • H04N1/32101—Display, printing, storage or transmission of additional information, e.g. ID code, date and time or title
    • H04N1/32144—Display, printing, storage or transmission of additional information, e.g. ID code, date and time or title embedded in the image data, i.e. enclosed or integrated in the image, e.g. watermark, super-imposed logo or stamp
    • H04N1/32149—Methods relating to embedding, encoding, decoding, detection or retrieval operations
    • H04N1/32267—Methods relating to embedding, encoding, decoding, detection or retrieval operations combined with processing of the image
    • H04N1/32277—Compression

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Computing Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Software Systems (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Multimedia (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Mathematical Physics (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Medical Informatics (AREA)
  • Signal Processing (AREA)
  • Probability & Statistics with Applications (AREA)
  • Image Processing (AREA)

Abstract

The invention discloses an information hiding method and system based on a generation countermeasure network. The sender selects proper mapping parameters, maps the secret information to be sent into a secret tensor through a mapping function, and inputs the secret tensor into a structure generator to generate a structure code; the generator generates a secret-carrying image according to the structure coding and the texture coding; sending the secret-carrying image to a receiving party; the receiver inputs the encrypted image into the encoder to extract the structural code; the secret tensor is recovered using an extractor, and the secret information is reversely recovered using a mapping function between the secret information and the secret tensor. The structure coding and the texture coding respectively and independently control the structure characteristic and the texture characteristic of the secret-carrying image, and simultaneously, the generated secret-carrying image has different texture characteristics by sampling different texture codes.

Description

Information hiding method and system based on generation countermeasure network
Technical Field
The invention relates to the field of information hiding, in particular to an information hiding method and system based on a generation countermeasure network.
Background
The image hiding technology is one of the most challenging and important subjects in the field of information security, the main purpose of the technology is to hide secret information into a digital image, and related methods are widely applied to the fields of image data enhancement, copyright protection, military secret communication and the like.
In recent years, steganalysis technology which is continuously developed makes traditional embedded image hiding technology face a security bottleneck, and such methods which realize hiding by modifying a carrier image can damage the naturalness of the carrier image, leave easily-perceived modification traces on image statistical characteristics, and are difficult to resist the detection of traditional steganalysis algorithms.
In order to solve the problem, the invention provides an information hiding method and system based on a generation countermeasure network, which uses a generated image as a secret-carrying image, can fundamentally resist the detection of the traditional steganalysis algorithm, and greatly improves the transmission efficiency, the safety and the secrecy of secret information transmission.
Disclosure of Invention
The invention provides an information hiding method and system based on a generation countermeasure network, which are used for solving the technical problems of low transmission efficiency and poor secrecy of an information hiding technology.
The embodiment of the invention also provides an information hiding method based on the generation countermeasure network, which comprises the following steps:
a sender selects proper mapping parameters according to an actual application scene, maps secret information to be sent into a secret tensor through a secret information-secret tensor mapping function, and inputs the secret information into a structure generator to generate a structure code; the generator generates a secret-carrying image according to the structural coding and the randomly sampled texture coding in the specific distribution; sending the secret-carrying image to a receiver;
after receiving the secret-carrying image, the receiver inputs the secret-carrying image into an encoder to extract a structure code; the secret tensor is recovered again by the extractor, and the secret information is reversely recovered by a mapping function between the secret information and the secret tensor.
Preferably, before the sender and the receiver communicate, the method further includes: training image structure-texture decoupling on a target data set to generate a countermeasure network; the trained counterpoise network includes a trained structure generator and generator for distribution to a sender of secret information, and an encoder and extractor for distribution to a receiver of secret information.
Preferably, the secret information to be sent is in the form of a binary information string; the secret tensor is in the form of a floating point value tensor.
Preferably, training the image "structure-texture" decoupling generation countermeasure network on the target dataset comprises:
the image 'structure-texture' decoupling generation countermeasure network comprises the following steps: the system comprises an encoder, a generator, a structure generator, an extractor, a truth discriminator, a distribution discriminator and a texture discriminator;
selecting an image from an image library as a reference image X, inputting the reference image X into an encoder, and outputting a first structural code S 1 And a first texture coding T 1 Encoding the first texture using a distribution discriminator 1 Supervising to make the texture coding space accord with specific distribution;
randomly sampling a second texture code T over a particular distribution 2 And a secret tensor Z;
inputting the secret tensor Z into the structure generator and outputting a second structure code S 2 ；
Encoding the first structure S 1 And a first texture coding T 1 An input generator for outputting a first image as a reconstructed image
Figure BDA0003545882070000021
Encoding the second structure S 2 And a first texture coding T 1 An input generator for outputting the second image
Figure BDA0003545882070000022
Encoding the second structure S 2 And a second texture coding T 2 An input generator for outputting the third image
Figure BDA0003545882070000023
Computing a first image using a texture discriminator
Figure BDA0003545882070000024
And a second image
Figure BDA0003545882070000025
Global texture feature similarity between the images helps the network to realize the structural and texture decoupling of the images; using a degree of realism discriminator on a first image
Figure BDA0003545882070000026
Second image
Figure BDA0003545882070000027
And a third image
Figure BDA0003545882070000028
Performing supervision to train the generator in terms of its ability to generate high quality images;
the second image
Figure BDA0003545882070000029
Input encoder for extracting the second image
Figure BDA00035458820700000210
Structure coding and texture coding of the third image
Figure BDA00035458820700000211
Input encoder for extracting a third image
Figure BDA00035458820700000212
Structure coding and texture coding of (1); the texture coding extracted by the monitoring encoder of the distribution discriminator is in accordance with the specific distribution, so that any sample on the specific distribution can correspond to reasonable texture characteristics to realize the generation of the texture coding;
encoding S according to a second structure using an extractor 2 And restoring the secret tensor Z, and training the extractor by using the average absolute error between the restored secret tensor and the secret tensor Z as a loss so as to train the secret tensor recovery capability of the extractor.
Preferably, the "secret information-secret tensor" mapping function divides the secret information into a plurality of binary information sections with equal length without overlapping, maps the binary information sections to fixed floating point values, and arranges the floating point values into a tensor form according to the sequence of the secret information sections to obtain the secret tensor after all the secret information sections are converted into the floating point values.
The embodiment of the invention also provides an information hiding system based on a generation countermeasure network, which comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor realizes the steps of the method when executing the computer program.
The invention has the following beneficial effects:
the invention relates to an information hiding method and an information hiding system based on a generated countermeasure network.A sender selects proper mapping parameters according to an actual application scene, maps secret information to be sent into a secret tensor through a secret information-secret tensor mapping function, and inputs the secret tensor into a structure generator to generate a structure code; the generator generates a secret-carrying image according to the structural coding and the randomly sampled texture coding in the specific distribution; sending the secret-carrying image to a receiving party; after receiving the secret-carrying image, the receiver inputs the secret-carrying image into an encoder to extract a structure code; and recovering the secret tensor by using the extractor, and reversely recovering the secret information by using a mapping function between the secret information and the secret tensor. The invention effectively improves the stability of the non-carrier image hiding technology by means of the stability of the image structure, and simultaneously generates the secret-carrying images with great difference in visual effect by sampling different texture codes, thereby greatly improving the diversity of the secret-carrying images, improving the information hiding transmission efficiency and improving the secrecy and the safety of the non-carrier image hiding technology.
In addition to the objects, features and advantages described above, other objects, features and advantages of the present invention are also provided. The present invention will be described in further detail below with reference to the accompanying drawings.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate embodiments of the invention and, together with the description, serve to explain the invention and not to limit the invention. In the drawings:
fig. 1 is a schematic diagram of an information hiding method and system based on a generation countermeasure network according to a preferred embodiment of the present invention;
fig. 2 is a diagram illustrating the correspondence between binary information segments and subintervals according to the preferred embodiment of the present invention.
Detailed Description
The embodiments of the invention will be described in detail below with reference to the drawings, but the invention can be implemented in many different ways as defined and covered by the claims.
Example 1:
referring to fig. 1, an information hiding method based on a generation countermeasure network includes:
s1, the sender selects proper mapping parameters according to the practical application scene, and maps the secret information to be sent into a secret tensor through a secret information-secret tensor mapping function;
s2 the sender inputs the secret tensor into the structure generator to generate the structure code; the generator generates a secret-carrying image according to the structure code and the randomly sampled texture code in the specific distribution; sending the secret-carrying image to a receiver;
s3, the receiver inputs the secret-carrying image into the coder to extract the structure code after receiving the secret-carrying image; recovering the secret tensor using the extractor;
s4 the receiver reversely recovers the secret information using a mapping function between the secret information and the secret tensor.
Optionally, the sender may send the secret-carrying image to the receiver through one or a combination of more than one of the internet of things, a wireless network, optical fiber communication, and a cloud-edge cooperative network.
Before the sender and the receiver communicate, the method further comprises: training image structure-texture decoupling on a target data set to generate a countermeasure network; the trained counterpoise network includes a trained structure generator and generator for distribution to a sender of secret information, and an encoder and extractor for distribution to a receiver of secret information.
The secret information to be sent is in a binary information string form; the secret tensor is in a floating point value tensor form;
in this alternative embodiment, training the image "structure-texture" decoupling generation countermeasure network on the target dataset includes:
the image 'structure-texture' decoupling generation countermeasure network comprises the following steps: the device comprises an encoder, a generator, a structure generator, an extractor, a truth discriminator, a distribution discriminator and a texture discriminator;
wherein:
the encoder is used for decoupling and encoding the structure and the texture features of the image and extracting the structure encoding and the texture encoding of the image;
the generator is composed of a plurality of style convolution blocks, structure coding is used as an input feature map, texture coding is used for modulating convolution kernel parameters in the style convolution blocks, and finally a high-quality dense-carrying image with structure features corresponding to the structure coding and texture features corresponding to the texture coding is generated;
the structure generator is used for generating reasonable structure codes which accord with distribution according to the secret tensor (the length and the width are consistent with the structure codes, and the number of channels can be set according to requirements);
the extractor is used for restoring the secret tensor according to the structure code and is mainly applied to the secret information extraction process;
the truth discriminator is used for distinguishing the generated image from the real image so as to help the generator generate an image which is distributed more closely to the real image;
the texture discriminator is used for calculating the global texture feature similarity between the images and helping the network to realize decoupling coding on the structural features and the texture features of the images;
the distribution discriminator is used for restricting the texture coding extracted by the encoder to accord with the specific distribution, so that any sample on the specific distribution can correspond to reasonable texture characteristics, and the generation of the texture coding is realized.
Selecting an image from an image library as a reference image X, inputting the reference image X into an encoder, and outputting a first structural code S 1 And a first texture coding T 1 Encoding the first texture T using a distribution discriminator 1 Supervising to make the texture coding space accord with specific distribution;
randomly sampling a second texture code T over a particular distribution 2 And a secret tensor Z;
inputting the secret tensor Z into the structure generator and outputting a second structure code S 2 ；
Encoding the first structure S 1 And a first texture coding T 1 An input generator for outputting a first image as a reconstructed image
Figure BDA0003545882070000041
Encoding the second structure S 2 And a first texture coding T 1 An input generator for outputting a second image
Figure BDA0003545882070000042
Encoding the second structure S 2 And a second texture coding T 2 An input generator for outputting the third image
Figure BDA0003545882070000043
Computing a first image using a texture discriminator
Figure BDA0003545882070000044
And a second image
Figure BDA0003545882070000045
Global texture feature similarity between the images helps the network to realize the structural and texture decoupling of the images; using a degree of truth discriminator for a first image
Figure BDA0003545882070000046
Second image
Figure BDA0003545882070000047
And a third image
Figure BDA0003545882070000048
Performing supervision to train the generator in terms of its ability to generate high quality images;
the second image
Figure BDA0003545882070000049
Input encoder for extracting the second image
Figure BDA00035458820700000410
Structure coding and texture coding of the third image
Figure BDA00035458820700000411
Input encoder for extracting a third image
Figure BDA00035458820700000412
Structure coding and texture coding of (3); texture coding using distributed arbiter supervised encoder extractionThe codes conform to specific distribution, so that any sample on the specific distribution can correspond to reasonable texture features to realize the generation of texture codes;
encoding S according to a second structure using an extractor 2 And restoring the secret tensor Z, and training the extractor by using the average absolute error between the restored secret tensor and the secret tensor Z as a loss so as to train the secret tensor recovery capability of the extractor.
In an alternative embodiment, the "secret information-secret tensor" mapping function divides the secret information into a plurality of binary information sections with equal length without overlapping, maps the binary information sections to fixed floating point values, and arranges the floating point values into a tensor form according to the sequence of the secret information sections to obtain the secret tensor after all the secret information sections are converted into the floating point values.
It should be noted that the mapping function first divides the secret information into a plurality of binary information segments with length σ without overlapping.
Specifically, the mapping function is in the interval [ -1,1]Is equidistantly provided with 2 σ And the sub-intervals have the length of 2 multiplied by delta multiplied by r, and when a binary information segment corresponding to a decimal value m needs to be mapped, a floating point value z is randomly selected in the mth (starting from 0) sub-interval to represent the information segment. For any binary information segment, the corresponding decimal value is m, and the mapping result z is expressed as:
Figure BDA0003545882070000051
where z is the floating point value, σ is the length of the binary segment, m is the corresponding decimal number, r-1/2 σ-1 Is the maximum length of the subinterval, Delta epsilon [ 0%, 50%]For balancing the stability and randomness of the mapping function, when Δ ═ 0%, the randomness in the mapping function is removed, and the binary information segment will be mapped to a fixed floating point value, i.e. the mapping result z is expressed as:
Figure BDA0003545882070000052
where z is the floating point value, σ is the length of the binary segment, and m is the corresponding decimal number.
Referring to fig. 2, when σ is 2 and Δ is 25%, binary information segment "00" corresponds to a random value in an interval (-0.875, -0.625), binary information segment "01" corresponds to a random value in an interval (-0.375, -0.125), binary information segment "10" corresponds to a random value in an interval (0.125, 0.375), and binary information segment "11" corresponds to a random value in an interval (0.625, 0.875).
And after all the secret information sections are converted into floating point values, arranging the floating point values into a tensor form according to the sequence of the secret information sections.
Example 2:
an information hiding system based on a generation countermeasure network comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor executes the computer program to realize the steps of the method of the embodiment 1.
In summary, the information hiding method and system based on the generation countermeasure network of the invention select proper mapping parameters by the sender, map the secret information to be sent into a secret tensor through a mapping function, and input the secret tensor into a structure generator to generate a structure code; the generator generates a secret-carrying image according to the structure coding and the texture coding; sending the secret-carrying image to a receiver; the receiver receives the encrypted image and inputs the encrypted image into the encoder to extract the structural code; the secret tensor is recovered using an extractor, and the secret information is reversely recovered using a mapping function between the secret information and the secret tensor. The invention effectively improves the stability of the carrier-free image hiding technology by means of the stability of the image structure, and simultaneously generates the secret-carrying images with great difference in visual effect by sampling different texture codes, thereby greatly improving the diversity of the secret-carrying images, improving the information hiding transmission efficiency and improving the secrecy and the safety of information hiding.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (6)

1. An information hiding method based on a generation countermeasure network is characterized by comprising the following steps:
a sender selects proper mapping parameters according to an actual application scene, maps secret information to be sent into a secret tensor through a secret information-secret tensor mapping function, and inputs the secret tensor into a structure generator to generate a structure code; the generator generates a secret-carrying image according to the structural coding and the randomly sampled texture coding in the specific distribution; sending the secret-carrying image to a receiving party;
after receiving the secret-carrying image, the receiver inputs the secret-carrying image into an encoder to extract a structure code; the secret tensor is then recovered using the extractor, and the secret information is recovered inversely using a mapping function between the secret information and the secret.
2. The information hiding method based on generation countermeasure network of claim 1, wherein before the sender and the receiver communicate, the method further comprises: training image structure-texture decoupling on a target data set to generate a confrontation network; the trained counterpoise network includes a trained structure generator and generator for distribution to a sender of secret information, and an encoder and extractor for distribution to a receiver of secret information.
3. The information hiding method based on generation countermeasure network of claim 1, wherein the secret information to be transmitted is in the form of binary information string; the secret tensor is in the form of a floating point value tensor.
4. The method for hiding information based on generation of a countermeasure network as claimed in claim 1, wherein the training of image structure-texture decoupling generation of countermeasure network on target data set comprises:
the image structure-texture decoupling generation countermeasure network comprises the following steps: the system comprises an encoder, a generator, a structure generator, an extractor, a truth discriminator, a distribution discriminator and a texture discriminator;
selecting an image from an image library as a reference image X, inputting the reference image X into the encoder, and outputting a first structural code S 1 And a first texture coding T 1 Encoding the first texture T using the distribution discriminator 1 Supervising to make the texture coding space accord with specific distribution;
randomly sampling a second texture code T over a particular distribution 2 And a secret tensor Z;
inputting the secret tensor Z into the structure generator and outputting a second structure code S 2 ；
Encoding the first structure S 1 And a first texture coding T 1 An input generator for outputting a first image as a reconstructed image
Figure FDA0003545882060000011
Encoding the second structure S 2 And a first texture coding T 1 An input generator for outputting a second image
Figure FDA0003545882060000012
Encoding the second structure S 2 And a second texture coding T 2 An input generator for outputting the third image
Figure FDA0003545882060000013
Computing the first image using the texture discriminator
Figure FDA0003545882060000014
And the second image
Figure FDA0003545882060000015
Global texture feature similarity between the images helps the network to realize the structural and texture decoupling of the images; using the truth degree discriminator to the first image
Figure FDA0003545882060000016
Second image
Figure FDA0003545882060000017
And a third image
Figure FDA0003545882060000018
Performing supervision to train the generator in accordance with the ability to generate high quality images;
the second image
Figure FDA0003545882060000019
Input encoder for extracting the second image
Figure FDA00035458820600000110
Structure coding and texture coding of the third image
Figure FDA00035458820600000111
Input encoder for extracting a third image
Figure FDA00035458820600000112
Structure coding and texture coding of (1); using a distribution discriminator to supervise the texture coding extracted by the encoder to accord with the specific distribution, so that any sample on the specific distribution can correspond to reasonable texture characteristics to realize the generation of the texture coding;
encoding S according to a second structure using the extractor 2 And restoring the secret tensor Z, and training the extractor by using the average absolute error between the restored secret tensor and the secret tensor Z as a loss so as to train the secret tensor recovery capability of the extractor.
5. The information hiding method based on the generative countermeasure network according to any one of claims 1 to 4, wherein the secret information-secret tensor mapping function divides the secret information into a plurality of binary information sections with equal length without overlapping, maps the binary information sections to fixed floating point values, and arranges the floating point values into tensor forms according to the order of the secret information sections after all the secret information sections are converted into the floating point values to obtain the secret tensor.
6. An information hiding system based on a spanning confrontation network, comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the method of any of the preceding claims 1 to 5 when executing the computer program.
CN202210263544.1A 2022-03-14 2022-03-14 Information hiding method and system based on generation countermeasure network Active CN114928681B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210263544.1A CN114928681B (en) 2022-03-14 2022-03-14 Information hiding method and system based on generation countermeasure network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210263544.1A CN114928681B (en) 2022-03-14 2022-03-14 Information hiding method and system based on generation countermeasure network

Publications (2)

Publication Number Publication Date
CN114928681A true CN114928681A (en) 2022-08-19
CN114928681B CN114928681B (en) 2024-02-27

Family

ID=82805735

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210263544.1A Active CN114928681B (en) 2022-03-14 2022-03-14 Information hiding method and system based on generation countermeasure network

Country Status (1)

Country Link
CN (1) CN114928681B (en)

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10198782B1 (en) * 2016-09-15 2019-02-05 Digimarc Corporation Detecting conflicts between multiple different encoded signals within imagery
US10212306B1 (en) * 2016-03-23 2019-02-19 Amazon Technologies, Inc. Steganographic camera communication
CN111275603A (en) * 2019-12-09 2020-06-12 中国科学院信息工程研究所 Security image steganography method based on style conversion and electronic device
CN111951149A (en) * 2020-08-14 2020-11-17 中国人民武装警察部队工程大学 Image information steganography method based on neural network
WO2021010896A1 (en) * 2019-07-12 2021-01-21 Nanyang Technological University Method and system for distributed data management
CN112615974A (en) * 2020-12-14 2021-04-06 中国科学院信息工程研究所 Carrier-free covert communication method and system based on depth discriminator
CN114140309A (en) * 2021-12-03 2022-03-04 中国人民武装警察部队工程大学 Novel image steganography method and system based on NICE model

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10212306B1 (en) * 2016-03-23 2019-02-19 Amazon Technologies, Inc. Steganographic camera communication
US10198782B1 (en) * 2016-09-15 2019-02-05 Digimarc Corporation Detecting conflicts between multiple different encoded signals within imagery
WO2021010896A1 (en) * 2019-07-12 2021-01-21 Nanyang Technological University Method and system for distributed data management
CN111275603A (en) * 2019-12-09 2020-06-12 中国科学院信息工程研究所 Security image steganography method based on style conversion and electronic device
CN111951149A (en) * 2020-08-14 2020-11-17 中国人民武装警察部队工程大学 Image information steganography method based on neural network
CN112615974A (en) * 2020-12-14 2021-04-06 中国科学院信息工程研究所 Carrier-free covert communication method and system based on depth discriminator
CN114140309A (en) * 2021-12-03 2022-03-04 中国人民武装警察部队工程大学 Novel image steganography method and system based on NICE model

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
ZHEN WANG; ZHEN ZHANG; JIANHUI JIANG: "Multi-Feature Fusion based Image Steganography using GAN", 《2021 IEEE INTERNATIONAL SYMPOSIUM ON SOFTWARE RELIABILITY ENGINEERING WORKSHOPS (ISSREW)》, pages 280 - 281 *
付章杰; 王帆; 孙星明; 王彦: "基于深度学习的图像隐写方法研究", 《计算机学报》, vol. 43, no. 9, pages 1656 - 1672 *
雷雨；刘佳；李军；柯彦: "一种基于条件生成对抗网络的图像隐写方法研究与实现", 《信息网络安全》, no. 11, pages 48 - 57 *

Also Published As

Publication number Publication date
CN114928681B (en) 2024-02-27

Similar Documents

Publication Publication Date Title
Cao et al. High capacity reversible data hiding in encrypted images by patch-level sparse representation
Zhang et al. Reference sharing mechanism for watermark self-embedding
CN113194213B (en) PNG image information hiding and recovering method based on secret sharing and chaotic mapping
CN106709854B (en) Image information fusion encryption method based on cat face transformation and chaos
CN111984942B (en) Robust video zero watermarking method based on polar complex exponential transformation and residual neural network
CN112615974A (en) Carrier-free covert communication method and system based on depth discriminator
Mokhtar et al. Colored image encryption algorithm using DNA code and chaos theory
CN110417551B (en) Character and picture generation type camouflage and recovery method combined with error diffusion
Chen et al. Steganographic secret sharing with GAN-based face synthesis and morphing for trustworthy authentication in IoT
CN108460023B (en) Method for disguising and recovering legal section double-authentication full-key dependence carrier-free test questions
CN114928681B (en) Information hiding method and system based on generation countermeasure network
CN116827509A (en) Image encryption method based on five-dimensional conserved hyperchaotic system and bit plane segmentation diffusion
CN112184841B (en) Block replacement generation type information hiding and recovering method, equipment and medium
CN108171665B (en) Multi-version backup and restrictive dual authentication master key (t, s, k, n) image separate storage method
CN110569945A (en) QR code encryption method based on single-intensity recording phase recovery process
CN108156136B (en) Secret-pattern carrier-free test question disguising method for indirect transmission and random codebook
Alosta et al. Design of effective lossless data compression technique for multiple genomic DNA sequences
CN115134142B (en) Information hiding method and system based on file segmentation
Sharma et al. Generative network based image encryption
Ch et al. Robust steganographic framework for securing sensitive healthcare data of telemedicine using convolutional neural network
CN116033086B (en) Reversible neural network-based image hiding method
Huang et al. Visual cryptography for compressed sensing of images with transmission over multiple channels
CN110727956B (en) Double-authentication test question backup disguising method combining codebook expansion and question stem hashing
KR20140091393A (en) Apparatus and method for steganography, apparatus and method for data restoration
CN114389787B (en) Carrier-free information hiding method and system based on chaotic system and computer storage medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant