CN115731089B - Dual-task image steganography method based on component energy - Google Patents
Dual-task image steganography method based on component energy Download PDFInfo
- Publication number
- CN115731089B CN115731089B CN202211619646.9A CN202211619646A CN115731089B CN 115731089 B CN115731089 B CN 115731089B CN 202211619646 A CN202211619646 A CN 202211619646A CN 115731089 B CN115731089 B CN 115731089B
- Authority
- CN
- China
- Prior art keywords
- image
- hidden
- energy
- secret
- representing
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 82
- 238000000605 extraction Methods 0.000 claims abstract description 33
- 238000013528 artificial neural network Methods 0.000 claims abstract description 22
- 230000002441 reversible effect Effects 0.000 claims abstract description 21
- 230000008569 process Effects 0.000 claims description 41
- 230000006870 function Effects 0.000 claims description 23
- 238000005457 optimization Methods 0.000 claims description 6
- 230000009977 dual effect Effects 0.000 claims description 4
- 238000011176 pooling Methods 0.000 claims description 4
- 210000004763 bicuspid Anatomy 0.000 claims 2
- 230000004931 aggregating effect Effects 0.000 claims 1
- 238000009954 braiding Methods 0.000 claims 1
- 238000006243 chemical reaction Methods 0.000 description 5
- 238000013527 convolutional neural network Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 238000004891 communication Methods 0.000 description 3
- 238000003062 neural network model Methods 0.000 description 3
- 238000012512 characterization method Methods 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 125000004122 cyclic group Chemical group 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000001174 ascending effect Effects 0.000 description 1
- 230000002457 bidirectional effect Effects 0.000 description 1
- 239000000470 constituent Substances 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000006386 memory function Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 239000000758 substrate Substances 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Landscapes
- Image Processing (AREA)
Abstract
The disclosure belongs to the field of image encryption, and particularly relates to a component energy-based dual-task image steganography method, which comprises the following steps: acquiring an image data pair to be hidden, wherein the image data pair to be hidden comprises a secret image to be hidden and an overlay image to be hidden; respectively obtaining first energy hidden variable pairs corresponding to the energy data pairs to be hidden based on the image data pairs to be hidden through the component energy extraction model; obtaining hidden variable pairs based on the first energy hidden variable pairs through a reversible neural network; the hidden variable pair obtains a hidden image through an image generation module; acquiring a noise image; obtaining a second energy hidden variable pair based on the hidden image and the noise image through the component energy extraction model; obtaining a revealing hidden variable pair based on the second energy hidden variable pair through a reversible neural network; the hidden variable pair is disclosed, and a secret image and an overlay image are obtained through an image generation module, so that hidden writing and disclosure can be accurately realized.
Description
Technical Field
The disclosure belongs to the field of image encryption, and particularly relates to a dual-task image steganography method based on component energy.
Background
Image steganography is a process of hiding a secret image containing sensitive information into an overlay image without sensitive information to generate a steganography image, and allowing only an informed receiver to obtain the overlay image and the secret image through image disclosure. For information communication security, it is often desirable that the steganographic image and the overlay image look visually indistinct. Unlike image encryption and conventional steganography, image steganography has higher requirements on the information capacity of steganographic images, the invisibility of secret images, and the security of information processing procedures, and thus the task is more challenging. The image steganography has high practicability, and the steganography image containing secret image information is required to be indistinguishable from the coverage image, so that the method is more suitable for encryption communication, and secret data communication safety is guaranteed.
The traditional image steganography method generally consists of two sub-functional modules, namely image steganography and image disclosure. The method is mainly realized by the following two technical means, namely, secret image information is hidden in the coverage image through complex replacement and conversion algorithms, but with the improvement of the performance of computing resources, the replacement and conversion algorithms cannot guarantee that the steganographic image has reasonable resolution and capacity, and the safety of data cannot be guaranteed. And secondly, the two functional modules are loosely coupled and connected based on the machine learning deep neural network, and the two functional modules are respectively provided with independent parameter sets, so that the data quality problems such as image color distortion, texture flaws and the like can be caused. In addition, the conventional image steganography method hardly considers security problems, so that hidden information is easily detected.
The method based on the reversible neural network uses the same deep neural network function module to perform bidirectional actions to respectively complete the image steganography and image disclosure processes. The image revealing is the reverse process of image hiding, so the functional module can obtain all network parameters required for completing the functions of image hiding and image revealing by only performing one training, which is the fundamental difference between the method and the traditional method. At present, the method reaches the most advanced level in terms of secret image recovery accuracy, hidden security and invisibility. However, the method has high model coupling degree, weak expansibility and limited characterization capability on the input image.
Disclosure of Invention
The present disclosure has been made in view of the above-mentioned needs of the prior art, and an object of the present disclosure is to provide a dual-task image steganography method based on component energy to precisely steganographically and reveal secret images.
In order to solve the above problems, the technical solution provided by the present disclosure includes:
provided is a dual-task image steganography method based on component energy, comprising: acquiring an image data pair to be hidden, wherein the image data pair to be hidden comprises a secret image to be hidden and an overlay image to be hidden; respectively obtaining component energy pairs of the to-be-hidden energy data pairs based on the to-be-hidden image data pairs through a component energy extraction model, wherein the energy extraction model comprises a two-dimensional convolution layer, a full connection layer and global average pooling, and the component energy of the to-be-hidden energy data pairs comprises the component energy of the to-be-hidden secret image and the component energy of the to-be-hidden overlay image; obtaining hidden variable pairs based on component energy pairs of the energy data pairs to be hidden by a reversible neural network; the hidden variable pair obtains a hidden image through an image generation module; acquiring a noise image; obtaining component energies of the steganographic image and the noise image based on the steganographic image and the noise image through the component energy extraction model respectively; obtaining a hidden variable pair based on the component energy of the hidden image and the component energy of the noise image through a reversible neural network; the disclosed hidden variable pair obtains a secret image and an overlay image through the image generation module.
Preferably, when the pair of image data to be steganographically is acquired, the physical scale information of the image in the pair of image data to be steganographically is acquired at the same time.
Preferably, the parameters of the component energy extraction model are self-adjusted according to the input physical scale information of the image to be hidden.
Preferably, the secret image to be hidden obtains an energy hidden variable set of the secret image to be hidden through a component energy extraction modelZ n ={z 1 ，z 2 ，…，z n And } wherein,z i =Enc θ (x secreti )，z i a hidden variable representing the i-th secret image to be hidden,Enc θ () Representing the encoding process of the secret image to be steganographically,x secreti represent the firstiA plurality of secret images to be steganographically; the to-be-hidden-written coverage image obtains an energy hidden variable set of the to-be-hidden-written coverage image through a component energy extraction modelZ' n ={z' 1 ，z' 2 ，…，z' n And } wherein,z' i =Enc θ '(x coveri )，z' i a hidden variable representing the i-th to-be-hidden overlay image,Enc θ '() Representing the encoding process of the overlay image to be steganographically,x coveri represent the firstiThe overlay image is to be steganographically.
Preferably, the loss function of the method comprises an energy loss function and a generation loss function.
Preferably, the energy loss function is expressed as:
wherein,the foreground information energy hidden variable representation representing the steganographic image,background information representing steganographic image an energy hidden variable representation,the foreground information energy hidden variable representation representing the secret image,background information representing the secret image is represented by an energy hidden variable.
Preferably, the generating loss function is expressed as:
wherein,Ɛ、μandis the weight in the optimization process,x cover representing an overlay image to be steganographically,x' cover an overlay image is represented and is displayed,x secret representing the secret image to be steganographically,x' secret a secret image is represented and,Mrepresenting the total number of frames of the image of each input model in the algorithm model loss iteration process,is the firstiThe frame input covers the average value of the image,is the revealing process ofiThe frame generates an average value of the overlay image,is the firstiThe frame input covers the variance of the image,is the revealing process ofiThe frame generates a variance of the overlay image,represent the firstiFrame input overlay image and reveal process NoiThe frame generates a covariance between the overlay images,is the firstiThe frame is input with an average value of the secret image,is the revealing process ofiThe frame generates an average value of the secret image,is the firstiThe variance of the frame input secret image isReveal Process NoiThe frame generates a variance of the secret image,represent the firstiFrame input secret image and revealing processiThe frame generates a covariance between the secret images,c 1 andc 2 is a constant for ensuring learning fairness.
Compared with the prior art, the invention provides a dual-task deep neural network model based on component energy, and the network consists of three parts: a component energy extraction model, a reversible neural network, and an image generation module. The component energy extraction model adopts an energy equation to encode an input image into combined energy components respectively representing a foreground and a background. The reversible neural network is composed of a deep neural network functional module composed of a cyclic neural network with a memory function and a one-dimensional convolutional neural network layer, so that the combined energy information is hidden, and the safety of the hidden information is further enhanced. Aiming at the functional requirements of image steganography and image disclosure, the image generation module respectively completes the decoding process of secret energy obtained by hidden transmission and outputs a target image. The data information transmitted among the three modules is the image characteristics containing high-level semantic information. And secondly, the dimension of energy information generated by the component energy extraction model can be adaptively adjusted according to the physical scale of the input image, the function of the image generation module is realized depending on an energy component decomposition network, and the model structure and the functional coupling degree are enhanced. Finally, the constructed reversible neural network function implementation mainly depends on a one-dimensional convolutional neural network layer, the structure is simple, the calculated amount is low, the initial layer structure can be adaptively adjusted along with the transformation of the image resolution, and the reversible network is lossless in information, so that the model can retain the detailed information of input data, and the adaptive energy adjustment information characterization capability is strong.
Drawings
In order to more clearly illustrate the embodiments of the present description or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments described in the embodiments of the present description, and other drawings may be obtained according to these drawings for a person having ordinary skill in the art.
FIG. 1 is a flow chart showing the steps of a method for steganography on a dual-task image based on component energy according to the present invention;
FIG. 2 is a schematic illustration of a steganography process of a dual task image steganography method based on component energy in accordance with the present invention;
FIG. 3 is a schematic diagram of a method for capturing a dual-task image based on component energy according to the present invention.
Detailed Description
For the purposes of making the objects, technical solutions and advantages of the embodiments of the present application more clear, the technical solutions of the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is apparent that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are within the scope of the present disclosure.
In describing the embodiments of the present disclosure, it should be noted that, unless explicitly stated and limited otherwise, the term "connected" should be construed broadly, for example, it may be a fixed connection, a detachable connection, or an integral connection, a mechanical connection, an electrical connection, a direct connection, or an indirect connection via an intermediary. The specific meaning of the terms in this disclosure will be understood by those of ordinary skill in the art as the case may be.
The terms "top," "bottom," "above … …," "below," and "on … …" are used throughout the description to refer to the relative positions of components of the device, such as the relative positions of the top and bottom substrates inside the device. It will be appreciated that the devices are versatile, irrespective of their orientation in space.
For the purpose of facilitating an understanding of the embodiments of the present application, reference will now be made to the following description of specific embodiments, taken in conjunction with the accompanying drawings, in which the embodiments are not intended to limit the embodiments of the present application.
The present embodiment provides a dual task image steganography method based on component energy, as shown in fig. 1-3.
As shown in fig. 1, the dual-task image steganography method based on component energy includes:
s1, acquiring a to-be-hidden image data pair, wherein the to-be-hidden image data pair comprises a to-be-hidden secret image and a to-be-hidden overlay image.
The secret image to be steganographic and the overlay image to be steganographic are from a secret image dataset and an overlay image dataset, respectively. Specifically, randomly extracting an image from each of the secret image data set and the overlay image data set to form a pair of image data to be hidden, wherein the pair of image data to be hidden comprises a secret image to be hiddenx secret And an overlay image to be steganographically overlaidx cover . In addition, in the above process, the secret image to be steganographically and the physical scale information of the overlay image to be steganographically are simultaneously acquired.
S2, respectively obtaining component energy pairs of the to-be-hidden energy data pairs based on the to-be-hidden image data pairs through a component energy extraction model, wherein the energy extraction model comprises a two-dimensional convolution layer, a full connection layer and global average pooling, and the component energy of the to-be-hidden energy data pairs comprises the component energy of the to-be-hidden secret image and the component energy of the to-be-hidden overlay image.
Inputting the acquired pair of image data to be hidden into the component energy extraction model E to respectively extract the secret images to be hiddenx secret And an overlay image to be steganographically overlaidx cover Component energy of (a). The component energy extraction model E comprises a neural network model which comprises a two-dimensional convolution layer, a full connection layer and a global average pooling layer. The component energy extraction model adjusts parameters of the component energy extraction model E according to the information of the input image,the information includes previously acquired physical scale information. Further, when the input of the component energy extraction model is a secret image to be steganographically, the component energy extraction model is E 1 The method comprises the steps of carrying out a first treatment on the surface of the When the input of the component energy extraction model is an image to be steganographically covered, the component energy extraction model is E 2 。E 1 And E is 2 And respectively carrying out self-adaptive adjustment according to the obtained physical scale information of the secret image to be hidden and the physical scale information of the coverage image to be hidden, so that the dimension of the energy variable which is extracted from each image and represents the foreground information and the background information is the same as the dimension of the image.
The component energy extraction model can realize a learnable encoder function, and encodes an input secret image to be hidden into a set of secret image energy hidden variables to be hidden, wherein the set of secret image energy hidden variables can represent component energy of the secret image to be hiddenZ n ={z 1 ，z 2 ，…，z n And } wherein,z i = Enc θ (x secreti )，z i a hidden variable representing the i-th secret image to be hidden,Enc θ () Representing the encoding process of the secret image to be steganographically,x secreti represent the firstiA plurality of secret images to be steganographically; encoding an input to-be-steganographically overlaid image into a set of to-be-steganographically overlaid image energy hidden variables capable of representing constituent energies thereofZ' n ={z' 1 ，z' 2 ，…，z' n And } wherein,z' i =Enc θ '(x coveri )，z' i a hidden variable representing the i-th to-be-hidden overlay image,Enc θ '() Representing the encoding process of the overlay image to be steganographically,x coveri represent the firstiThe overlay image is to be steganographically. Gathering hidden variablesZ n Conversion to component energy of secret image to be steganographicallyEn 1 Gathering hidden variablesZ' n Conversion to component energy of an image to be steganographically overlaidEn 2 。
Thereafter through unsupervised learningL 1 The parameter optimization process of norms minimizes the model parameter θ:
L 1 (θ) For a norm where the parameter theta is minimized,arg x min() In order to minimize the function of the parameter,E θ () The parameter set in the energy extraction model for the component isθA function of the time of day and,x i for the ith secret image to be steganographically or the thiThe overlay image is to be steganographically.
S3, obtaining hidden variable pairs based on component energy pairs of the energy data pairs to be hidden through a reversible neural network.
Component energy of secret image to be hiddenEn 1 And component energy of an overlay image to be steganographicallyEn 2 At the same time input to the reversible neural network Inver through supervised model parametersMinimizing energy constraint learning for forward energy concealment associated transmission to obtain a concealment variable pairR n AndR' n expressed as:
wherein,representation of post-concealment associationiZhang Daiyin the foreground energy of the secret image or the overlay image to be steganographically,representing the first corresponding foreground energy after the hidden associationiZhang Daiyin write the background energy of the secret image or the image to be steganographically overlaid. Since the two images are included after the concealment association, the two images areiThe value of (2) is 1 or 2.
By optimizing, the model is forced to bias towards learning foreground features of the image, thereby concentrating energy in the foreground information. In particular, a reversible neural networkInverFrom the following componentsMAnd the hidden sub-functional modules are used for completing the energy hiding process. For the firstjA hidden sub-function module for inputting the firstjComponent energy of individual secret images to be steganographicallyAnd (d)jComponent energy of individual to-be-steganographically overlaid imagesThe output isAndexpressed as:
wherein,βmultiplying a sigmoid function by a constant factor,the point-to-point operation is indicated,ρ()，andthe residual model ResNet functional block proposed in the countermeasure network is generated with enhanced super resolution. The output of the last hidden sub-function module isAndthe image is input into an image generator to complete image generation. In order to simplify the representation of the drawing,，。
further, the neural network model includes 8 hidden sub-function algorithm modules with the same structure, and the 8 algorithm modules are arranged in ascending order, namely [ y ] when image steganography is performed 1 ；y 2 ；…；y 8 ]. In addition, each ResNet functional component comprises 5 one-dimensional convolutional neural network layers and 1-layer cyclic neural network layer LSTM, each functional component extracts high-level energy features with different visual receptive field scales, the features of the 5 one-dimensional convolutional neural network layers are cascaded, the receptive field of the energy features is enhanced, and therefore the correlation granularity of the energy features is finer, nonlinearity is further increased from the algorithm level, and the energy feature expression capability of the reversible neural network is improved.
And S4, the hidden variable pair obtains a hidden image through an image generation module.
Hiding hidden variables obtained in the processR n AndR' n input image generation moduleGIn generating a steganographic imagex stego Sum-difference imagex error To achieve image steganography. Wherein when the input image is an energy hidden variable of the secret image to be hidden, the image generation module G 1 The method comprises the steps of carrying out a first treatment on the surface of the When the input image is to be displayedWhen the energy hidden variable of the image is covered by the hidden writing, the image generating module G 2 Two hidden variables respectively pass through G 1 And G 2 Obtaining a steganographic image and a difference image, and G 1 And G 2 All are realized by the supervised model parameter optimization process through random gradient descent.
S5, acquiring a noise image.
Acquiring randomly generated noise image of the same scale as the steganographic image obtained abovex noise 。
S6, obtaining component energy of the steganographic image and the noise image based on the steganographic image and the noise image through the component energy extraction model.
Inputting the steganographic image and the noise image into the component energy extraction model E to extract steganographic images respectivelyx stego And noise imagex noise Is a hidden energy variable of (a). And the component energy extraction model performs self-adaptive adjustment according to the physical scale of the steganographic image and the noise image, namely, the component energy is mapped in the same energy conversion mode as S2. Component energy of the steganographic image and the noise image is finally obtainedEn' 1 AndEn' 2 。
s7, obtaining the hidden variable pair based on the component energy of the hidden image and the component energy of the noise image through the reversible neural network.
Component energy of the steganographic imageEn' 1 And component energy of noise imageEn' 2 And simultaneously inputting the energy into a reversible neural network Inver, and carrying out reverse energy hiding association transmission by the same energy hiding method as S3 to obtain the hidden variable pair. The 8 algorithm modules in the current reversible neural network are arranged in descending order, i.e., [ y ] 8 ；y 7 ；…；y 1 ]. The hidden variable pair is obtained through the processT n AndT' n 。
s8, the hidden variable pair is revealed, and a secret image and an overlay image are obtained through the image generation module.
Hiding hidden variables obtained in the processT n AndT' n input image generation moduleGIn (1) generating a secret imagex' secret Overlay imagex' cover To complete the reverse image disclosure. The image generation module in this step shares the same functional module as the image generation module in S3.
Through the above process, as shown in fig. 2 and 3, the same complete method model is used to realize image steganography and image disclosure, and parameters of each functional module of the algorithm model are adjusted through training of the deep learning model, so that the model function is gradually optimized, and the dual tasks of image steganography and image disclosure are completed.
Further, in the above process, the energy generation loss function isL EG And constraining the overall end-to-end adjustment and optimization of the parameters of the model, wherein the energy generation loss function comprises an energy loss function and a generation loss function, and is expressed as:
L
EG
=L
energy
+L
generation
wherein the method comprises the steps of，Ɛ、μAndis the weight in the optimization process,x cover representing an overlay image to be steganographically,x' cover an overlay image is represented and is displayed,x secret representing the secret image to be steganographically,x' secret a secret image is represented and,Mrepresenting the total number of frames of the image of each input model in the algorithm model loss iteration process,is the firstiThe frame input covers the average value of the image,is the revealing process ofiThe frame generates an average value of the overlay image,is the firstiThe frame input covers the variance of the image,is the revealing process ofiThe frame generates a variance of the overlay image,represent the firstiFrame input overlay image and reveal process NoiThe frame generates a covariance between the overlay images,is the firstiThe frame is input with an average value of the secret image,is the revealing process ofiThe frame generates an average value of the secret image,is the firstiVariance of frame input secret imageIs thatReveal Process NoiThe frame generates a variance of the secret image,represent the firstiFrame input secret image and revealing processiThe frame generates a covariance between the secret images,c 1 andc 2 is a constant for ensuring learning fairness.
The foregoing embodiments have been provided for the purpose of illustrating the general principles of the present application, and are not meant to limit the scope of the invention, but to limit the scope of the invention.
Claims (7)
1. A method for steganography of a dual-task image based on component energy, comprising:
acquiring an image data pair to be hidden, wherein the image data pair to be hidden comprises a secret image to be hidden and an overlay image to be hidden;
respectively obtaining component energy pairs of the to-be-hidden energy data pairs based on the to-be-hidden image data pairs through a component energy extraction model, wherein the energy extraction model comprises a two-dimensional convolution layer, a full connection layer and global average pooling, and the component energy of the to-be-hidden energy data pairs comprises the component energy of the to-be-hidden secret image and the component energy of the to-be-hidden overlay image; the component energy extraction model encodes the input secret image to be hidden into a set Z of secret image energy hidden variables capable of representing the component energy thereof n ＝{z 1 ，z 2 ，…，z n -wherein z i ＝Enc θ (x secreti )，z i Hidden variable, enc, representing the ith secret image to be hidden θ () Representing a plaiting of a secret image to be steganographicallyCode process, x secreti Representing an ith secret image to be steganographically; encoding an input to-be-steganographically overlaid image into a to-be-steganographically overlaid image energy hidden variable set Z 'capable of representing component energy thereof' n ＝{z' 1 ，z' 2 ，…，z' n Z 'where' i ＝Enc θ '(x coveri )，z' i Hidden variables, enc, representing the ith to-be-hidden overlay image θ ' () denotes the encoding process of the image to be covered by steganography, x coveri Representing the ith to-be-hidden overlay image, aggregating hidden variables Z n En converted into component energy of secret image to be steganographically 1 Collecting hidden variables Z' n En converted into component energy of an overlay image to be steganographically 2 ；
Obtaining hidden variable pairs based on component energy pairs of the energy data pairs to be hidden by a reversible neural network;
the hidden variable pair obtains a hidden image through an image generation module;
acquiring a noise image;
obtaining component energies of the steganographic image and the noise image based on the steganographic image and the noise image through the component energy extraction model respectively;
obtaining a hidden variable pair based on the component energy of the hidden image and the component energy of the noise image through a reversible neural network;
the disclosed hidden variable pair obtains a secret image and an overlay image through the image generation module.
2. A dual task image steganography method based on component energy according to claim 1, characterized in that when the pair of image data to be steganographically is acquired, the physical scale information of the image in the pair of image data to be steganographically is acquired at the same time.
3. The method for steganography of a dual-task image based on component energy according to claim 2, wherein parameters of the component energy extraction model are self-adjusted according to the input physical scale information of the image to be steganographically.
4. The method of claim 1, wherein the method further comprises the step of,
the secret image to be hidden is subjected to component energy extraction model to obtain an energy hidden variable set Z of the secret image to be hidden n ＝{z 1 ，z 2 ，...，z n -wherein z i ＝Enc θ (x secreti )，z i Hidden variable, enc, representing the ith secret image to be hidden θ () Representing the encoding process of a secret image to be steganographically, x secreti Representing an ith secret image to be steganographically;
the to-be-hidden-written coverage image obtains an energy hidden variable set Z 'of the to-be-hidden-written coverage image through a component energy extraction model' n ＝{z′ 1 ，z′ 2 ，...，z′ n Z 'where' i ＝Enc θ ′(x coveri )，z′ i Hidden variables, enc, representing the ith to-be-hidden overlay image θ ' () denotes the encoding process of the image to be covered by steganography, x coveri Representing the ith to-be-steganographically overlaid image.
5. A method of bi-task image steganography based on component energy according to claim 1, characterized in that the loss function of the method comprises an energy loss function and a generation loss function.
6. The method of component energy-based bicuspid image steganography of claim 5, wherein the energy loss function is expressed as:
L energy ＝||En′ 1_pos -En′ 1_neg ||+||En′ 2_pos -En′ 2_neg ||
wherein En' 1pos Foreground information energy hidden variable representation, en ', representing a steganographic image' 1_neg Background information representative of a steganographic image energy hidden variable representation, en' 2_pos Foreground information energy hidden variable representation representing secret image，En′ 2_neg Background information representing the secret image is represented by an energy hidden variable.
7. The method of component energy-based bicuspid image steganography of claim 5, wherein the generating a loss function is expressed as:
L generation ＝εL g +μL re +θL ssim
L g ＝||x′ cover -x cover ||+||x′ secret -x secret ||
L re ＝||x′ cover -x cover || 2 +||x′ secret -x secret || 2
where ε, μ and θ are weights in the optimization process, x cover Representing an overlay image to be steganographically, x' cover Representing an overlay image, x secret Representing a secret image to be steganographically, x' secret Represents a secret image, M represents the total number of frames of the image of each input model in the algorithm model loss iteration process,is the average value of the i-th frame input overlay image, is->Is revealing that process i frame generates the average value of the overlay image,/->Is the variance of the i-th frame input overlay image,/>Is the disclosure of the variance of the process i frame generation overlay image，/>Representing covariance between an i-th frame input overlay image and a reveal process i-th frame generation overlay image,/for>Is the average value of the input secret image of the ith frame, is->Is revealing the average value of the process i frame generated secret image,/-, and>is the variance of the input secret image of the ith frame, is +.>Revealing Process the ith frame generates the variance, delta, of the secret image secret，secret i′ Representing the covariance between the input secret image of the ith frame and the generated secret image of the ith frame of the revealing process, c 1 And c 2 Is a constant for ensuring learning fairness.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211619646.9A CN115731089B (en) | 2022-12-16 | 2022-12-16 | Dual-task image steganography method based on component energy |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211619646.9A CN115731089B (en) | 2022-12-16 | 2022-12-16 | Dual-task image steganography method based on component energy |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115731089A CN115731089A (en) | 2023-03-03 |
| CN115731089B true CN115731089B (en) | 2024-02-23 |
Family
ID=85301438
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211619646.9A Active CN115731089B (en) | 2022-12-16 | 2022-12-16 | Dual-task image steganography method based on component energy |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115731089B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107992466A (en) * | 2017-10-09 | 2018-05-04 | 中国电子科技集团公司第二十八研究所 | A kind of such poems of the Song Dynasty carrier text information hiding method based on clause control, system |
| CN109859091A (en) * | 2018-12-24 | 2019-06-07 | 中国人民解放军国防科技大学 | Image steganography detection method based on Gabor filtering and convolutional neural network |
| CN111598762A (en) * | 2020-04-21 | 2020-08-28 | 中山大学 | Generating type robust image steganography method |
| CN112581345A (en) * | 2020-12-04 | 2021-03-30 | 华南理工大学 | Image block-based image steganalysis method, system, device and medium |
| CN112884630A (en) * | 2021-02-04 | 2021-06-01 | 南开大学 | Reversible neural network-based high-capacity image steganography and recovery method |
| 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 |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11080809B2 (en) * | 2017-05-19 | 2021-08-03 | Google Llc | Hiding information and images via deep learning |
| US11526959B2 (en) * | 2019-11-15 | 2022-12-13 | Booz Allen Hamilton Inc. | System and method for digital image steganography detection using an ensemble of neural spatial rich models |
-
2022
- 2022-12-16 CN CN202211619646.9A patent/CN115731089B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107992466A (en) * | 2017-10-09 | 2018-05-04 | 中国电子科技集团公司第二十八研究所 | A kind of such poems of the Song Dynasty carrier text information hiding method based on clause control, system |
| CN109859091A (en) * | 2018-12-24 | 2019-06-07 | 中国人民解放军国防科技大学 | Image steganography detection method based on Gabor filtering and convolutional neural network |
| CN111598762A (en) * | 2020-04-21 | 2020-08-28 | 中山大学 | Generating type robust image steganography method |
| CN112581345A (en) * | 2020-12-04 | 2021-03-30 | 华南理工大学 | Image block-based image steganalysis method, system, device and medium |
| CN112884630A (en) * | 2021-02-04 | 2021-06-01 | 南开大学 | Reversible neural network-based high-capacity image steganography and recovery method |
| 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)
| Title |
|---|
| 《一种基于ＦａｓｔＩＣＡ的图像隐写提取算法》;李祁云等;《计算机应用》;第27卷(第5期);全文 * |
| 《基于深度学习的构造式信息隐藏方法研究》;苏日成;《中国优秀硕士学位论文全文库》;全文 * |
| 图像隐写分析算法研究概述;彭伟;胡宁;胡璟璟;;计算机科学(第S1期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115731089A (en) | 2023-03-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2022267641A1 (en) | Image defogging method and system based on cyclic generative adversarial network | |
| CN111145131B (en) | Infrared and visible light image fusion method based on multiscale generation type countermeasure network | |
| CN111243066B (en) | Facial expression migration method based on self-supervision learning and confrontation generation mechanism | |
| CN109191382B (en) | Image processing method, device, electronic equipment and computer readable storage medium | |
| CN110378844B (en) | Image blind motion blur removing method based on cyclic multi-scale generation countermeasure network | |
| CN110084734B (en) | Big data ownership protection method based on object local generation countermeasure network | |
| CN110310221B (en) | Multi-domain image style migration method based on generation countermeasure network | |
| CN110543846B (en) | Multi-pose face image obverse method based on generation countermeasure network | |
| CN110084121A (en) | Implementation method based on the human face expression migration for composing normalized circulation production confrontation network | |
| CN110706302A (en) | System and method for text synthesis image | |
| CN111861880B (en) | Image super-fusion method based on regional information enhancement and block self-attention | |
| CN110599395A (en) | Target image generation method, device, server and storage medium | |
| CN115484410B (en) | Event camera video reconstruction method based on deep learning | |
| CN114240810B (en) | Face sketch-photo synthesis method based on progressive generation network | |
| CN107545277A (en) | Model training, auth method, device, storage medium and computer equipment | |
| CN109949217B (en) | Video super-resolution reconstruction method based on residual learning and implicit motion compensation | |
| CN112115490A (en) | Carrier image synthesis steganography method based on GAN | |
| CN111726472A (en) | Image anti-interference method based on encryption algorithm | |
| CN116645598A (en) | Remote sensing image semantic segmentation method based on channel attention feature fusion | |
| CN113421185B (en) | StyleGAN-based mobile terminal face age editing method | |
| CN115731089B (en) | Dual-task image steganography method based on component energy | |
| CN117173025A (en) | Single-frame image super-resolution method and system based on cross-layer mixed attention transducer | |
| CN117274059A (en) | Low-resolution image reconstruction method and system based on image coding-decoding | |
| CN115393176A (en) | Portrait cartoon method, system and electronic equipment | |
| JP2023092185A (en) | Image processing apparatus, learning method, and program |
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 |