CN113744113A - Image reversible information hiding method driven by convolutional neural network - Google Patents
Image reversible information hiding method driven by convolutional neural network Download PDFInfo
- Publication number
- CN113744113A CN113744113A CN202111055396.6A CN202111055396A CN113744113A CN 113744113 A CN113744113 A CN 113744113A CN 202111055396 A CN202111055396 A CN 202111055396A CN 113744113 A CN113744113 A CN 113744113A
- Authority
- CN
- China
- Prior art keywords
- image
- pixel
- original
- predictor
- optimal prediction
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 49
- 238000013527 convolutional neural network Methods 0.000 title claims abstract description 39
- 230000002441 reversible effect Effects 0.000 title claims abstract description 31
- 230000002194 synthesizing effect Effects 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 3
- 229910003460 diamond Inorganic materials 0.000 claims description 2
- 239000010432 diamond Substances 0.000 claims description 2
- 238000011084 recovery Methods 0.000 claims description 2
- 230000007547 defect Effects 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 3
- 235000002566 Capsicum Nutrition 0.000 description 2
- 241000282537 Mandrillus sphinx Species 0.000 description 2
- 241000758706 Piperaceae Species 0.000 description 2
- 230000000052 comparative effect Effects 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- OLBCVFGFOZPWHH-UHFFFAOYSA-N propofol Chemical compound CC(C)C1=CC=CC(C(C)C)=C1O OLBCVFGFOZPWHH-UHFFFAOYSA-N 0.000 description 2
- 229960004134 propofol Drugs 0.000 description 2
- -1 carrier Substances 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 239000000306 component Substances 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 238000013075 data extraction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 239000004615 ingredient Substances 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T1/00—General purpose image data processing
- G06T1/0021—Image watermarking
-
- 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
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Processing (AREA)
- Editing Of Facsimile Originals (AREA)
Abstract
The invention discloses an image reversible information hiding method driven by a convolutional neural network. The image reversible information hiding method combines an image pixel predictor based on a convolutional neural network and a traditional image pixel predictor to construct an image pixel predictor meeting the protection condition of optimal prediction performance. The image reversible information hiding method makes up the defect of using a single image pixel predictor, better utilizes the spatial correlation of image pixels, and improves the pixel prediction precision and the secret-loaded image quality of the image reversible information hiding method.
Description
Technical Field
The invention relates to the technical field of information hiding, in particular to an image reversible information hiding method driven by a convolutional neural network.
Background
Image reversible information hiding, which is a technology capable of reversibly embedding data into a carrier image, plays an important role in the fields of archive management, forensic, medical image processing, and the like.
Reversibility means that the original carrier image can be recovered losslessly after the embedded data is extracted. The core problem of image invertible information concealment is to construct a steep prediction error histogram with an image pixel predictor.
However, most of the conventional image reversible information hiding methods only use a single image pixel predictor, cannot fully utilize the spatial correlation of the image pixels, and have limited pixel prediction accuracy.
Disclosure of Invention
The invention aims to provide an image reversible information hiding method driven by a convolutional neural network, which improves the pixel prediction precision and the secret-carrying image quality of the image reversible information hiding method.
The purpose of the invention is realized by the following technical scheme:
a convolutional neural network driven image reversible information hiding method comprises the following steps:
segmenting the carrier image into a first original pixel set and a second original pixel set which are not overlapped;
constructing a first image pixel predictor meeting a set optimal prediction performance protection condition by using the second original pixel set, performing optimal prediction on the first original pixel set, and reversibly embedding data in the first original pixel set to generate a first secret-carrying pixel set; using the first secret-carrying pixel set to make a second image pixel predictor meeting the set optimal prediction performance protection condition, performing optimal prediction on the second original pixel set, and reversibly embedding data in the second original pixel set to generate a second secret-carrying pixel set;
and synthesizing the first secret-carrying pixel set and the second secret-carrying pixel set into a secret-carrying image.
According to the technical scheme provided by the invention, the image pixel predictor meeting the optimal prediction performance protection condition is constructed by combining the image pixel predictor based on the convolutional neural network and the traditional image pixel predictor. The image reversible information hiding method makes up the defect of using a single image pixel predictor, better utilizes the spatial correlation of image pixels, and improves the pixel prediction precision and the secret-loaded image quality of the image reversible information hiding method.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on the drawings without creative efforts.
Fig. 1 is a flowchart of an image reversible information hiding method driven by a convolutional neural network according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a convolutional neural network-driven image reversible information hiding method according to an embodiment of the present invention;
FIG. 3 is a block diagram of a convolutional neural network-based image pixel predictor provided by an embodiment of the present invention;
FIG. 4 is a schematic illustration of a carrier image used in a comparative experiment provided by an embodiment of the present invention;
fig. 5 is a prediction error histogram obtained by performing pixel prediction on an image Lena by using different image reversible information hiding methods according to an embodiment of the present invention;
fig. 6 is a graph showing a relationship between PSNR values of the secret-carrying image and embedding rates according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention are clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present invention without making any creative effort, shall fall within the protection scope of the present invention.
The terms that may be used herein are first described as follows:
the terms "comprising," "including," "containing," "having," or other similar terms of meaning should be construed as non-exclusive inclusions. For example: including a feature (e.g., material, component, ingredient, carrier, formulation, material, dimension, part, component, mechanism, device, process, procedure, method, reaction condition, processing condition, parameter, algorithm, signal, data, product, or article of manufacture), is to be construed as including not only the particular feature explicitly listed but also other features not explicitly listed as such which are known in the art.
The method for hiding reversible information of an image provided by the present invention is described in detail below. Details which are not described in detail in the embodiments of the invention belong to the prior art which is known to the person skilled in the art. Those not specifically mentioned in the examples of the present invention were carried out according to the conventional conditions in the art or conditions suggested by the manufacturer.
Fig. 1 is a flowchart of an image reversible information hiding method driven by a convolutional neural network according to an embodiment of the present invention, and fig. 2 is a schematic diagram, where the method mainly includes the following steps:
and 11, dividing the carrier image into a first original pixel set and a second original pixel set which are not overlapped.
The preferred embodiment of this step is as follows:
1) giving a carrier image with the size of H multiplied by W, and sequentially traversing to obtain an original pixel value y at a position (i, j) in the carrier imagei,jWhere, i 1., H, j 1., W, H, W are the height and width of the carrier image, respectively.
2) The original pixel y of the carrier imagei,jIs divided into a first non-overlapping original pixel set y according to the position of the pixelA＝{yi,jI mod (i + j,2) ═ 0} and a second original set of pixels yB＝{yi,jI mod (i + j,2) ═ 1}, where mod { · } is the remainder operation.
In the embodiment of the invention, a first image pixel predictor meeting the set optimal prediction performance protection condition is constructed by utilizing the second original pixel set, the optimal prediction mode of the first original pixel set is the same as that of a second image pixel predictor meeting the set optimal prediction performance protection condition and constructed by utilizing the first secret-carrying pixel set, and the optimal prediction mode of the second original pixel set is the same; similarly, the manner in which data is reversibly embedded in the first original pixel set to generate the first secret-carrying pixel set is the same as the manner in which data is reversibly embedded in the second original pixel set to generate the second secret-carrying pixel set.
In the following, an example is described in which a first image pixel predictor that satisfies a set optimal prediction performance protection condition is constructed by using the second original pixel set, the first original pixel set is optimally predicted, data is reversibly embedded in the first original pixel set, and a first secret-loaded pixel set is generated, where a preferred embodiment is as follows:
1) using a second original set of pixels yBConstructing a first image pixel predictor meeting the protection condition of the optimal prediction performance, and performing prediction on a first original pixel set yAAnd performing optimal prediction.
The first image pixel predictor and the second image pixel predictor in the embodiment of the invention have the same structure, and both comprise: the image pixel predictor consists of a diamond-shaped image pixel predictor, a horizontal image pixel predictor, a vertical image pixel predictor and an image pixel predictor based on a convolutional neural network;
for the first image pixel predictor, the prediction result is expressed as:
wherein ,
and
respectively obtaining predicted values calculated by a diamond image pixel predictor, a horizontal image pixel predictor, a vertical image pixel predictor and an image pixel predictor based on a convolutional neural network; y isi,jDenotes the original pixel value at the (i, j) position in the carrier image, i 1.. and H, j 1.. and W, H, W are the height and width of the carrier image, respectively.
In the embodiment of the present invention, the diamond-shaped image pixel predictor, the horizontal image pixel predictor and the vertical image pixel predictor are conventional image pixel predictors, and thus are not described in detail.
As shown in fig. 3, the convolutional neural network-based image pixel predictor takes a convolutional neural network denoiser as a core network, and models pixel prediction as a reconstruction problem from a low-resolution image to a high-resolution image; after the second original pixel set or the first secret-carrying pixel set is input to an image pixel predictor based on a convolutional neural network, performing bicubic difference, adding noise to generate a corresponding noise image, inputting the noise image to a convolutional neural network de-noising device, and generating a predicted value by using a residual image output by the convolutional neural network de-noising device and the noise image; the convolutional neural network de-noising device comprises a plurality of times of expanding convolution operation, a linear rectification function is added between the first time of expanding convolution operation and the second time of expanding convolution operation, and batch normalization operation and linear rectification function are added between the second time of expanding convolution operation and the penultimate expanding convolution operation.
In the structure shown in fig. 3, s-DConv is an expansion convolution operation (s is an expansion rate of the expansion convolution, and a value range is from 1 to 4), BNorm is a batch normalization operation, ReLU is a linear rectification function, the number of expansion convolution operations in the convolutional neural network de-noising device is only an example, and is not a limitation, and the specific number may be set by a user according to actual conditions or experience.
In the embodiment of the present invention, the optimal predicted performance protection condition is expressed as:
wherein ,
and is
In the optimal prediction process, only the pixels meeting the optimal prediction performance protection condition are predicted, and the prediction is represented as follows:
wherein ,ei,jFor the prediction error (i.e. the difference between the original pixel value and the corresponding predicted value) obtained using the diamond-shaped image pixel predictor, the horizontal image pixel predictor, the vertical image pixel predictor and the image pixel predictor based on the convolutional neural network,
is the optimal prediction error.
In addition, taking the first original pixel set as an example, performing optimal prediction may be expressed as:
wherein ,yARepresenting a first original set of pixels, yi,jRepresenting the original pixel value at the (i, j) position in the carrier image,
to the pixel yi,j∈yAThe optimal predicted value of (a) is,
is the optimal prediction error.
2) Embedding reversible data in a first original set of pixels y using prediction error extension basedAReversibly embedding data to generate a first dense set of pixels
The method comprises the following steps:
first, an extended optimal prediction error is calculated
Expressed as:
wherein ,bkThe k-th data bit to be embedded is, and t is a parameter for controlling the data embedding capacity;
for optimal prediction error, based on optimal prediction
Calculating to obtain; i 1., H, j 1., W, H, W are the height and width, respectively, of the carrier image;
then, based on the optimal prediction
And the extended optimal prediction error
Computing secret-carrying pixels
To representComprises the following steps:
finally, a series of dense pixels is used
Forming a first dense set of pixels
Furthermore, the extended optimal prediction error
Capable of lossless recovery to optimal prediction error
According to the optimal predicted value
And optimal prediction error
The original pixel value y can be calculatedi,jExpressed as:
and step 13, synthesizing the first secret-carrying pixel set and the second secret-carrying pixel set into a secret-carrying image.
In an embodiment of the invention, a first secret-carrying pixel set is combined
And a second dense set of pixels
And synthesizing the secret-carrying image according to the pixel position (i, j)Secret pixel at position (i, j) in a secret image
Expressed as:
wherein mod {. is a remainder operation, i is 1,.. multidot.h, j is 1,.. multidot.w, H, W are the height and width of the secret image, respectively, and the size of the secret image is the same as that of the carrier image.
As described above, the process of processing the carrier image into the secret-carrying image and the process of processing the secret-carrying image into the carrier image are the inverse process described above, as shown in the lower half of fig. 2, the secret-carrying image is firstly divided, and then the second original pixel set and the first original pixel set are sequentially restored by the optimal prediction and the reversible data extraction method based on the optimal prediction error expansion in the manner described in the foregoing step 12, and then are merged into the carrier image.
The above solution provided by the embodiments of the present invention is directed to solving the problem of constructing an optimal image pixel predictor. The image reversible information hiding method combines an image pixel predictor based on a convolutional neural network and a traditional image pixel predictor to construct an image pixel predictor meeting the protection condition of optimal prediction performance. The image reversible information hiding method makes up the defect of using a single image pixel predictor, better utilizes the spatial correlation of image pixels, and improves the pixel prediction precision and the secret-loaded image quality of the image reversible information hiding method.
In order to test the pixel prediction precision and the dense image quality of the image reversible information hiding method provided by the invention, an optimal predictor used by the image reversible information hiding method provided by the invention is marked as a deployed predictor. Four other existing image reversible information hiding methods are selected for comparison, and the image pixel predictors used by them are respectively marked as A, B, C and D. The image super-resolution reconstruction method based on the convolutional neural network comprises the following steps of A, B, C and D, wherein A is a diamond-shaped image pixel predictor, B is a composite predictor with higher pixel prediction precision in a horizontal image pixel predictor and a vertical image pixel predictor, C is an image super-resolution reconstruction pixel predictor, and D is the image pixel predictor based on the convolutional neural network.
The carrier images used in the comparative experiments were Jetplane, Lena, Livingroom, Mandrill, Peppers, and Pirate, as shown in parts (a) to (f) of fig. 4, in that order. The embedding rate Payload is measured in bpi using the number of data Bits (Bits per image) embedded per image. The pixel prediction accuracy is measured using the mean square error MSE of the predicted pixels and the original pixels. The quality of the carrier image is measured by the peak signal-to-noise ratio PSNR (unit: dB) and the carrier image is taken as a reference image.
Fig. 5 shows a prediction error histogram obtained by performing pixel prediction on an image Lena using different image invertible information hiding methods. As shown in part (a), the mean square errors MSE of the image pixel predictor A, B, C, D and the propofol are 24.5584, 43.1920, 15.0220, 13.7655 and 11.4276, respectively, at an embedding rate of 50000 bpi. As shown in part (b), the mean square errors MSE of the image pixel predictor A, B, C, D and the propofol are 25.1875, 45.8287, 14.8854, 13.6084 and 11.4281 at an embedding rate of 100000bpi, respectively. Therefore, the image reversible information hiding method provided by the invention can obtain the highest pixel prediction precision.
Fig. 6 shows a graph of PSNR values of the secret-carrying image versus embedding rate. (a) And (f) carrier images are Jetplane, Lena, Livingrom, Mandrill, Peppers and Pirate in sequence, and the result shown in FIG. 6 shows that the image reversible information hiding method provided by the invention can obtain the best dense image quality at most embedding rates for the given carrier images.
Through the above description of the embodiments, it is clear to those skilled in the art that the above embodiments can be implemented by software, and can also be implemented by software plus a necessary general hardware platform. With this understanding, the technical solutions of the embodiments can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a usb disk, a removable hard disk, etc.), and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods according to the embodiments of the present invention.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (8)
1. An image reversible information hiding method driven by a convolutional neural network is characterized by comprising the following steps:
segmenting the carrier image into a first original pixel set and a second original pixel set which are not overlapped;
constructing a first image pixel predictor meeting a set optimal prediction performance protection condition by using the second original pixel set, performing optimal prediction on the first original pixel set, and reversibly embedding data in the first original pixel set to generate a first secret-carrying pixel set; using the first secret-carrying pixel set to make a second image pixel predictor meeting the set optimal prediction performance protection condition, performing optimal prediction on the second original pixel set, and reversibly embedding data in the second original pixel set to generate a second secret-carrying pixel set;
and synthesizing the first secret-carrying pixel set and the second secret-carrying pixel set into a secret-carrying image.
2. The convolutional neural network driven image invertible information hiding method of claim 1, wherein the segmenting the carrier image into the first and second non-overlapping original pixel sets comprises:
giving a carrier image with the size of H multiplied by W, and sequentially traversing to obtain an original pixel value y at a position (i, j) in the carrier imagei,jWherein, i 1, 1., H, j 1., W, H, W are respectivelyHeight, width of the carrier image;
the original pixel y of the carrier imagei,jIs divided into a first non-overlapping original pixel set y according to the position of the pixelA＝{yi,jI mod (i + j,2) ═ 0} and a second original set of pixels yB＝{yi,jI mod (i + j,2) ═ 1}, where mod { · } is the remainder operation.
3. The convolutional neural network driven image invertible information hiding method according to claim 1, wherein performing optimal prediction on the first original pixel set is represented as:
wherein ,yARepresenting a first original set of pixels, yi,jRepresenting the original pixel value at the (i, j) position in the carrier image,
to the pixel yi,j∈yAThe optimal predicted value of (a) is,
is the optimal prediction error.
4. The method as claimed in claim 1, wherein the second original pixel set is used to construct a first image pixel predictor meeting the protection condition of the set optimal prediction performance, and the optimal prediction of the first original pixel set is performed in the same manner as the optimal prediction of the second original pixel set by using a second image pixel predictor meeting the protection condition of the set optimal prediction performance;
wherein the first image pixel predictor and the second image pixel predictor have the same structure and both comprise: the image pixel predictor consists of a diamond-shaped image pixel predictor, a horizontal image pixel predictor, a vertical image pixel predictor and an image pixel predictor based on a convolutional neural network;
for the first image pixel predictor, the prediction result is expressed as:
wherein ,
and
respectively obtaining predicted values calculated by a diamond image pixel predictor, a horizontal image pixel predictor, a vertical image pixel predictor and an image pixel predictor based on a convolutional neural network; y isi,jDenotes the original pixel value at the position (i, j) in the carrier image, i 1.. and H, j 1.. and W, H, W are the height and width, y, respectively, of the carrier imageBRepresenting a second original set of pixels;
the optimal predicted performance protection condition is expressed as:
wherein ,
and is
In the optimal prediction process, only the pixels meeting the optimal prediction performance protection condition are predicted, and the prediction is represented as follows:
wherein ,ei，jFor prediction errors obtained using a diamond-shaped image pixel predictor, a horizontal image pixel predictor, a vertical image pixel predictor, and a convolutional neural network-based image pixel predictor,
is the optimal prediction error.
5. The method as claimed in claim 4, wherein the convolutional neural network-driven image reversible information hiding method is characterized in that the convolutional neural network-based image pixel predictor takes a convolutional neural network denoiser as a core network,
after the second original pixel set or the first secret-carrying pixel set is input to an image pixel predictor based on a convolutional neural network, performing bicubic difference, adding noise to generate a corresponding noise image, inputting the noise image to a convolutional neural network de-noising device, and generating a predicted value by using a residual image output by the convolutional neural network de-noising device and the noise image;
the convolutional neural network de-noising device comprises a plurality of times of expanding convolution operation, a linear rectification function is added between the first time of expanding convolution operation and the second time of expanding convolution operation, and batch normalization operation and linear rectification function are added between the second time of expanding convolution operation and the penultimate expanding convolution operation.
6. The method according to claim 4 or 5, wherein the first dense set of pixels is generated by reversibly embedding data in the first original set of pixels, and the second dense set of pixels is generated by reversibly embedding data in the second original set of pixels;
wherein reversibly embedding data in the first original set of pixels, the manner of generating a first dense set of pixels comprises:
calculating an extended optimal prediction error
Expressed as:
wherein ,bkThe k-th data bit to be embedded is, and t is a parameter for controlling the data embedding capacity;
for optimal prediction error, based on optimal prediction
Calculating to obtain; i 1., H, j 1., W, H, W are the height and width, respectively, of the carrier image;
according to optimal prediction
And the extended optimal prediction error
Computing secret-carrying pixels
Expressed as:
wherein ,yi,jRepresenting the original pixel value at the (i, j) position in the carrier image;
from a series of dense pixels
Forming a first dense set of pixels
7. The method as claimed in claim 6, wherein the extended optimal prediction error is
Capable of lossless recovery to optimal prediction error
According to the optimal predicted value
And optimal prediction error
The original pixel value y can be calculatedi,jExpressed as:
8. the convolutional neural network-driven image invertible information hiding method of claim 1, wherein synthesizing the first dense set of pixels and the second dense set of pixels into a dense image comprises:
combining a first dense set of pixels
And a second dense set of pixels
And synthesizing a secret image according to the pixel position (i, j), wherein the secret pixel at the position (i, j) in the secret image
Expressed as:
wherein mod {. is a remainder operation, i is 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111055396.6A CN113744113B (en) | 2021-09-09 | 2021-09-09 | Image reversible information hiding method driven by convolutional neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111055396.6A CN113744113B (en) | 2021-09-09 | 2021-09-09 | Image reversible information hiding method driven by convolutional neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113744113A true CN113744113A (en) | 2021-12-03 |
| CN113744113B CN113744113B (en) | 2023-08-29 |
Family
ID=78737739
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111055396.6A Active CN113744113B (en) | 2021-09-09 | 2021-09-09 | Image reversible information hiding method driven by convolutional neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113744113B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117557807A (en) * | 2024-01-11 | 2024-02-13 | 齐鲁工业大学(山东省科学院) | Convolutional neural network image prediction method based on weighted filtering enhancement |
| CN118230075A (en) * | 2024-05-23 | 2024-06-21 | 齐鲁工业大学(山东省科学院) | Digital image reversible information hiding error prediction method based on multiple loss optimization |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060034483A1 (en) * | 2000-08-14 | 2006-02-16 | Au Oscar C | Methods and apparatus for hiding data in halftone images |
| US20080285790A1 (en) * | 2007-05-18 | 2008-11-20 | The Hong Kong University Of Science And Technology | Generalized lossless data hiding using multiple predictors |
| CN110162986A (en) * | 2019-05-10 | 2019-08-23 | 广西赛联信息科技股份有限公司 | Reversible information hidden method based on adjacent pixel prediction model |
-
2021
- 2021-09-09 CN CN202111055396.6A patent/CN113744113B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060034483A1 (en) * | 2000-08-14 | 2006-02-16 | Au Oscar C | Methods and apparatus for hiding data in halftone images |
| US20080285790A1 (en) * | 2007-05-18 | 2008-11-20 | The Hong Kong University Of Science And Technology | Generalized lossless data hiding using multiple predictors |
| CN110162986A (en) * | 2019-05-10 | 2019-08-23 | 广西赛联信息科技股份有限公司 | Reversible information hidden method based on adjacent pixel prediction model |
Non-Patent Citations (1)
| Title |
|---|
| 李亚翔;张显全;俞春强;唐振军;: "采用相邻像素预测的可逆信息隐藏算法", 华侨大学学报(自然科学版), no. 02 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117557807A (en) * | 2024-01-11 | 2024-02-13 | 齐鲁工业大学(山东省科学院) | Convolutional neural network image prediction method based on weighted filtering enhancement |
| CN117557807B (en) * | 2024-01-11 | 2024-04-02 | 齐鲁工业大学(山东省科学院) | Convolutional neural network image prediction method based on weighted filtering enhancement |
| CN118230075A (en) * | 2024-05-23 | 2024-06-21 | 齐鲁工业大学(山东省科学院) | Digital image reversible information hiding error prediction method based on multiple loss optimization |
| CN118230075B (en) * | 2024-05-23 | 2024-08-13 | 齐鲁工业大学(山东省科学院) | Digital image reversible information hiding error prediction method based on multiple loss optimization |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113744113B (en) | 2023-08-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wu et al. | Reversible image watermarking on prediction errors by efficient histogram modification | |
| CN113744113A (en) | Image reversible information hiding method driven by convolutional neural network | |
| JP4775756B2 (en) | Decoding device and program thereof | |
| CN107689026B (en) | Reversible steganography method based on optimal coding | |
| JP2015513151A (en) | Method and apparatus for performing hierarchical super-resolution of input images | |
| Wahed et al. | Reversible data hiding with interpolation and adaptive embedding | |
| Luo et al. | Convolutional neural networks-based stereo image reversible data hiding method | |
| US7463782B2 (en) | Data encoding with an amplitude model and path between the data and corresponding decoding | |
| US7778468B2 (en) | Decoding apparatus, dequantizing method, and program thereof | |
| US20160044328A1 (en) | Apparatus and method of compressing and restoring image using filter information | |
| Kouhi et al. | Prediction error distribution with dynamic asymmetry for reversible data hiding | |
| Hu et al. | Deep inter prediction with error-corrected auto-regressive network for video coding | |
| Chang et al. | An image zooming technique based on vector quantization approximation | |
| JP2024511920A (en) | Video super-resolution based on compression information | |
| JP4945533B2 (en) | Image processing apparatus and image processing method | |
| JP2016535382A (en) | Method and apparatus for constructing an original image estimate from a low quality version and epitome of the original image | |
| Arezoomand et al. | Perceptually optimized loss function for image super-resolution | |
| Wahed et al. | A simplified parabolic interpolation based reversible data hiding scheme | |
| JP4726040B2 (en) | Encoding processing device, decoding processing device, encoding processing method, decoding processing method, program, and information recording medium | |
| JP2009049627A (en) | Method of encoding scalable image, decoding method, encoder, decoder, program thereof, and recording medium thereof | |
| Liu et al. | Hyperspectral image super-resolution employing nonlocal block and hybrid multiscale three-dimensional convolution | |
| Jhong et al. | Exploring more capacity for grayscale-invariance reversible data hiding | |
| TWI397299B (en) | Method of high capacity reversible data hiding scheme using edge prediction error | |
| JP2004289284A (en) | Image processing method, image processing apparatus, and image processing program | |
| JP2009118303A (en) | Image encoding apparatus, image encoding method, image decoding apparatus, and image decoding method |
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 |