CN112711763A - Color image scrambling encryption method based on countermeasure automatic encoder - Google Patents
Color image scrambling encryption method based on countermeasure automatic encoder Download PDFInfo
- Publication number
- CN112711763A CN112711763A CN202011553826.2A CN202011553826A CN112711763A CN 112711763 A CN112711763 A CN 112711763A CN 202011553826 A CN202011553826 A CN 202011553826A CN 112711763 A CN112711763 A CN 112711763A
- Authority
- CN
- China
- Prior art keywords
- image
- matrix
- encoder
- color image
- scrambling
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/60—Protecting data
- G06F21/602—Providing cryptographic facilities or services
-
- 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
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Bioethics (AREA)
- Computer Hardware Design (AREA)
- Computer Security & Cryptography (AREA)
- Facsimile Transmission Control (AREA)
Abstract
The invention discloses a color image scrambling encryption method based on an automatic antagonistic encoder, which relates to the technical field of image processing and comprises the following steps: carrying out matrix conversion on a source image to be processed in advance, and mapping a pixel value to [0,1 ]]Within range, a matrix is acquired
Matrix to be acquired
And inputting the image into a neural network model to obtain encrypted or decrypted image output. The invention realizes the automation and safety of the resolution ratio of the color imageThe scrambling is complete, and the adaptability is strong.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a color image scrambling encryption method based on a countermeasure automatic encoder.
Background
The rapid development of network technology has led to an increasing demand for multimedia transmission of images, sounds and the like capable of visually expressing information, and the multimedia of the images, the sounds and the like may relate to security, confidentiality, privacy and other aspects, and how to ensure the information to be transmitted to a receiver safely and accurately is a basic content of network security research. The image scrambling conceals the original information of an image part by randomly distributing the pixel position, the color and the like of an image, and converts the hidden information into a meaningless image, so that the safety of the image transmission process is improved, and the image scrambling is widely applied to image information hiding and image encryption.
As an important means of image encryption processing, image scrambling aims to scramble pixel values and pixel positions so that a person can be an image with randomly distributed positions or an image with randomly distributed pixels. And the deep learning model can effectively and automatically extract image features which can be used for image encryption or auxiliary image encryption.
Therefore, a color image scrambling encryption method based on a counter-automatic encoder is needed.
The patent CN112036416A for retrieving the invention discloses the technical field of image processing, and particularly relates to an image processing system and method based on deep learning, wherein the system comprises: the image gray histogram processing unit is configured to perform image gray histogram processing on an image to be processed based on preset 3 different gray value ranges, and respectively obtain corresponding gray histogram characteristic distributions in the different gray value ranges; the neural network processing unit is configured to obtain input image feature expressions of images to be processed in10 scales, 3 depths and 3 types based on an image processing model established by pre-training, wherein each scale corresponds to 3 depths and 3 types, multilayer sampling and image truncation are performed on the image feature expressions of each scale, each depth and each type, then full connection is performed, and the fully connected images are output; and the image truncation unit is configured to be provided for the neural network processing unit, perform image truncation on the image to be processed based on the obtained gray level histogram characteristic distribution, and screen out interference of irrelevant information. The method has the advantages of higher accuracy of image recognition and high image processing efficiency. But the adaptability is poor and has certain limitation.
An effective solution to the problems in the related art has not been proposed yet.
Disclosure of Invention
Aiming at the problems in the related art, the invention provides a color image scrambling encryption method based on a countering automatic encoder, so as to overcome the technical problems in the prior related art.
The technical scheme of the invention is realized as follows:
a color image scrambling encryption method based on a countermeasure automatic encoder comprises the following steps:
step S1, the source image to be processed is matrix-converted in advance, and the pixel value is mapped to [0,1 ]]Within range, a matrix is acquired
Step S2, obtaining the matrix
Inputting the image into a neural network model to obtain encrypted or decrypted image output, and comprising the following steps:
step S201, each trained L-th layer convolution kernel WLTraversing through the image matrix and associating it with matrix XLRespective or fill element multiplication;
step S202, all obtained Ws*×X*Adding to determine the new predicted value of the (L +1) th convolutional layer
Step S203, all the predicted values are calculated
The input feature matrixes are collected and combined into a new matrix to form the input feature matrix of the next convolutional layer;
step S204, output matrix X of the last convolution layerL*And converting the image into an image to obtain a final encrypted or decrypted image.
The method comprises the following steps of:
acquiring a 256 × 256 × 3 plaintext or ciphertext image in advance;
the acquired 256 × 256 × 3 plaintext or ciphertext image is converted into a 256 × 256 × 3 matrix.
Wherein, build neural network model, include:
scaling encoder and decoder losses, which scales codec losses, expressed as:
calibration discriminator loss, expressed as:
wherein x is an original image, y is an image with uniformly distributed pixels, and x-pdata(x) Is the distribution of the original image x, y-pdata(y) is uniform distribution, G is encoder, F is decoder and D is discriminator.
Wherein, still include the following step:
the parameters for initializing each convolution sum by performing the neural network model in advance are expressed as:
Wn＝random[wn,1,wn,2,…,wn,j]；
obtaining an encryption key W comprising a parameter W from each convolutional layernComposition, expressed as: w ═ W1,W2,...,Wn]；
Wherein, WnIs a parameter of the nth convolution layer, wn,jIs the jth parameter of the nth convolutional layer;
wherein, the gradient descending process of the calibration encoder is also included, which is expressed as:
wherein, thetajIs the value of the parameter theta at the training jth generation, alpha is the learning rate of the control error variation range, and is V (theta, thetaj) Is passed to the parameter thetajOf the gradient of (c).
Wherein, the step is that each trained Lth layer convolution kernel W is processedLTraversing through the image matrix and associating it with matrix XLCorresponding or filler elements, where L ≦ 24.
The invention has the beneficial effects that:
the invention is based on the color image scrambling encryption method of the anti-automatic encoder, which carries out matrix conversion on a source image to be processed in advance and maps a pixel value to [0,1 ]]Within range, a matrix is acquired
Matrix to be acquired
The image is input to a neural network model to obtain encrypted or decrypted image output, so that automatic and safe scrambling of color image resolution is realized, and adaptability is high.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed in the embodiments will be briefly described 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 without creative efforts.
FIG. 1 is a flow chart of a color image scrambling encryption method based on a countering automatic encoder according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of an encoder based on a color image scrambling encryption method against an automatic encoder according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a decoder based on a color image scrambling encryption method against an automatic encoder according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of a residual block structure of a color image scrambling encryption method based on a countermeasure automatic encoder according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of a discriminator based on a color image scrambling encryption method against an automatic encoder according to an embodiment of the present invention;
FIG. 6 is a schematic diagram of a neural network training flow based on a color image scrambling encryption method against an automatic encoder according to an embodiment of the present invention;
fig. 7 is a schematic diagram of an image encryption and decryption process based on a color image scrambling encryption method against an automatic encoder according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be 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 of the embodiments. All other embodiments that can be derived by one of ordinary skill in the art from the embodiments given herein are intended to be within the scope of the present invention.
According to an embodiment of the present invention, there is provided a color image scrambling encryption method based on a countering automatic encoder.
As shown in fig. 1-7, the color image scrambling encryption method based on the countermeasure automatic encoder according to the embodiment of the present invention includes the following steps:
step S1, the source image to be processed is matrix-converted in advance, and the pixel value is mapped to [0,1 ]]Within range, a matrix is acquired
Step S2, obtaining the matrix
Inputting the image into a neural network model to obtain encrypted or decrypted image output, and comprising the following steps:
step S201, each trained L-th layer convolution kernel WLTraversing through the image matrix and associating it with matrix XLRespective or fill element multiplication;
step S202, all obtained Ws*×X*Adding to determine the new predicted value of the (L +1) th convolutional layer
Step S203, all the predicted values are calculated
The input feature matrixes are collected and combined into a new matrix to form the input feature matrix of the next convolutional layer;
step S204, output matrix X of the last convolution layerL*And converting the image into an image to obtain a final encrypted or decrypted image.
By means of the technical scheme, the color image scrambling encryption method based on the anti-automatic encoder performs matrix conversion on a source image to be processed in advance and maps pixel values to [0,1 ]]Within range, a matrix is acquired
Matrix to be acquired
The image is input to a neural network model to obtain encrypted or decrypted image output, so that automatic and safe scrambling of color image resolution is realized, adaptability is strong, and limitation is wide.
The method comprises the following steps of:
acquiring a 256 × 256 × 3 plaintext or ciphertext image in advance;
the acquired 256 × 256 × 3 plaintext or ciphertext image is converted into a 256 × 256 × 3 matrix.
Wherein, build neural network model, include:
scaling encoder and decoder losses, which scales codec losses, expressed as:
calibration discriminator loss, expressed as:
wherein x is an original image, y is an image with uniformly distributed pixels, and x-pdata(x) Is the distribution of the original image x, y-pdata(y) is uniform distribution, G is encoder, F is decoder and D is discriminator.
Wherein, still include the following step:
the parameters for initializing each convolution sum by performing the neural network model in advance are expressed as:
Wn＝random[wn,1,wn,2,…,wn,j]；
obtaining an encryption key W comprising a parameter W from each convolutional layernComposition, expressed as: w ═ W1,W2,...,Wn]；
Wherein, WnIs a parameter of the nth convolution layer, wn,jIs the jth parameter of the nth convolutional layer;
wherein, the gradient descending process of the calibration encoder is also included, which is expressed as:
wherein, thetajFor the parameter theta in the j-th generation of trainingValue, α is the learning rate of the control error variation range, V.J. (θ)j) Is passed to the parameter thetajOf the gradient of (c).
Wherein, the step is that each trained Lth layer convolution kernel W is processedLTraversing through the image matrix and associating it with matrix XLCorresponding or filler elements, where L ≦ 24.
Specifically, as shown in fig. 2 to 5, for the above neural network model, in the training, the value of λ and the training algebra select Adam solution as a parameter optimizer, the batch size is 1, and the initial learning rate is 0.0002. Running environment win10, GPU NVIDA RTX 2070. During the training process, the value of λ is gradually increased. Within a training period, we flip the training image with a probability not lower than 50%. We set λ 10 to 100 cycles for our model training, then λ 100 to 150 cycles, then λ 1000 to 150 cycles, then λ 2000 to 300 cycles, then λ 10000 to 300 cycles, finally reduce the learning rate by 50%, retrain 550 cycles, and train the images in the training set one cycle at a time. The discriminator and the encoder, the decoder are alternately trained once for each image in the training set.
In summary, with the above technical solution of the present invention, based on the color image scrambling encryption method of the anti-autoencoder, the source image is subjected to matrix conversion in advance, and the pixel values are mapped to [0,1 ]]Within range, a matrix is acquired
Matrix to be acquired
The image is input to a neural network model to obtain encrypted or decrypted image output, so that automatic and safe scrambling of color image resolution is realized, and adaptability is high.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (6)
1. A color image scrambling encryption method based on a countermeasure automatic encoder is characterized by comprising the following steps:
carrying out matrix conversion on a source image to be processed in advance, and mapping a pixel value to [0,1 ]]Within range, a matrix is acquired
Matrix to be acquired
Inputting the image into a neural network model to obtain encrypted or decrypted image output, and comprising the following steps:
each trained Lth layer convolution kernel WLTraversing through the image matrix and associating it with matrix XLRespective or fill element multiplication;
all W obtained*×X*Adding to determine the new predicted value of the (L +1) th convolutional layer
All the predicted values are compared
The input feature matrixes are collected and combined into a new matrix to form the input feature matrix of the next convolutional layer;
the output matrix X of the last convolutional layerL*And converting the image into an image to obtain a final encrypted or decrypted image.
2. The method for color image scrambling encryption based on countering automatic encoders according to claim 1, characterized in that the step of matrix-transforming the source image to be processed comprises the steps of:
acquiring a 256 × 256 × 3 plaintext or ciphertext image in advance;
the acquired 256 × 256 × 3 plaintext or ciphertext image is converted into a 256 × 256 × 3 matrix.
3. The automatic color image scrambling encryption method based on the countermeasure automatic encoder according to claim 1, characterized in that a neural network model is constructed, comprising:
scaling encoder and decoder losses, which scales codec losses, expressed as:
calibration discriminator loss, expressed as:
wherein x is an original image, y is an image with uniformly distributed pixels, and x-pdata(x) Is the distribution of the original image x, y-pdata(y) is uniform distribution, G is encoder, F is decoder and D is discriminator.
4. The method of color image scrambling encryption based on a competing auto-encoder as claimed in claim 3, further comprising the steps of:
the parameters for initializing each convolution sum by performing the neural network model in advance are expressed as:
Wn＝random[wn，1，wn，2，...，wn，j]；
obtaining an encryption key W comprising a parameter W from each convolutional layernComposition, expressed as: w ═ W1，W2，...，Wn]；
Wherein, WnIs a parameter of the nth convolution layer, wn，jIs the jth parameter of the nth convolutional layer;
5. the method of color image scrambling encryption based on countering automatic encoders of claim 4 further comprising a gradient descent process of the calibration encoder represented as:
wherein, thetajIs the value of the parameter theta in the training j generation, alpha is the learning rate of the control error variation range, VJ (theta)j) Is passed to the parameter thetajOf the gradient of (c).
6. The method of claim 1 wherein each of said trained Lth convolutional kernels W is used as a basis for scrambling a color imageLTraversing through the image matrix and associating it with matrix XLCorresponding or filler elements, where L ≦ 24.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011553826.2A CN112711763B (en) | 2020-12-24 | 2020-12-24 | Color image scrambling encryption method based on countermeasure automatic encoder |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011553826.2A CN112711763B (en) | 2020-12-24 | 2020-12-24 | Color image scrambling encryption method based on countermeasure automatic encoder |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112711763A true CN112711763A (en) | 2021-04-27 |
| CN112711763B CN112711763B (en) | 2023-04-07 |
Family
ID=75544350
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011553826.2A Active CN112711763B (en) | 2020-12-24 | 2020-12-24 | Color image scrambling encryption method based on countermeasure automatic encoder |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112711763B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113407968A (en) * | 2021-06-29 | 2021-09-17 | 平安国际智慧城市科技股份有限公司 | Encryption method, device, equipment and storage medium of target detection model |
| CN114117514A (en) * | 2021-10-29 | 2022-03-01 | 香港理工大学深圳研究院 | Encrypted face recognition method and system based on optical speckle |
| DE102021133923A1 (en) | 2021-12-20 | 2023-06-22 | Deutsches Zentrum für Luft- und Raumfahrt e.V. | Process for encrypting and decrypting user data |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120047097A1 (en) * | 2005-11-23 | 2012-02-23 | Beyondcore, Inc. | Secure Handling of Documents with Fields that Possibly Contain Restricted Information |
| CN107610191A (en) * | 2017-09-28 | 2018-01-19 | 广东工业大学 | A kind of color image encrypting method and device |
| CN111105338A (en) * | 2019-12-20 | 2020-05-05 | 长春理工大学 | Image encryption method based on GAN and chaotic system |
-
2020
- 2020-12-24 CN CN202011553826.2A patent/CN112711763B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120047097A1 (en) * | 2005-11-23 | 2012-02-23 | Beyondcore, Inc. | Secure Handling of Documents with Fields that Possibly Contain Restricted Information |
| CN107610191A (en) * | 2017-09-28 | 2018-01-19 | 广东工业大学 | A kind of color image encrypting method and device |
| CN111105338A (en) * | 2019-12-20 | 2020-05-05 | 长春理工大学 | Image encryption method based on GAN and chaotic system |
Non-Patent Citations (4)
| Title |
|---|
| KENTA IIDA ETC.: "An Image Identification Scheme Of Encrypted Jpeg Images For Privacy-Preserving Photo Sharing Services", 《IEEE》 * |
| 万伟: "一种抗攻击的网络加密算法研究", 《现代电子技术》 * |
| 李彤等: "一种基于耦合映像格子的彩色图像加密算法", 《计算机应用研究》 * |
| 薛茹等: "基于二维最大熵和教与学优化算法的图像分割", 《电视技术》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113407968A (en) * | 2021-06-29 | 2021-09-17 | 平安国际智慧城市科技股份有限公司 | Encryption method, device, equipment and storage medium of target detection model |
| CN114117514A (en) * | 2021-10-29 | 2022-03-01 | 香港理工大学深圳研究院 | Encrypted face recognition method and system based on optical speckle |
| DE102021133923A1 (en) | 2021-12-20 | 2023-06-22 | Deutsches Zentrum für Luft- und Raumfahrt e.V. | Process for encrypting and decrypting user data |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112711763B (en) | 2023-04-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112711763B (en) | Color image scrambling encryption method based on countermeasure automatic encoder | |
| Golts et al. | Unsupervised single image dehazing using dark channel prior loss | |
| CN111260584A (en) | Underwater degraded image enhancement method based on GAN network | |
| CN113313644B (en) | Underwater image enhancement method based on residual double-attention network | |
| CN112712488B (en) | Remote sensing image super-resolution reconstruction method based on self-attention fusion | |
| CN106980791B (en) | Secret sharing method for extended weighting threshold ocean remote sensing image in cloud environment | |
| CN110717868A (en) | Video high dynamic range inverse tone mapping model construction and mapping method and device | |
| CN112132737B (en) | Image robust steganography method without reference generation | |
| Jiang et al. | Perception-driven deep underwater image enhancement without paired supervision | |
| CN113283577A (en) | Industrial parallel data generation method based on meta-learning and generation countermeasure network | |
| CN116188274A (en) | Image super-resolution reconstruction method | |
| CN116486408A (en) | Cross-domain semantic segmentation method and device for remote sensing image | |
| CN114782462A (en) | Semantic weighting-based image information hiding method | |
| CN115511722A (en) | Remote sensing image denoising method based on depth feature fusion network and joint loss function | |
| CN118072034A (en) | Neural network-based pore-throat feature extraction method and system | |
| CN109672885A (en) | A kind of video image encoding and decoding method for mine intelligent monitoring | |
| CN117689704A (en) | Monocular image depth estimation method based on diffusion model | |
| CN116452469A (en) | Image defogging processing method and device based on deep learning | |
| CN116596788A (en) | Multi-stage underwater image enhancement method based on transducer | |
| CN113947512B (en) | High-capacity controllable carrier-free image steganography method | |
| CN116091918A (en) | Land utilization classification method and system based on data enhancement | |
| CN113949880B (en) | Extremely-low-bit-rate man-machine collaborative image coding training method and coding and decoding method | |
| CN115222609A (en) | Underwater image restoration method based on confrontation network model generation and confrontation network model generation training method | |
| Babaheidarian et al. | Decode and transfer: A new steganalysis technique via conditional generative adversarial networks | |
| CN111104868A (en) | Cross-quality face recognition method based on convolutional neural network characteristics |
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 |