Image processing method and device, electronic equipment and storage medium
The ciphertext information is generated through the anti-interference model and image fusion model, which solves the problems of leakage, tampering and illegal use during image transmission, and achieves high security of image transmission.
Patent Information
- Application Number
- CN202411467797.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-21
- Publication Date
- 2025-02-14
AI Technical Summary
Images are susceptible to attacks by hackers, malware and criminals during the Internet transmission process, resulting in image leakage, tampering and illegal use, especially in the field of smart medical care to protect patient privacy.
The host image of the target image is obtained through the anti-interference model, and the target image is embedded into the host image through the image fusion model to generate a container image. Then, a modulation matrix is generated cyclically, the container image is modulated, and the light intensity measurement is performed to generate ciphertext information, including the modulation matrix and light intensity value.
The dual encryption of images is achieved, which significantly improves the security of image transmission, prevents image leakage and tampering, and ensures the protection of patient privacy.
Smart Images
Figure CN119444542A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of information security technology, and in particular to an image processing method, device, electronic device and storage medium. Background Art
[0002] On the Internet, images have become the main carrier of information transmission. However, during the transmission process, images face security threats from hackers, malware and criminals, and are prone to image leakage, tampering and illegal use. For example, in the field of smart medical care, medical images must be transmitted through the Internet, which may lead to the leakage of medical images or patients' personal privacy. Whether it is remote diagnosis or data sharing, the protection of patients' privacy information is crucial. Therefore, a secure image encryption solution is urgently needed. Summary of the invention
[0003] Embodiments of the present application provide an image processing method, an apparatus, an electronic device, and a storage medium, for providing a secure image encryption solution.
[0004] In a first aspect, an embodiment of the present application provides an image processing method, comprising:
[0005] Acquire a host image of a target image to be encrypted through an anti-interference model;
[0006] Embedding the target image into the host image through an image fusion model to obtain a container image;
[0007] The process is repeated N times: generating a modulation matrix, performing modulation processing on the container image using the modulation matrix to obtain a modulated image, and measuring the light intensity of the modulated image to obtain a light intensity value of the modulated image, where N is an integer greater than 1;
[0008] The ciphertext information of the target image is sent, where the ciphertext information includes N modulation matrices and light intensity values of N modulated images.
[0009] In some embodiments, the anti-interference model includes a first deep learning network, M second deep learning networks, a feature fusion component, and a revealing network component, the M second deep learning networks correspond to different interference attacks, M is an integer greater than 1, and
[0010] The host image of the target image to be encrypted is obtained through the anti-interference model, including:
[0011] Input the target image into the first deep learning network for feature extraction, and input M interference images after the target image is subjected to M interference attacks into the second deep learning networks corresponding to the M interference attacks respectively for feature extraction;
[0012] Inputting the image features output by the first deep learning network and the M second deep learning networks into the feature fusion component for feature fusion to obtain fused features of the target image;
[0013] The fused features are input into the revealing network component to perform image restoration, thereby obtaining the host image similar to the target image.
[0014] In some embodiments, the embedding target of the image fusion model is that the host image is similar to the container image, and the target image is similar to the decrypted target image.
[0015] In a second aspect, an embodiment of the present application provides an image processing method, including:
[0016] Receive ciphertext information of a target image, the ciphertext information including N modulation matrices and light intensity values of N modulated images, the light intensity value of each modulated image being obtained by measuring the light intensity of the corresponding modulated image, the modulated image being obtained by modulating a container image using a modulation matrix, the container image being obtained by embedding the target image to be encrypted into a host image through an image fusion model, and the host image being obtained through an anti-interference model;
[0017] Performing image restoration according to the N modulation matrices and the light intensity values of the N modulation images to obtain the container image;
[0018] The container image is subjected to image extraction by revealing the model to obtain the decrypted target image.
[0019] In some embodiments, performing image restoration according to the N modulation matrices and the light intensity values of the N modulation images to obtain the container image includes:
[0020] Determine the light intensity difference between the light intensity value of the i-th modulated image and the light intensity average value of the N modulated images, and take the product of the light intensity difference and the i-th modulation matrix as the i-th restored image;
[0021] The pixel values at the same position of the N restored images are summed and averaged to obtain the container image.
[0022] In some embodiments, the weight parameters of the revealing network components in the anti-interference model and the revealing network components in the revealing model are the same.
[0023] In a third aspect, an embodiment of the present application provides an image processing device, including:
[0024] An acquisition module, used for acquiring a host image of a target image to be encrypted through an anti-interference model;
[0025] A first encryption module, used for embedding the target image into the host image through an image fusion model to obtain a container image;
[0026] The second encryption module is used to execute N times in a loop: generate a modulation matrix, use the modulation matrix to modulate the container image to obtain a modulated image, measure the light intensity of the modulated image to obtain a light intensity value of the modulated image, where N is an integer greater than 1;
[0027] The sending module is used to send the ciphertext information of the target image, wherein the ciphertext information includes N modulation matrices and light intensity values of N modulated images.
[0028] In a fourth aspect, an embodiment of the present application provides an image processing device, including:
[0029] A receiving module, used to receive ciphertext information of a target image, wherein the ciphertext information includes N modulation matrices and light intensity values of N modulated images, wherein the light intensity value of each modulated image is obtained by measuring the light intensity of the corresponding modulated image, wherein the modulated image is obtained by modulating a container image using a modulation matrix, wherein the container image is obtained by embedding the target image to be encrypted into a host image through an image fusion model, and wherein the host image is obtained through an anti-interference model;
[0030] A first decryption module, used for performing image restoration according to the N modulation matrices and the light intensity values of the N modulated images to obtain the container image;
[0031] The second decryption module is used to perform image extraction on the container image by revealing the model to obtain the decrypted target image.
[0032] In a fifth aspect, an embodiment of the present application provides an electronic device, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein:
[0033] The memory stores a computer program that can be executed by at least one processor. The computer program is executed by the at least one processor so that the at least one processor can perform the above-mentioned image processing method.
[0034] In a sixth aspect, an embodiment of the present application provides a storage medium. When a computer program in the storage medium is executed by a processor of an electronic device, the electronic device can execute any of the above-mentioned image processing methods.
[0035] In a seventh aspect, an embodiment of the present application provides a computer program product, including a computer program, which implements any of the above-mentioned image processing methods when executed by a processor.
[0036] In the embodiment of the present application, the host image of the target image to be encrypted is obtained through the anti-interference model, and the target image is embedded in the host image through the image fusion model to obtain the container image, and the first encryption of the target image is completed, and then the cycle is performed N times: a modulation matrix is generated, and the container image is modulated by the modulation matrix to obtain a modulated image, and the light intensity of the modulated image is measured to obtain the light intensity value of the modulated image to complete the secondary encryption of the target image, and then the ciphertext information of the target image is sent, and the ciphertext information includes N modulation matrices and N light intensity values of the modulated image. In this way, the image can be effectively protected by two encryptions, thereby improving the security of image transmission. BRIEF DESCRIPTION OF THE DRAWINGS
[0037] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:
[0038] Figure 1 An interactive flow chart of an image processing method provided in an embodiment of the present application;
[0039] Figure 2 A schematic diagram of an image processing process provided in an embodiment of the present application;
[0040] Figure 3 A schematic diagram of a processing process of an anti-interference model provided in an embodiment of the present application;
[0041] Figure 4 A schematic diagram of the structure of an image processing device provided in an embodiment of the present application;
[0042] Figure 5 A schematic diagram of the structure of another image processing device provided in an embodiment of the present application;
[0043] Figure 6 A schematic diagram of the hardware structure of an electronic device for implementing any image processing method provided in an embodiment of the present application. DETAILED DESCRIPTION
[0044] In order to provide a secure image encryption solution, embodiments of the present application provide an image processing method, device, electronic device, and storage medium.
[0045] The preferred embodiments of the present application are described below in conjunction with the drawings in the specification. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present application, and are not used to limit the present application. In addition, the embodiments and features in the embodiments of the present application may be combined with each other if there is no conflict.
[0046] In order to facilitate the understanding of this application, the technical terms involved in this application are:
[0047] Adversarial attacks: attackers can deceive the model by making slight perturbations to the image, causing it to produce incorrect outputs, which may cause the image protection mechanism to be bypassed or fail, thereby leaking sensitive information in the image.
[0048] Reconstruction Error refers to the difference between the reconstructed image and the original data when the image is reconstructed (or decompressed) after some processing or compression during the image processing process. This difference can usually be quantified by some measurement method to reflect the similarity between the reconstructed image and the original image.
[0049] Diffusion Model is a type of probabilistic model used to generate data. The basic idea of the diffusion model is to gradually diffuse the image to a simple distribution (usually a Gaussian distribution), and then gradually generate the image from the noise through the inverse process.
[0050] Peak Signal-to-Noise Ratio (PSNR) is a commonly used image quality evaluation indicator. It is used to measure the difference between the original image and the compressed or processed image. It is usually used to evaluate the quality of generated images in tasks such as image compression, image denoising, and image restoration. The larger the PSNR, the closer the generated image is to the original image, and the higher the quality of the generated image.
[0051] The solution of the embodiment of the present application can be applied to scenarios where images need to be encrypted, such as remote consultation scenarios in the field of smart medical care, medical image sharing, etc. Another example is face recognition scenarios in the financial field. In addition, it should be noted that the target image to be encrypted in the embodiment of the present application can be a grayscale image or a color image.
[0052] After introducing the application scenarios of the embodiments of the present application, the image processing method proposed in the present application is described below with specific embodiments.
[0053] See also Figure 1 , Figure 1 An interactive flow chart of an image processing method provided in an embodiment of the present application, the method includes the following steps.
[0054] In step 101, the sending end obtains a host image of a target image to be encrypted through an anti-interference model.
[0055] In some embodiments, the anti-interference model includes M+1 different deep learning networks, a feature fusion component, and a revealing network component, the M+1 deep learning networks include a first deep learning network and M second deep learning networks, and the interference attacks corresponding to the M second deep learning networks are different, then the host image can be obtained according to the following steps:
[0056] The target image is input into the first deep learning network for feature extraction, and the M interference images after the target image is subjected to M interference attacks are input into the second deep learning networks corresponding to the M interference attacks for feature extraction. Then, the image features output by the first deep learning network and the image features output by the M second deep learning networks are input into the feature fusion component for feature fusion to obtain the fused features of the target image, and then the fused features are input into the revealing network component for image restoration to obtain a host image similar to the target image, where M is an integer greater than 1.
[0057] It should be noted that when the target image is a grayscale image, M+1 different deep learning networks are selected from the set of deep learning networks suitable for processing grayscale images; when the target image is a color image, M+1 different deep learning networks are selected from the set of deep learning networks suitable for processing color images.
[0058] In step 102, the sending end embeds the target image into the host image through an image fusion model to obtain a container image.
[0059] Among them, the embedding goal of the image fusion model is: the host image and the container image are similar, and the target image and the decrypted target image (referring to the image finally displayed after decryption at the receiving end) are similar. That is, the image fusion model can not only embed the target image into the host image, but also ensure that the host image and the container image are similar, and the target image and the image finally displayed after decryption at the receiving end are similar, so as to improve the accuracy of subsequent image restoration at the receiving end. It should be noted that the embedding goal is generally achieved in the training stage of the image fusion model. In the application stage, the trained image fusion model can achieve the embedding goal.
[0060] Here, the target image and the host image are input into the image fusion model, and the image fusion model completes the embedding. The output of the image fusion model is the container image. The container image is similar to the target image but is not the target image, thus completing the first encryption of the target image.
[0061] In step 103, the transmitting end executes the following steps N times in a loop: generating a modulation matrix, performing modulation processing on the container image using the modulation matrix to obtain a modulated image, and performing light intensity measurement on the modulated image to obtain a light intensity value of the modulated image.
[0062] Wherein, N is an integer greater than 1.
[0063] In practical applications, a matrix can be randomly generated each time as a modulation matrix. The size of the modulation matrix is the same as the size of the container image. For example, if the size of the container image is 256×256, the modulation matrix also contains 256×256 elements. The modulation matrix is regarded as a spot field, and each element in the modulation matrix is a pixel value in the spot field. Using the modulation matrix to modulate the container image is to multiply the pixel values at the same position in the modulation matrix and the container image. The light intensity measurement of the modulated image is to record the encrypted total light intensity value. The formula is as follows:
[0064] D = ∫O′(x, y)dxdy, O′(x, y) represents the modulated image.
[0065] In this way, the container image is encrypted by means of modulation and light intensity measurement, thereby completing the secondary encryption of the target image.
[0066] In step 104, the transmitting end sends the ciphertext information of the target image to the receiving end, where the ciphertext information includes N modulation matrices and N light intensity values of the modulated images.
[0067] That is, the encrypted target image is sent, and the encrypted target image is represented by N modulation matrices and N light intensity values of the modulation images.
[0068] In step 105, the receiving end performs image restoration according to the N modulation matrices and the light intensity values of the N modulation images to obtain the container image.
[0069] In specific implementation, the light intensity difference between the light intensity value of the i-th modulated image and the average light intensity value of N modulated images can be determined, and the product of the light intensity difference and the i-th modulation matrix can be used as the i-th restored image. Then, the pixel values at the same position of the N restored images are summed and averaged to obtain the container image.
[0070] For example, the container image G(x,y) is obtained according to the following formula:
[0071]
[0072] Among them, {·} is the average operation, and {D} represents the light intensity value D 1 ,D 2 ...D N The average value of , Ai(x,y) represents the i-th modulation matrix.
[0073] In step 106, the receiving end extracts the container image by revealing the model to obtain a decrypted target image.
[0074] In practical applications, the revelation model will also include a revelation network component (of course, other network components are also included). The function of the revelation network component is to recover the image from a feature vector containing noise. The weight parameters of the revelation network component in the anti-interference model and the revelation network component in the revelation model are made the same, so that the revelation model has a certain anti-interference ability, thereby improving the image recovery ability of the receiving end.
[0075] In an embodiment of the present application, a target image and interference images subjected to different interference attacks (noise disturbance attack, color shift attack, single pixel attack, etc.) are input into different deep learning networks for feature extraction, and the image features output by these deep learning networks are fused to enhance the representation capability of the image features. Subsequently, a revelation network component (the revelation network component is also a component of a subsequent revelation model) is used to restore the fused image features to obtain a host image. The target image and the host image are combined through an image fusion model to generate a container image, which is then modulated and encoded by a ghost imaging system to generate ciphertext information for transmission. Subsequently, at the receiving end, the received ciphertext information can be first restored to a container image, and then the container image can be extracted through the revelation model to obtain a decrypted target image, wherein the revelation network component and the revelation network component in the revelation model share weights to enhance the algorithm's anti-interference and anti-attack capabilities.
[0076] See also Figure 2 , Figure 2 A schematic diagram of an image processing process provided by an embodiment of the present application, at the sending end, an anti-interference model is used to obtain a host image similar to the target image to be encrypted, and the target image is embedded (i.e., fused) into the host image through an image fusion model to obtain a container image, thereby completing the first encryption, and then the container image is modulated with speckle by a ghost imaging system to obtain a ciphertext image (i.e., ciphertext information), and the second encryption is completed, and the ciphertext information is sent to the receiving end. After receiving the ciphertext information, the ciphertext information is first modulated with speckle by a ghost imaging system to restore the container image, and then the container image is extracted by the revealing model to obtain the decrypted target image.
[0077] The main parts of the above process are introduced below.
[0078] 1. Anti-interference model.
[0079] See also Figure 3 , Figure 3A schematic diagram of the processing process of an anti-interference model provided in an embodiment of the present application, the anti-interference model includes different deep learning networks 0-M, feature fusion components and revealing network components. In practical applications, the target image can be input into deep learning network 0 for feature extraction, the interference image 1 after the target image is subjected to the first interference attack is input into deep learning network 1 for feature extraction, the interference image 2 after the target image is subjected to the second interference attack is input into deep learning network 2 for feature extraction... The interference image M after the target image 1 is subjected to the Mth interference attack is input into deep learning network M for feature extraction, and then the image features output by deep learning networks 0-M are fused through the feature fusion component to obtain the fusion features of the target image, and the fusion features are input into the revealing network component for image restoration to obtain a host image similar to the target image. Among them, deep learning network 0 is the above-mentioned first deep learning network, deep learning networks 1-M are the above-mentioned M second deep learning networks, and various interference attacks such as noise perturbation attack, color shift attack, single pixel attack, etc.
[0080] The following is an introduction to the training process of the anti-interference model.
[0081] (1) Data preparation and preprocessing.
[0082] Select a large image set of uniform type for the image type to be encrypted. For example, if you need to develop a system for encrypting a patient's brain MRI image, you can select a large-scale brain MRI image set and perform standard processing on the images in the image set, such as edge cropping, resizing, grayscale normalization, etc.
[0083] (2) Select and train the basic model.
[0084] Select multiple deep learning networks for image feature extraction, train these deep learning networks with the image sets prepared in (1) and verify to prevent overfitting and achieve parameter fine-tuning.
[0085] (3) Adversarial attack data generation and training.
[0086] For each trained deep learning network in (2), use the corresponding adversarial attack method to generate adversarial samples, add the adversarial samples and the original samples to the training set, and retrain the deep learning network to ensure that the deep learning network can adapt to and extract features from the adversarial samples.
[0087] (4) Feature extraction and fusion.
[0088] Using the deep learning network trained in (3), image features are extracted from the original samples and adversarial samples. The image features extracted by each deep learning network are fused, such as splicing, through the feature fusion component. The fused feature vector contains information from multiple deep learning networks and multiple sample types (original samples and adversarial samples), which is a more comprehensive feature representation of the original sample.
[0089] (5) Image restoration.
[0090] Through the inverse process of the diffusion model, the training reveals that the network component starts from a noisy feature vector obtained by (4) and recovers the original sample by gradually removing noise and restoring features.
[0091] 2. Image fusion model.
[0092] 1) The input is two images: a target image and a host image. The target image is the image that needs to be encrypted and protected; the host image is the restored image obtained after the target image is processed through the above steps.
[0093] A Segment Anything Model (SAM) based on an improved segmentation model is designed and trained as an image fusion model to embed the target image into the host image to obtain a container image. Since the target image needs to be restored from the container image later, the container image should be as similar to the host image as possible.
[0094] To this end, the training objective of the image fusion model is: L(H,C,S,R)=||HC||+β||SR||, where H, C, S, and R represent the host image, container image, target image, and target image after decryption at the receiving end, respectively, and β is a coefficient used to equalize the reconstruction error, which can be determined in advance.
[0095] In addition, the cross entropy loss function can be used, the formula is as follows:
[0096]
[0097] Among them, B represents the total number of image samples, y c represents the true value of the Cth image sample, p c Represents the predicted value of the Cth image sample.
[0098] 3. Ghost imaging system.
[0099] The operations at the transmitting end mainly include random modulation and light intensity measurement. These two steps are introduced below.
[0100] 1) Random modulation.
[0101] Here, a random matrix modulation pattern is used as the modulation spot field, i.e., the key. The randomly generated matrix A(x,y) can be regarded as a two-dimensional array containing random phase information, and Element Delta i (x, y) represents a pixel value in the spot field, and these pixel values are multiplied one by one with the pixel values in the container image to generate a modulated image O'(x, y). That is, A(x, y) and the container image have the same size, and the pixel values at the same position in A(x, y) and the container image are multiplied to form the modulated image O'(x, y).
[0102] 2) Light intensity measurement.
[0103] The light intensity of the modulated image O′(x,y) is measured using the following formula:
[0104] D = ∫O′(x,y)dxdy;
[0105] The encrypted total light intensity value D is recorded and used as ciphertext information.
[0106] Repeat the above modulation and light intensity measurement process N times, and the random matrix used for the i-th time is recorded as A i (x, y), thus obtaining a series of light intensity values D 1 ,D 2 ...D N , light intensity value D 1 ,D 2 ...D N The N modulation matrices are the ciphertext information of the target image. 1 ,D 2 ...D N And N modulation matrices.
[0107] The operation at the receiving end is mainly decryption and restoration, which is introduced below.
[0108] 3) Decryption and restoration.
[0109] After receiving the ciphertext information, the receiver recovers the container image G(x,y) based on the ciphertext information. The calculation formula is:
[0110]
[0111] Among them, {·} is the average operation, and {D} represents the light intensity value D 1 ,D 2 ...D N The average value of .
[0112] 4. Reveal the model.
[0113] The revelation model is also implemented using a diffusion model, and the revelation network component used to restore the image from image features in the revelation model can be the same as the above-mentioned revelation network component. During the training process, the two share weights. In this way, the trained revelation model can have certain anti-interference and anti-attack capabilities, thereby improving the image restoration capability of the receiving end.
[0114] After calculation, in the present application scheme, the structural similarity index (SSIM) between the image restored at the receiving end (i.e., the decrypted target image) and the target image is higher than 0.96, and the PSNR is still higher than 36 at a noise intensity of 0.5.
[0115] The solution of this application has the following advantages:
[0116] (1) The target image to be encrypted is embedded into the host image to generate a container image. The container image is almost the same as the host image. This is the first-level encryption. The ciphertext information of the target image is obtained by modulation and encoding through the modulation matrix. This is the second-level decryption. In this way, double encryption has higher security.
[0117] (2) It integrates multiple deep learning networks that can cope with various adversarial attacks and has strong anti-interference capabilities. The revealing network components used when restoring the host image and the revealing network components in the revealing model when restoring the container image share weights, which can make the revealing model also have strong anti-noise and anti-interference capabilities.
[0118] (3) The target image is hidden in the host image, that is, the image is used as ciphertext. This conforms to the characteristics of computer vision algorithms and overcomes the shortcomings of limited image carrying capacity of methods such as injecting image watermarks and key encryption.
[0119] Based on the same technical concept, the embodiments of the present application also provide two image processing devices. The principles of solving the problems by the two image processing devices are similar to those of the above-mentioned two image processing methods. Therefore, the implementation of the two image processing devices can refer to the implementation of the two image processing methods, and the repeated parts will not be repeated.
[0120] Figure 4 A schematic diagram of the structure of an image processing device provided in an embodiment of the present application includes:
[0121] An acquisition module 401 is used to acquire a host image of a target image to be encrypted through an anti-interference model;
[0122] A first encryption module 402, configured to embed the target image into the host image through an image fusion model to obtain a container image;
[0123] The second encryption module 403 is used to execute N times in a loop: generate a modulation matrix, modulate the container image using the modulation matrix to obtain a modulated image, measure the light intensity of the modulated image to obtain a light intensity value of the modulated image, where N is an integer greater than 1;
[0124] The sending module 404 is used to send the ciphertext information of the target image, where the ciphertext information includes N modulation matrices and light intensity values of N modulated images.
[0125] In some embodiments, the anti-interference model includes a first deep learning network, M second deep learning networks, a feature fusion component and a revealing network component, the M second deep learning networks correspond to different interference attacks, M is an integer greater than 1, and the acquisition module 401 is specifically used to:
[0126] Input the target image into the first deep learning network for feature extraction, and input M interference images after the target image is subjected to M interference attacks into the second deep learning networks corresponding to the M interference attacks respectively for feature extraction;
[0127] Inputting the image features output by the first deep learning network and the M second deep learning networks into the feature fusion component for feature fusion to obtain fused features of the target image;
[0128] The fused features are input into the revealing network component to perform image restoration, thereby obtaining the host image similar to the target image.
[0129] In some embodiments, the embedding target of the image fusion model is that the host image is similar to the container image, and the target image is similar to the decrypted target image.
[0130] Figure 5 A schematic diagram of the structure of another image processing device provided in an embodiment of the present application includes:
[0131] A receiving module 501 is used to receive ciphertext information of a target image, wherein the ciphertext information includes N modulation matrices and light intensity values of N modulated images, wherein the light intensity value of each modulated image is obtained by measuring the light intensity of the corresponding modulated image, wherein the modulated image is obtained by modulating a container image using a modulation matrix, wherein the container image is obtained by embedding the target image to be encrypted into a host image through an image fusion model, and wherein the host image is obtained through an anti-interference model;
[0132] A first decryption module 502, configured to perform image restoration according to the N modulation matrices and the light intensity values of the N modulated images to obtain the container image;
[0133] The second decryption module 503 is used to perform image extraction on the container image by revealing the model to obtain the decrypted target image.
[0134] In some embodiments, the first decryption module 502 is specifically configured to:
[0135] Determine the light intensity difference between the light intensity value of the i-th modulated image and the light intensity average value of the N modulated images, and take the product of the light intensity difference and the i-th modulation matrix as the i-th restored image;
[0136] The pixel values at the same position of the N restored images are summed and averaged to obtain the container image.
[0137] In some embodiments, the weight parameters of the revealing network components in the anti-interference model and the revealing network components in the revealing model are the same.
[0138] The division of modules in the embodiments of the present application is schematic and is only a logical function division. There may be other division methods in actual implementation. In addition, the functional modules in the embodiments of the present application may be integrated in a processor, or may exist physically separately, or two or more modules may be integrated in one module. The coupling between the modules may be achieved through some interfaces, which are usually electrical communication interfaces, but it is not excluded that they may be mechanical interfaces or other forms of interfaces. Therefore, the modules described as separate components may or may not be physically separated, and may be located in one place or distributed to different locations of the same or different devices. The above-mentioned integrated modules may be implemented in the form of hardware or in the form of software functional modules.
[0139] After introducing the image processing method and apparatus according to the exemplary embodiment of the present application, next, an electronic device according to another exemplary embodiment of the present application is introduced.
[0140] Refer to the following Figure 6 The electronic device 130 implemented according to this embodiment of the present application is described. Figure 6 The electronic device 130 shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.
[0141] like Figure 6 As shown, the electronic device 130 is in the form of a general electronic device. The components of the electronic device 130 may include but are not limited to: the at least one processor 131, the at least one memory 132, and a bus 133 connecting different system components (including the memory 132 and the processor 131).
[0142] Bus 133 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, a processor, or a local bus using any of a variety of bus architectures.
[0143] The memory 132 may include a readable medium in the form of a volatile memory, such as a random access memory (RAM) 1321 and / or a cache memory 1322 , and may further include a read-only memory (ROM) 1323 .
[0144] The memory 132 may also include a program / utility 1325 having a set (at least one) of program modules 1324, such program modules 1324 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.
[0145] The electronic device 130 may also communicate with one or more external devices 134 (e.g., keyboards, pointing devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 130, and / or communicate with any device that enables the electronic device 130 to communicate with one or more other electronic devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 135. Furthermore, the electronic device 130 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 136. As shown, the network adapter 136 communicates with other modules for the electronic device 130 via a bus 133. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 130, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.
[0146] In an exemplary embodiment, a storage medium is also provided, and when a computer program in the storage medium is executed by a processor of an electronic device, the electronic device can perform any of the above-mentioned image processing methods. Optionally, the storage medium can be a non-temporary computer-readable storage medium, for example, the non-temporary computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.
[0147] In an exemplary embodiment, the electronic device of the present application may include at least one processor and a memory communicatively connected to the at least one processor, wherein the memory stores a computer program executable by the at least one processor, and when the computer program is executed by the at least one processor, the at least one processor may perform the steps of any image processing method provided in the embodiments of the present application.
[0148] In an exemplary embodiment, a computer program product is also provided. When the computer program product is executed by an electronic device, the electronic device can implement any exemplary method provided in the present application.
[0149] Furthermore, the computer program product may use any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a RAM, a ROM, an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a compact disk read only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.
[0150] The program product for image processing in the embodiment of the present application may adopt a CD-ROM and include program code, and may be run on a computing device. However, the program product of the present application is not limited thereto, and in this document, a readable storage medium may be any tangible medium containing or storing a program, and the program may be used by or in combination with an instruction execution system, apparatus, or device.
[0151] The readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, wherein the readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The readable signal medium may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.
[0152] The program code contained on the readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, radio frequency (RF), etc., or any suitable combination of the foregoing.
[0153] Program code for performing the operations of the present application may be written in any combination of one or more programming languages, including object-oriented programming languages such as Java, C++, etc., and conventional procedural programming languages such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device via any type of network such as a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., via the Internet using an Internet service provider).
[0154] It should be noted that, although several units or subunits of the device are mentioned in the above detailed description, this division is merely exemplary and not mandatory. In fact, according to the embodiments of the present application, the features and functions of two or more units described above can be embodied in one unit. Conversely, the features and functions of one unit described above can be further divided into multiple units to be embodied.
[0155] In addition, although the operations of the method of the present application are described in a specific order in the drawings, this does not require or imply that the operations must be performed in this specific order, or that all the operations shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps.
[0156] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.
[0157] The present application is described with reference to the flowcharts and / or block diagrams of the methods, apparatus (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0158] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0159] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0160] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.
[0161] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application also includes these modifications and variations.
Claims
1. An image processing method, characterized in that: include: Acquire a host image of a target image to be encrypted through an anti-interference model; Embedding the target image into the host image through an image fusion model to obtain a container image; The process is repeated N times: generating a modulation matrix, performing modulation processing on the container image using the modulation matrix to obtain a modulated image, and measuring the light intensity of the modulated image to obtain a light intensity value of the modulated image, where N is an integer greater than 1; The ciphertext information of the target image is sent, where the ciphertext information includes N modulation matrices and light intensity values of N modulated images.
2. The method according to claim 1, characterized in that The anti-interference model includes a first deep learning network, M second deep learning networks, a feature fusion component and a revealing network component, the M second deep learning networks correspond to different interference attacks, and M is an integer greater than 1, and The host image of the target image to be encrypted is obtained through the anti-interference model, including: Input the target image into the first deep learning network for feature extraction, and input M interference images after the target image is subjected to M interference attacks into the second deep learning networks corresponding to the M interference attacks respectively for feature extraction; Inputting the image features output by the first deep learning network and the M second deep learning networks into the feature fusion component for feature fusion to obtain fused features of the target image; The fused features are input into the revealing network component to perform image restoration, thereby obtaining the host image similar to the target image.
3. The method according to claim 1 or 2, characterized in that The embedding goal of the image fusion model is that the host image is similar to the container image, and the target image is similar to the decrypted target image.
4. An image processing method, characterized in that: include: Receive ciphertext information of a target image, the ciphertext information including N modulation matrices and light intensity values of N modulated images, the light intensity value of each modulated image being obtained by measuring the light intensity of the corresponding modulated image, the modulated image being obtained by modulating a container image using a modulation matrix, the container image being obtained by embedding the target image to be encrypted into a host image through an image fusion model, and the host image being obtained through an anti-interference model; Performing image restoration according to the N modulation matrices and the light intensity values of the N modulation images to obtain the container image; The container image is subjected to image extraction by revealing the model to obtain the decrypted target image.
5. The method according to claim 4, characterized in that Performing image restoration according to the N modulation matrices and the light intensity values of the N modulation images to obtain the container image includes: Determine the light intensity difference between the light intensity value of the i-th modulated image and the light intensity average value of the N modulated images, and take the product of the light intensity difference and the i-th modulation matrix as the i-th restored image; The pixel values at the same position of the N restored images are summed and averaged to obtain the container image.
6. The method according to claim 4 or 5, characterized in that The weight parameters of the revealing network component in the anti-interference model and the revealing network component in the revealing model are the same.
7. An image processing device, characterized in that: include: An acquisition module, used for acquiring a host image of a target image to be encrypted through an anti-interference model; A first encryption module, used for embedding the target image into the host image through an image fusion model to obtain a container image; The second encryption module is used to execute N times in a loop: generate a modulation matrix, modulate the container image using the modulation matrix to obtain a modulated image, measure the light intensity of the modulated image to obtain a light intensity value of the modulated image, where N is an integer greater than 1; The sending module is used to send the ciphertext information of the target image, wherein the ciphertext information includes N modulation matrices and light intensity values of N modulated images.
8. An image processing device, characterized in that: include: A receiving module, used to receive ciphertext information of a target image, wherein the ciphertext information includes N modulation matrices and light intensity values of N modulated images, wherein the light intensity value of each modulated image is obtained by measuring the light intensity of the corresponding modulated image, wherein the modulated image is obtained by modulating a container image using a modulation matrix, wherein the container image is obtained by embedding the target image to be encrypted into a host image through an image fusion model, and wherein the host image is obtained through an anti-interference model; A first decryption module, used for performing image restoration according to the N modulation matrices and the light intensity values of the N modulated images to obtain the container image; The second decryption module is used to perform image extraction on the container image by revealing the model to obtain the decrypted target image.
9. An electronic device, characterized in that: include: at least one processor, and a memory communicatively connected to the at least one processor, wherein: The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the method according to any one of claims 1 to 6.
10. A storage medium, characterized in that: When the computer program in the storage medium is executed by a processor of an electronic device, the electronic device can execute the method according to any one of claims 1 to 6.
11. A computer program product, characterized in that The method comprises a computer program, wherein when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.