Image authentic identification method and device, electronic equipment and storage medium
By combining image reconstruction models and binary classification models, high-level semantic information of images is automatically captured, solving the problem of poor recognition accuracy of traditional image identification methods when generating images using deep generative models, and achieving efficient identification of natural and synthetic images.
Patent Information
- Application Number
- CN202510896882.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-30
- Publication Date
- 2025-11-04
AI Technical Summary
Traditional image discrimination methods have poor accuracy when dealing with images generated by deep generative models, and it is difficult to capture enough discrimination information.
By using a trained image reconstruction model to add and remove noise from the image to be detected, a reconstructed residual image is generated. A binary classification model is then used to automatically judge the fake detection results, capturing the subtle differences in the intrinsic mechanisms between natural and synthetic images.
It improves the accuracy of image authentication and enables automated identification of natural and synthetic images.
Smart Images
Figure CN120894601A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present disclosure relates to image authentication technology, and more particularly, to an image authentication method and device, an electronic device, and a storage medium. BACKGROUND
[0003] Traditional image authentication methods mainly rely on low-level features extracted manually, such as pixel statistical features, noise features, etc. However, since a deep generative model can accurately learn the pixel distribution of an image, the generated image is close to the real image in statistical features, and therefore, the traditional image authentication method cannot capture enough discriminative information when facing the image generated by the deep generative model, resulting in poor authentication accuracy.
[0004] Therefore, there is a need to provide a new technical solution to improve the accuracy of image authentication. SUMMARY
[0005] An object of the present disclosure is to provide a new technical solution for an image authentication method.
[0006] According to a first aspect of the present disclosure, an image authentication method is provided, comprising:
[0007] obtaining a to-be-detected image;
[0008] performing noise adding processing on the to-be-detected image by using a trained image reconstruction model to obtain a noise vector;
[0009] performing noise removing processing on the noise vector by using the trained image reconstruction model to obtain a first reconstructed image;
[0010] obtaining a reconstructed residual image according to the to-be-detected image and the first reconstructed image;
[0011] inputting the reconstructed residual image into a trained binary classification model to obtain an authentication result of the to-be-detected image, wherein the authentication result is one of a natural image and a synthetic image.
[0012] Optionally, the method further comprises:
[0013] obtaining a first image training sample set, wherein each image in the first image training sample set is a synthetic image;
[0014] inputting each synthetic image in the first image training sample set into an image reconstruction model to be trained to obtain a plurality of second reconstructed images;
[0015] obtaining a loss function value according to the pixel value of each pixel point in the synthetic image and the pixel value of each pixel point in the corresponding second reconstructed image;
[0016] The parameters of the image reconstruction model to be trained are adjusted according to the loss function value until the loss function value is less than a preset threshold. Then, the training of the image reconstruction model to be trained is stopped, and a trained image reconstruction model is obtained.
[0017] Optionally, the step of adding noise to the image to be detected using a trained image reconstruction model to obtain a noise vector includes:
[0018] Using a trained image reconstruction model, noise is added to the image to be detected to obtain a noise vector that follows a normal distribution.
[0019] Optionally, the trained image reconstruction model is a trained VAE model or a trained diffusion model.
[0020] Optionally, obtaining the reconstructed residual image based on the image to be detected and the first reconstructed image includes:
[0021] The pixel values of the same position in the image to be detected and the first reconstructed image are subtracted to obtain the pixel values of each pixel after subtraction.
[0022] The reconstructed residual image is obtained based on the pixel values of each pixel after the subtraction process.
[0023] Optionally, the method further includes:
[0024] Obtain a second image training sample set, where each image in the second image training sample set is a reconstructed residual image, and each reconstructed residual image is labeled with a fake detection result;
[0025] Each image in the second image training sample set is input into the binary classification model to be trained, and the trained binary classification model is obtained.
[0026] Optionally, the binary classification model to be trained is a ResNet model.
[0027] According to a second aspect of this disclosure, an image authentication device is provided, comprising:
[0028] The acquisition module is used to acquire the image to be detected;
[0029] The noise processing module is used to add noise to the image to be detected using a trained image reconstruction model to obtain a noise vector.
[0030] The noise reduction module is used to perform noise reduction processing on the noise vector using the trained image reconstruction model to obtain the first reconstructed image.
[0031] The residual image determination module is used to obtain a reconstructed residual image based on the image to be detected and the first reconstructed image;
[0032] The image authentication result determination module is used to input the reconstructed residual image into a trained binary classification model to obtain the authentication result of the image to be detected, wherein the authentication result is one of a natural image and a synthetic image.
[0033] According to a third aspect of this disclosure, an electronic device is provided, including a memory and a processor, the memory storing a computer program for controlling the processor to operate in order to perform an image authentication method according to any one of the first aspects of this disclosure.
[0034] According to a fourth aspect of this disclosure, a non-volatile computer-readable storage medium is provided, having stored thereon computer program instructions that, when executed by a processor, implement the image authentication method according to any one of the first aspects.
[0035] The image authentication method provided in this disclosure uses a trained image reconstruction model to add noise to the image to be detected, obtaining a noise vector. The trained image reconstruction model then denoises the noise vector to obtain a first reconstructed image. Based on the image to be detected and the first reconstructed image, a reconstructed residual image is obtained. This method can automatically mine high-level semantic information of the image, thereby capturing subtle differences in the intrinsic mechanisms between natural and synthetic images. These subtle differences are reflected in the reconstructed residual image. A trained binary classification model is then used to classify the reconstructed residual image to obtain the authentication result of the image to be detected. This achieves automated image authentication. Compared to existing technologies, this method improves the accuracy of image authentication by capturing subtle differences in the intrinsic mechanisms between natural and synthetic images.
[0036] The features and advantages of the embodiments of this specification will become clear from the following detailed description of exemplary embodiments with reference to the accompanying drawings. Attached Figure Description
[0037] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments of this specification and, together with their description, serve to explain the principles of these embodiments.
[0038] Figure 1 A flowchart of an image authentication method according to an embodiment of the present disclosure is shown.
[0039] Figure 2 A flowchart of an image authentication method according to an embodiment of the present disclosure is shown.
[0040] Figure 3A schematic block diagram of an image authentication device according to an embodiment of the present disclosure is shown.
[0041] Figure 4 A structural block diagram of an electronic device according to an embodiment of the present disclosure is shown. Detailed Implementation
[0042] Various exemplary embodiments of this specification will now be described in detail with reference to the accompanying drawings.
[0043] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the embodiments of this specification or their application or use.
[0044] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.
[0045] Figure 1 A flowchart of an image authentication method according to an embodiment of the present disclosure is shown. Figure 1 As shown, the method includes steps S110 to S150.
[0046] Step S110: Obtain the image to be detected.
[0047] Step S120: Using the trained image reconstruction model, noise is added to the image to be detected to obtain a noise vector.
[0048] In one embodiment, the trained image reconstruction model is a trained VAE (Variational Autoencoder) or a trained diffusion model.
[0049] In one embodiment, step S120 specifically includes using a trained image reconstruction model to add noise to the image to be detected, thereby obtaining a noise vector that follows a normal distribution.
[0050] Step S120 involves mapping the image to be detected to a noise space to obtain a noise vector that follows a normal distribution.
[0051] Step S130: Using the trained image reconstruction model, the noise vector is denoised to obtain the first reconstructed image.
[0052] Taking a trained image reconstruction model, specifically a trained VAE model, as an example, the encoder of the trained VAE model maps the image to be detected to a noise space. This involves converting the image to be detected into mean and variance parameters in the noise space, and then sampling a noise vector from this distribution. This noise vector follows a normal distribution. This operation is implemented through a series of convolutional layers and / or fully connected layers. Then, the decoder of the VAE model denoises the noise vector to obtain the first reconstructed image, essentially mapping the noise vector back to the image data space. This operation is also implemented through a series of network layers (convolutional or fully connected layers), ultimately outputting a first reconstructed image with the same size and number of channels as the image to be detected.
[0053] Taking a trained diffusion model as an example, noise is first added to the image data to be detected, and gradually diffused into a noise vector. Then, the noise vector is removed in reverse by using neural network learning to reconstruct an image, which is the first reconstructed image.
[0054] Step S140: Obtain the reconstructed residual image based on the image to be detected and the first reconstructed image.
[0055] In one embodiment, the pixel values of pixels at the same location in the image to be detected and the first reconstructed image are subtracted to obtain the pixel values of each pixel after subtraction. Based on the pixel values of each pixel after subtraction, the reconstructed residual image is obtained.
[0056] When the image to be detected is a color image, the first reconstructed image is also a color image. The pixels at the same location in both the image to be detected and the first reconstructed image are traversed, and the differences in the RGB color components of the pixels at the same location in both images are calculated. For example, if the RGB value of a pixel at a certain location in the image to be detected is (R1, G1, B1), and the RGB value of the corresponding pixel at the same location in the first reconstructed image is (R2, G2, B2), then the difference in the RGB values of that pixel can be calculated using the following formula.
[0057] ΔR = R1 - R2,
[0058] ΔG = G1 - G2,
[0059] ΔB = B1 - B2.
[0060] The RGB values of each pixel after the difference processing can be obtained using the above formula. Based on the RGB values of each pixel after the difference processing, the reconstructed residual image is obtained.
[0061] When the image to be detected is a grayscale image, the first reconstructed image is also a grayscale image. In a grayscale image, the brightness of each pixel is represented by a single numerical value, which is usually within a specific range, such as 0 to 255. This value is called the grayscale value. The pixels at the same location in both the image to be detected and the first reconstructed image are iterated through, and the difference in grayscale values between the pixels at the same location in both images is calculated. Based on the grayscale values of each pixel after the difference processing, the reconstructed residual image is obtained.
[0062] Step S150: Input the reconstructed residual image into the trained binary classification model to obtain the authentication result of the image to be detected, wherein the authentication result is either a natural image or a synthetic image.
[0063] Natural images are images captured by imaging equipment that have not been modified or processed by humans. Synthetic images are images that have been artificially modified or synthesized using machine learning models.
[0064] It should be noted that when the image to be detected is a synthetic image, the first reconstructed image obtained using the trained image reconstruction model has little difference from the original image. When the image to be detected is a natural image, the first reconstructed image obtained using the trained image reconstruction model has a significant difference from the original image. This difference can be represented using the reconstruction residual image. The trained binary classification model can determine the image's authentication result based on the degree of difference in the reconstruction residual image.
[0065] The image authentication method provided in this disclosure uses a trained image reconstruction model to add noise to the image to be detected, obtaining a noise vector. The trained image reconstruction model then denoises the noise vector to obtain a first reconstructed image. Based on the image to be detected and the first reconstructed image, a reconstructed residual image is obtained. This method can automatically mine high-level semantic information of the image, thereby capturing subtle differences in the intrinsic mechanisms between natural and synthetic images. These subtle differences are reflected in the reconstructed residual image. A trained binary classification model is then used to classify the reconstructed residual image to obtain the authentication result of the image to be detected. This achieves automated image authentication. Compared to existing technologies, this method improves the accuracy of image authentication by capturing subtle differences in the intrinsic mechanisms between natural and synthetic images.
[0066] In one embodiment, the image authentication method further includes: acquiring a first image training sample set, wherein each image in the first image training sample set is a synthetic image; inputting each synthetic image in the first image training sample set into an image reconstruction model to be trained to obtain multiple second reconstructed images; obtaining a loss function value based on the pixel value of each pixel in the synthetic image and the pixel value of each pixel in the corresponding second reconstructed image; adjusting the parameters of the image reconstruction model to be trained based on the loss function value until the loss function value is less than a preset threshold, stopping the training of the image reconstruction model to be trained, and obtaining a trained image reconstruction model.
[0067] During the training process of the image reconstruction model, based on the pixel values of each pixel in the synthesized image and the corresponding pixel values in the second reconstructed image, the Mean Squared Error (MSE) is calculated using the following formula and used as the loss function value. MSE can predict the model's accuracy; the smaller the MSE value, the more accurate the fit.
[0068]
[0069] Where m is the total number of pixels, y i To synthesize the pixel value of a pixel at a specific location in the image, This represents the pixel value of the corresponding pixel in the second reconstructed image.
[0070] When the synthesized image is a color image, the MSE (Mean Sequence Size) of each of the RGB channels is calculated separately, and then the average of the MSEs of the three RGB channels is taken as the final MSE. When calculating the MSE of the R channel using the above formula, y... i The R value is the pixel value at a specific location in the composite image. Let y be the R value of the corresponding pixel in the second reconstructed image. When calculating the MSE on the G channel using the above formula, y i To obtain the G value of a pixel at a specific location in the composite image, Let y be the G value of the corresponding pixel in the second reconstructed image. When calculating the MSE on the B channel using the above formula, y i To determine the B value of a pixel at a specific location in the composite image, This is the B value of the pixel at the corresponding position in the second reconstructed image.
[0071] When the synthesized image is a grayscale image, when calculating MSE using the above formula, y i To synthesize the grayscale value of a pixel at a specific location in the image. This represents the grayscale value of the corresponding pixel in the second reconstructed image.
[0072] If the loss function value is greater than or equal to a preset threshold, the parameters of the image reconstruction model are adjusted using the backpropagation algorithm until the loss function value of the image reconstruction model meets the requirements. Then, training stops, and a well-trained image reconstruction model is obtained.
[0073] In one embodiment, the image authentication method further includes: obtaining a second image training sample set, wherein each image in the second image training sample set is a reconstructed residual image, and each reconstructed residual image is labeled with an authentication result; inputting each image in the second image training sample set into a binary classification model to be trained for training, thereby obtaining a trained binary classification model.
[0074] The method for determining the reconstructed residual image can refer to the above embodiments, and will not be elaborated further here. The authentication result is either a natural image or a synthetic image.
[0075] In one embodiment, the binary classification model to be trained is a ResNet model.
[0076] The binary classification model to be trained can also be one of the Support Vector Machine (SVM) or Logistic Regression models.
[0077] Figure 2 This is a schematic diagram of the processing flow of an image authentication method according to an embodiment of the present disclosure.
[0078] See Figure 2 Let P(x) represent a synthetic image generated by a diffusion model, and R(y) represent a natural image. First, using a trained image reconstruction model, the synthetic image P(x) is back-mapped into a noise space following a normal distribution (0,1), yielding a noise vector v(x). Similarly, the natural image R(y) is back-mapped into the same noise space, yielding a noise vector v(y). Then, using the same model, the noise vector v(x) is denoised to obtain a reconstructed image p(x'), and the noise vector v(y) is denoised to obtain a reconstructed image p(y'). The synthetic image P(x) and the reconstructed image p(x') differ significantly, while the natural image R(y) and the reconstructed image p(y') differ less. This difference can be represented using the reconstructed residual image. The corresponding reconstructed residual images are then input into a trained binary classification model to obtain the image authentication results.
[0079] This disclosure also provides an image authentication apparatus for implementing any of the above method embodiments. Figure 3 A structural block diagram of an image authentication device according to some embodiments is shown. For example... Figure 3As shown, the image authentication device 300 may include an acquisition module 310, a noise processing module 320, a noise reduction processing module 330, a residual image determination module 340, and an image authentication result determination module 350.
[0080] The acquisition module 310 is used to acquire the image to be detected.
[0081] The noise processing module 320 is used to add noise to the image to be detected using a trained image reconstruction model to obtain a noise vector.
[0082] The denoising module 330 is used to denoise the noise vector using the trained image reconstruction model to obtain the first reconstructed image.
[0083] The residual image determination module is used to obtain the reconstructed residual image 340 based on the image to be detected and the first reconstructed image.
[0084] The image authentication result determination module 350 is used to input the reconstructed residual image into the trained binary classification model to obtain the authentication result of the image to be detected, wherein the authentication result is either a natural image or a synthetic image.
[0085] In one embodiment, the image authentication device further includes an image reconstruction model training module.
[0086] The image reconstruction model training module is used to obtain a first image training sample set, in which each image is a synthetic image. The synthetic images in the first image training sample set are input into the image reconstruction model to be trained to obtain multiple second reconstructed images. The loss function value is obtained based on the pixel value of each pixel in the synthetic image and the pixel value of each pixel in the corresponding second reconstructed image. The parameters of the image reconstruction model to be trained are adjusted according to the loss function value until the loss function value is less than a preset threshold. Then, the training of the image reconstruction model to be trained is stopped, and the trained image reconstruction model is obtained.
[0087] In one embodiment, the noise processing module 320 is further configured to use a trained image reconstruction model to perform noise processing on the image to be detected, thereby obtaining a noise vector that follows a normal distribution.
[0088] In one embodiment, the trained image reconstruction model is a trained VAE model or a trained diffusion model.
[0089] In one embodiment, the residual image determination module 340 is further configured to perform subtraction processing on the pixel values of pixels at the same position in the image to be detected and the first reconstructed image to obtain the pixel values of each pixel after subtraction processing; and obtain the reconstructed residual image based on the pixel values of each pixel after subtraction processing.
[0090] In one embodiment, the image authentication device further includes a binary classification model training module.
[0091] The binary classification model training module is used to obtain a second image training sample set. Each image in the second image training sample set is a reconstructed residual image, and each reconstructed residual image is labeled with a fake detection result. Each image in the second image training sample set is input into the binary classification model to be trained for training, and a trained binary classification model is obtained.
[0092] In one embodiment, the binary classification model to be trained is a ResNet model.
[0093] This disclosure also provides an electronic device for implementing any of the above method embodiments. Figure 4 A structural block diagram of an electronic device 4 according to some embodiments is shown. The electronic device 4 may be a PC, workstation, laptop, server, etc., and is not limited thereto.
[0094] like Figure 4 As shown, the electronic device 4 includes a processor 410 and a memory 420 for storing executable instructions of the processor 410. The processor 410 is configured to implement an image authentication method according to any embodiment of the present disclosure when executing the instructions stored in the memory 420.
[0095] The processor 410 is used to execute computer instructions, which can be written using instruction sets of architectures such as x86, Arm, RISC, MIPS, and SSE. The memory 420 includes, for example, ROM (Read-Only Memory), RAM (Random Access Memory), and non-volatile memory such as a hard disk, etc., and is not limited thereto.
[0096] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. For the apparatus embodiments, relevant parts can be referred to the descriptions in the method embodiments.
[0097] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.
[0098] Embodiments of this specification may be systems, methods, and / or computer program products. A computer program product may include a computer-readable storage medium having computer instructions stored thereon for causing a processor to implement various aspects of the embodiments of this specification.
[0099] Computer-readable storage media can be tangible devices capable of holding and storing computer instructions for use by computer instruction execution devices. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing computer instructions thereon, and any suitable combination thereof. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.
[0100] The computer instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network layer, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network layer may include copper cables, fiber optic cables, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network layer adapter card or network layer interface in each computing / processing device receives computer instructions from the network layer and forwards those instructions for storage on computer-readable storage media within the respective computing / processing device.
[0101] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this specification. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of computer instructions, which contains one or more executable computer instructions for implementing a specified logical function. In some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions. It will be known to those skilled in the art that implementation in hardware, implementation in software, and implementation using a combination of software and hardware are equivalent.
[0102] Various embodiments of this specification have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or improvement of the technology in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.
Claims
1. An image authentication method, characterized in that, include: Acquire the image to be detected; The trained image reconstruction model is used to add noise to the image to be detected, resulting in a noise vector; The trained image reconstruction model is used to denoise the noise vector to obtain the first reconstructed image; Based on the image to be detected and the first reconstructed image, a reconstructed residual image is obtained; The reconstructed residual image is input into a trained binary classification model to obtain the authentication result of the image to be detected, wherein the authentication result is either a natural image or a synthetic image.
2. The method according to claim 1, characterized in that, The method further includes: Obtain a first image training sample set, wherein each image in the first image training sample set is a synthetic image; Each synthesized image from the first image training sample set is input into the image reconstruction model to be trained to obtain multiple second reconstructed images; The loss function value is obtained based on the pixel values of each pixel in the synthesized image and the corresponding pixel values of each pixel in the second reconstructed image; The parameters of the image reconstruction model to be trained are adjusted according to the loss function value until the loss function value is less than a preset threshold. Then, the training of the image reconstruction model to be trained is stopped, and a trained image reconstruction model is obtained.
3. The method according to claim 1, characterized in that, The step of adding noise to the image to be detected using a trained image reconstruction model to obtain a noise vector includes: Using a trained image reconstruction model, noise is added to the image to be detected to obtain a noise vector that follows a normal distribution.
4. The method according to claim 1, characterized in that, The trained image reconstruction model is either a trained VAE model or a trained diffusion model.
5. The method according to claim 1, characterized in that, The step of obtaining the reconstructed residual image based on the image to be detected and the first reconstructed image includes: The pixel values of the same position in the image to be detected and the first reconstructed image are subtracted to obtain the pixel values of each pixel after subtraction. The reconstructed residual image is obtained based on the pixel values of each pixel after the subtraction process.
6. The method according to claim 1, characterized in that, The method further includes: Obtain a second image training sample set, where each image in the second image training sample set is a reconstructed residual image, and each reconstructed residual image is labeled with a fake detection result; Each image in the second image training sample set is input into the binary classification model to be trained, and the trained binary classification model is obtained.
7. The method according to claim 6, characterized in that, The binary classification model to be trained is a ResNet model.
8. An image authentication device, characterized in that, include: The acquisition module is used to acquire the image to be detected; The noise processing module is used to add noise to the image to be detected using a trained image reconstruction model to obtain a noise vector. The noise reduction module is used to perform noise reduction processing on the noise vector using the trained image reconstruction model to obtain the first reconstructed image. The residual image determination module is used to obtain a reconstructed residual image based on the image to be detected and the first reconstructed image; The image authentication result determination module is used to input the reconstructed residual image into a trained binary classification model to obtain the authentication result of the image to be detected, wherein the authentication result is one of a natural image and a synthetic image.
9. An electronic device, characterized in that, It includes a memory and a processor, the memory storing a computer program for controlling the processor to operate in order to perform the image authentication method according to any one of claims 1 to 7.
10. A non-volatile computer-readable storage medium storing computer program instructions thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the image authentication method according to any one of claims 1 to 7.