A forged image detection method, system and device

Abstract

The present application discloses a forged image detection method, system and device, which relates to the field of image processing technology and is used to obtain more accurate forged image detection results. The method includes: inputting an image to be detected into multiple model sets to obtain first detection results of each of the multiple model sets; wherein each model set includes a first model and multiple second models for detecting the same type of forged images, the second model is used to detect the image to be detected to obtain the second detection result of the image to be detected, and the first model is used to fuse each second detection result in the model set to obtain the first detection result; the second models in different set models are used to detect different types of forged images; based on the first detection results of each of the multiple model sets, a target detection result of the image to be detected is obtained, and the target detection result is used to characterize the probability that the image to be detected is a forged image.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a forged image detection method, system and device. Background Art

[0002] With the rapid development of digital media technology, users can not only easily capture images through various powerful image processing software, but also tamper with the content of images with the help of powerful editing functions and algorithms. This type of tampering process often includes splicing, removal, copying and moving, or image enhancement. Such behavior will reduce the credibility of image information and bring negative impacts to society.

[0003] In order to reduce the risks brought by image forgery technology, more effective deep learning methods are constantly proposed and widely used. At present, deep learning methods usually analyze by extracting various high-dimensional features (such as noise, frequency domain, and pixels) from forged images, including: 1. Image block-based methods; 2. Methods based on manually designed features; 3. Methods based on constrained convolutional layers; 4. Methods based on pixels. These methods have improved the performance of deep learning methods, but the means of tampering with forged images are varied and there are also large differences between the tampering means. The commonly used forged image detection models have poor detection effects on forged images. Summary of the invention

[0004] The present application provides a forged image detection method, system and device for obtaining more accurate forged image detection results.

[0005] In order to achieve the above technical objectives, this application adopts the following technical solutions:

[0006] In a first aspect, an embodiment of the present application provides a method for detecting a forged image, the method comprising:

[0007] Acquire the image to be detected;

[0008] Inputting an image to be detected into multiple model sets to obtain first detection results of each of the multiple model sets; wherein the first detection result is used to characterize the probability of detecting that the image to be detected is a forged image based on the model set, each model set includes a first model and multiple second models for detecting the same type of forged images, the second model is used to detect the image to be detected to obtain a second detection result of the image to be detected, the second detection result is used to characterize the probability that the image to be detected is detected as a forged image by the second model, and the first model is used to fuse the second detection results in the model set to obtain the first detection result; the second models in different set models are used to detect different types of forged images;

[0009] Based on the first detection results of each of the multiple model sets, an object detection result of the image to be detected is obtained, and the object detection result is used to characterize the probability that the image to be detected is a forged image.

[0010] The technical solution provided by the present application brings at least the following beneficial effects: first, the second model in each model set detects the image to be detected, and then the first model in the model set fuses the second detection result of the second model to obtain the first detection result, which can avoid the detection deviation of a single model and improve the accuracy of the first detection result. The second models in different model sets are used to detect different types of forged images. By inputting the image to be detected into the second model of each model set, the first detection results of different model sets characterize the possibility that the forged type corresponding to the model set exists in the image to be detected. This solution is suitable for the detection of forged images of various forged types and has a wider range of applications. In addition, based on the first detection results of each of the multiple model sets, the target detection result of the image to be detected is obtained. The target detection result fully considers the first detection results of each model set, the false alarm rate is lower, and the detection result of the forged image is more accurate.

[0011] In one possible implementation, based on the first detection results of each of the multiple model sets, a target detection result of the image to be detected is obtained, including: determining the weight values ​​of the first detection results of each of the multiple model sets; based on the weight values ​​of the first detection results of each of the multiple model sets, weighted fusion of the first detection results of each of the multiple model sets is performed to obtain the target detection result of the image to be detected.

[0012] In one possible implementation, the sum of the weight values ​​of all first detection results is 1; the weight value of the first detection result is positively correlated with the first detection result; or, the weight value of the first detection result greater than or equal to the first threshold is the first weight value, the weight value of the first detection result less than the first threshold is the second weight value, and the first weight value is greater than the second weight value; or, the weight value of the first detection result greater than or equal to the first threshold is the third weight value, the weight value of the first detection result less than the first threshold is the fourth weight value, the fourth weight value is the quotient of the first numerical value and the first quantity, the first numerical value is the difference between 1 and the second numerical value, the second numerical value is the sum of all third weight values, and the first quantity is the number of all first detection results less than the first threshold.

[0013] In a possible implementation, determining the weight values ​​of the first detection results of each of the plurality of model sets includes: receiving and responding to a first setting value input by a user for each first detection result, determining the first setting value as the weight value of the first detection result, and the target detection result, the first detection result, and the weight value of the first detection result satisfy the following relationship: ;in, Used to characterize target detection results, Used to characterize the i-th first detection result, The weight value used to characterize the i-th first detection result, is also used to characterize the first setting value corresponding to the i-th first detection result, and n is used to characterize the number of model sets; or, receiving and responding to a second setting value input by a user for each first detection result, determining a weight value of the first detection result based on the second setting value, and the target detection result, the first detection result and the weight value of the first detection result satisfy the following relationship: ;in, Used to characterize target detection results, Used to characterize the i-th first detection result, The weight value used to characterize the i-th first detection result, is used to represent the second set value corresponding to the i-th first detection result, and n is used to represent the number of model sets.

[0014] In a possible implementation manner, the method further includes: determining the type of forged image corresponding to the model set that outputs the first detection result greater than or equal to the first threshold as the type of forged image to which the image to be detected belongs.

[0015] In a possible implementation, the method further includes: when the target detection result is greater than or equal to a second threshold, outputting first prompt information, the first prompt information being used to indicate that the probability that the image to be detected is a forged image is the target detection result; when the target detection result is less than the second threshold, outputting second prompt information, the second prompt information being used to indicate that the probability that the image to be detected is not a forged image is the difference between 1 and the target detection result.

[0016] In a possible implementation manner, each model set further includes at least one third model, and the third model is used to detect multiple types of forged images.

[0017] In a possible implementation, the method also includes: for any model set, inputting a preset number of sample images into each second model in the model set to obtain sample detection results output by each second model, each sample image corresponding to a sample true value; inputting the sample detection results output by each second model into the preset model, and training the preset model based on the sample true value corresponding to each sample image; and when the iteration end condition is met, determining that the preset model that has completed training is the first model.

[0018] In a second aspect, the present application provides a forged image detection device, comprising:

[0019] A transmission module, used for acquiring an image to be detected;

[0020] A processing module, used for inputting an image to be detected into a plurality of model sets to obtain first detection results of each of the plurality of model sets; wherein the first detection result is used to characterize the probability of detecting that the image to be detected is a forged image based on the model set, each model set includes a first model and a plurality of second models for detecting the same type of forged images, the second model is used to detect the image to be detected to obtain a second detection result of the image to be detected, the second detection result is used to characterize the probability that the image to be detected is detected as a forged image by the second model, the first model is used to fuse the second detection results in the model set to obtain the first detection result; the second models in different set models are used to detect different types of forged images;

[0021] The processing module is also used to obtain a target detection result of the image to be detected based on the first detection results of each of the multiple model sets, and the target detection result is used to characterize the probability that the image to be detected is a forged image.

[0022] In one possible implementation, the processing module is specifically used to: determine the weight value of the first detection result of each of the multiple model sets; based on the weight value of the first detection result of each of the multiple model sets, perform weighted fusion on the first detection results of each of the multiple model sets to obtain the target detection result of the image to be detected.

[0023] In one possible implementation, the sum of the weight values ​​of all first detection results is 1; the weight value of the first detection result is positively correlated with the first detection result; or, the weight value of the first detection result greater than or equal to the first threshold is the first weight value, the weight value of the first detection result less than the first threshold is the second weight value, and the first weight value is greater than the second weight value; or, the weight value of the first detection result greater than or equal to the first threshold is the third weight value, the weight value of the first detection result less than the first threshold is the fourth weight value, the fourth weight value is the quotient of the first numerical value and the first quantity, the first numerical value is the difference between 1 and the second numerical value, the second numerical value is the sum of all third weight values, and the first quantity is the number of all first detection results less than the first threshold.

[0024] In a possible implementation, the processing module is specifically used to: receive and respond to a first setting value input by a user for each first detection result, determine the first setting value as a weight value of the first detection result, and the target detection result, the first detection result, and the weight value of the first detection result satisfy the following relationship: ;in, Used to characterize target detection results, Used to characterize the i-th first detection result, The weight value used to characterize the i-th first detection result, is also used to characterize the first setting value corresponding to the i-th first detection result, and n is used to characterize the number of model sets; or, receiving and responding to a second setting value input by a user for each first detection result, determining a weight value of the first detection result based on the second setting value, and the target detection result, the first detection result and the weight value of the first detection result satisfy the following relationship: ;in, Used to characterize target detection results, Used to characterize the i-th first detection result, The weight value used to characterize the i-th first detection result, is used to represent the second set value corresponding to the i-th first detection result, and n is used to represent the number of model sets.

[0025] In a possible implementation, the processing module is further configured to determine the type of forged image corresponding to the model set that outputs the first detection result greater than or equal to the first threshold as the type of forged image to which the image to be detected belongs.

[0026] In a possible implementation, the transmission module is also used to: when the target detection result is greater than or equal to a second threshold, output a first prompt information, the first prompt information being used to indicate that the probability that the image to be detected is a forged image is the target detection result; when the target detection result is less than the second threshold, output a second prompt information, the second prompt information being used to indicate that the probability that the image to be detected is not a forged image is the difference between 1 and the target detection result.

[0027] In a possible implementation manner, each model set further includes at least one third model, and the third model is used to detect multiple types of forged images.

[0028] In one possible implementation, the processing module is also used to: for any model set, input a preset number of sample images into each second model in the model set to obtain sample detection results output by each second model, where each sample image corresponds to a sample true value; input the sample detection results output by each second model into the preset model, and train the preset model based on the sample true value corresponding to each sample image; and when the iteration end condition is met, determine that the preset model that has completed the training is the first model.

[0029] In a third aspect, the present application provides an electronic device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement any one of the forged image detection methods provided in the first aspect.

[0030] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and when the computer instructions are executed by a processor, any one of the forged image detection methods provided in the first aspect is implemented.

[0031] In a fifth aspect, the present application provides a computer program product, which includes computer instructions. When the computer instructions are executed by a processor, any one of the forged image detection methods provided in the first aspect is implemented.

[0032] In a sixth aspect, the present application provides a forged image detection system, which includes an electronic device and a prompt device; the electronic device is used to execute any one of the forged image detection methods provided in the first aspect above; the prompt device is used to display type information of the forged image.

[0033] For the specific description of the second to sixth aspects and their various implementations in the present application, reference may be made to the detailed description in the first aspect and its various implementations; and for the beneficial effects of the second to sixth aspects and their various implementations, reference may be made to the beneficial effects analysis in the first aspect and its various implementations, which will not be repeated here.

[0034] These and other aspects of the present application will become more apparent from the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 A schematic diagram of the structure of a forged image detection system applicable to a forged image detection method provided in an embodiment of the present application;

[0036] Figure 2 A schematic diagram of the hardware composition of a computing device provided in an embodiment of the present application;

[0037] Figure 3 A flowchart of a forged image detection method provided in an embodiment of the present application;

[0038] Figure 4 A schematic diagram of a use scenario of a forged image detection method provided in an embodiment of the present application Figure 1 ;

[0039] Figure 5 A schematic diagram of a use scenario of a forged image detection method provided in an embodiment of the present application Figure 2 ;

[0040] Figure 6 A schematic diagram of a use scenario of a forged image detection method provided in an embodiment of the present application Figure 3 ;

[0041] Figure 7An overall logic diagram of a forged image detection method provided in an embodiment of the present application;

[0042] Figure 8 A schematic diagram of the structure of a forged image detection device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0043] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0044] It should be noted that, in the embodiments of the present application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a concrete way. The terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present application, unless otherwise specified, "multiple" means two or more.

[0045] With the rapid development of digital media technology, users can not only easily capture images through various powerful image processing software, but also tamper with the content of images with the help of powerful editing functions and algorithms. This type of tampering process often includes splicing, removal, copying and moving, or image enhancement. Such behavior will reduce the credibility of image information and bring negative impacts to society.

[0046] In order to reduce the risks brought by image forgery technology, more effective deep learning methods are constantly proposed and widely used. At present, deep learning methods usually analyze by extracting various high-dimensional features (such as noise, frequency domain, and pixels) from forged images, including: 1. Image block-based methods; 2. Methods based on manually designed features; 3. Methods based on constrained convolutional layers; 4. Methods based on pixels. These methods have improved the performance of deep learning methods, but the means of tampering with forged images are varied and there are also large differences between the tampering means. The commonly used forged image detection models have poor detection effects on forged images.

[0047] In this regard, an embodiment of the present application provides a forged image detection method, in which an image to be detected is input into multiple model sets, and the second model in each model set first detects the image to be detected, and then the first model in the model set fuses the second detection result of the second model to obtain the first detection result, which can avoid the detection deviation of a single model and improve the accuracy of the first detection result. The second models in different model sets are used to detect different types of forged images. By inputting the image to be detected into the second models of each model set, the first detection results of the different model sets characterize the possibility that the forged type corresponding to the model set exists in the image to be detected. This solution is suitable for forged image detection of various forged types and has a wider range of applications. In addition, based on the first detection results of each of the multiple model sets, the target detection result of the image to be detected is obtained, and the target detection result fully considers the first detection results of each model set, with a lower false alarm rate and a more accurate detection result of the forged image.

[0048] Please refer to Figure 1 , which shows a forged image detection system to which the forged image detection method provided by the present application is applicable. Figure 1 As shown, the forged image detection system 1 includes: an electronic device 10 and a prompt device 20 .

[0049] A communication connection is established between the electronic device 10 and the prompt device 20. It should be understood that the connection mode may be a wireless connection, such as a Bluetooth connection, a wireless fidelity (Wi-Fi) connection, etc., or a wired connection, such as an optical fiber connection, etc., without limitation.

[0050] In some embodiments, the electronic device 10 is used to obtain an image to be detected. Specifically, it can be: a user obtains an image that is suspected to be a forged image, and the user inputs the image to be detected into the electronic device 10 through the terminal device used by the user. It can also be: the electronic device 10 uses each frame of the image received as the image to be detected.

[0051] In some embodiments, multiple model sets are deployed in the electronic device 10, and the electronic device 10 is used to input the image to be detected into the multiple model sets to obtain the first detection results of each of the multiple model sets. Among them, the first detection result is used to characterize the probability of detecting that the image to be detected is a forged image based on the model set, each model set includes a first model and multiple second models for detecting the same type of forged images, the second model is used to detect the image to be detected to obtain the second detection result of the image to be detected, the second detection result is used to characterize the probability that the second model detects that the image to be detected is a forged image, and the first model is used to fuse the second detection results in the model set to obtain the first detection result; the second models in different set models are used to detect different types of forged images.

[0052] In some embodiments, the electronic device 10 is used to detect whether the image to be detected is a forged image. If the image to be detected is detected to be a forged image, the electronic device 10 is further used to determine type information of the forged image and send the type information of the forged image to the prompt device 20.

[0053] In some embodiments, the prompt device 20 receives the type information of the forged image sent by the electronic device 10, and displays the type information of the forged image to the user, so that the user can better understand the forged image. Exemplarily, the prompt device 20 can be a display device, and the prompt device 20 displays the type information of the forged image to the user in the form of a picture or text showing the type information of the forged image. Alternatively, the prompt device 20 can also be a broadcasting device, and displays the type information of the forged image to the user in the form of an audio corresponding to the type information of the forged image.

[0054] In some embodiments, the counterfeit image detection system 1 may include one or more prompting devices 20 .

[0055] In some embodiments, the electronic device 10 and the prompt device 20 may be as follows: Figure 1 As shown in FIG. 1 , they are two independent devices; alternatively, the electronic device 10 and the prompt device 20 may also be integrated together.

[0056] The hardware structure of the electronic device 10 includes: Figure 2 The computing device shown in FIG. Figure 2 Taking the computing device shown as an example, the hardware structure of the electronic device 10 is introduced.

[0057] like Figure 2 As shown, the computing device may include a processor 301, a memory 302, a communication interface 303, and a bus 304. The processor 301, the memory 302, and the communication interface 303 may be connected via a bus 304.

[0058] The processor 301 is the control center of the computing device, which can be a processor or a general term for multiple processing elements. For example, the processor 301 can be a general-purpose central processing unit (CPU) or other general-purpose processors. Among them, the general-purpose processor can be a microprocessor or any conventional processor.

[0059] As an embodiment, the processor 301 may include one or more CPUs, such as Figure 2 CPU 0 and CPU 1 are shown in .

[0060] The memory 302 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, an electrically erasable programmable read-only memory (EEPROM), a disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto.

[0061] In a possible implementation, the memory 302 may exist independently of the processor 301, and the memory 302 may be connected to the processor 301 via a bus 304 to store instructions or program codes. When the processor 301 calls and executes the instructions or program codes stored in the memory 302, the model deployment method provided in the embodiment of the present application can be implemented.

[0062] In another possible implementation, the memory 302 may also be integrated with the processor 301 .

[0063] The communication interface 303 is used for the computing device to connect with other devices through a communication network, which may be Ethernet, a radio access network (RAN), a wireless local area network (WLAN), etc. The communication interface 303 may include a receiving unit for receiving data and a sending unit for sending data.

[0064] The bus 304 may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 2 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0065] It should be pointed out that Figure 2 The structure shown in the figure does not constitute a limitation on the computing device, except Figure 2 In addition to the components shown, the computing device may include more or fewer components than shown, or combine certain components, or arrange the components differently.

[0066] The implementation of the embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0067] like Figure 3 As shown, an embodiment of the present application provides a forged image detection method, which can be performed by the above electronic device, and the method includes the following steps:

[0068] S101, obtaining an image to be detected.

[0069] The image to be detected may be an image suspected of being forged obtained by a user. The user inputs the image to be detected into the electronic device through a terminal device used by the user, so that the electronic device obtains the image to be detected.

[0070] S102: Input the image to be detected into multiple model sets to obtain first detection results of each of the multiple model sets.

[0071] Among them, the first detection result is used to characterize the probability of detecting that the image to be detected is a forged image based on the model set, each model set includes a first model and multiple second models for detecting the same type of forged images, the second model is used to detect the image to be detected to obtain a second detection result of the image to be detected, the second detection result is used to characterize the probability that the second model detects that the image to be detected is a forged image, the first model is used to fuse the second detection results in the model set to obtain the first detection result; the second models in different set models are used to detect different types of forged images.

[0072] In some embodiments, different second models in the same model set may be constructed by different machine learning algorithms.

[0073] In some embodiments, for any model set, the electronic device inputs a preset number of sample images into each second model in the model set to obtain sample detection results output by each second model, where each sample image corresponds to a sample true value; the electronic device inputs the sample detection results output by each second model into the preset model, and trains the preset model based on the sample true value corresponding to each sample image; when the iteration end condition is met, the electronic device determines that the trained preset model is the first model.

[0074] In specific implementation, the preset model can be a model constructed based on machine learning algorithms such as bagging, boosting, and stacking.

[0075] Optionally, after acquiring the sample image, the electronic device first performs unified format processing on the sample image so that the format of the sample image can be applicable to various second models.

[0076] Exemplarily, the electronic device obtains a forged image training data set and formats it. For example, the resolution of each image in the forged image training data set is set to 1024×1024, and the format is set to portable network graphics (PNG). The formatted sample images are input into each second model, and all second models can detect the sample images normally. For each model set, N second models should be included, and the N second models are used to detect the sample images, and the detection results of the N second models on the sample images are obtained. , is a floating point number between [0, 1]. The closer the result is to 1, the higher the probability that the sample image is a forged image of the type corresponding to the second model detection, and vice versa. The first model is trained using a machine learning algorithm, and the second detection results of all the second models in the model set are integrated through the first model. Each sample image will obtain a first detection result in the range of [0, 1] in each model set. The first detection result of each sample image in the range of [0, 1] is used as the feature value of the sample image, and the authenticity of the sample image pre-calibrated by the user is used as the label (the label of the sample image is 1 when the sample image is a forged image, and the label of the sample image is 0 when the sample image is not a forged image). The machine learning algorithm is used for fitting training to construct a first model for fusing the second detection results output by each second model in the model set. Therefore, in the actual detection process, after each second model in each model set outputs the second detection result, the first model in the model set fuses each second detection result to obtain the first detection result of the model set.

[0077] As another example, the process of inputting the image to be detected into the model set can be as follows: Figure 4 As shown, the model set includes three second models, and the three second models output three second detection results. The first model fuses the three second detection results together and outputs the first detection result corresponding to the model set.

[0078] By pre-training the first model, the first model can fuse the second detection results output by each second model into one first detection result. The first detection result then takes into account the second detection results output by different second models, thereby avoiding errors that may be caused by forged image detection by a single second model. Ultimately, the first detection result obtained by fusing the second detection results is more accurate and reliable.

[0079] In some embodiments, the iteration end condition can be set by the user based on computing resources, computing time, whether the calculated result reaches a preset result, whether the calculation result has converged, etc. For example, the user sets that the iterative training can be ended when the difference between the result output by the model and the true value is less than a preset difference, then the iteration end condition is that the difference between the first detection result output by the first model and the true value corresponding to the sample image is less than the preset difference.

[0080] In some embodiments, the present application constructs each model set based on prior knowledge, which can be specifically implemented as follows: ① Consult multiple relevant papers to determine the detection models that are clearly introduced in the papers and are good at detecting a certain type or multiple types of forged images, classify and summarize the detection models that detect different types of forged images, and construct a model set. ② Use data sets of different forged image types for testing, determine which types of forged images various detection models are good at detecting based on experimental data, classify and summarize the detection models, and construct a model set.

[0081] In some embodiments, each model set also includes at least one third model, and the third model is used to detect multiple types of forged images. The second model in the model set can also be called a basic model, and the third model can also be called a general model. The basic model is a detection model that can detect one type of forged image, and the general model is a detection model that can detect multiple types of forged images. The third models in different model sets can be the same or different, and the second models in different model sets are different.

[0082] For example, the user knows through prior knowledge that model 1 is used to detect forged images of type a, model 2 is used to detect forged images of type b, model 3 is used to detect forged images of type c, model 4 is used to detect forged images of type a, model 5 is used to detect forged images of type c, model 6 is used to detect forged images of type a, type b, and type c, model 7 is used to detect forged images of type b, and model 8 is used to detect forged images of type a, type b, and type c. Then the model set A constructed by the user may include model 1, model 4, and model 6 (wherein model 6 may also be replaced by model 8), the model set B may include model 2, model 7, and model 8 (wherein model 8 may also be replaced by model 6), and the model set C may include model 3, model 5, and model 6 (wherein model 6 may also be replaced by model 8).

[0083] In this way, the model set includes both the basic model and the general model. The model set can not only detect a single type of forged image, but also detect other types of forged images through the general model in the model set. The model set has the ability to detect multiple types of forged images. The first detection result corresponding to the model set not only takes into account the type of forged image corresponding to the basic model, but also takes into account other types of forged images, thereby avoiding detection deviation of a single type of forged image, and the detection result is more accurate and reliable.

[0084] S103: Obtaining target detection results of the image to be detected based on the first detection results of each of the multiple model sets.

[0085] Among them, the target detection result is used to characterize the probability that the image to be detected is a forged image.

[0086] In some embodiments, the electronic device first determines the weight values ​​of the first detection results of each of the multiple model sets, and then based on the weight values ​​of the first detection results of each of the multiple model sets, performs weighted fusion on the first detection results of each of the multiple model sets to obtain the target detection result of the image to be detected.

[0087] In one example, the electronic device receives and responds to the first setting value input by the user for each first detection result, determines the first setting value as the weight value of the first detection result, and the sum of the first setting values ​​corresponding to all the first detection results is 1. Based on this, the target detection result, the first detection result and the weight value of the first detection result satisfy the following relationship:

[0088] ;

[0089] in, Used to characterize target detection results, Used to characterize the i-th first detection result, The weight value used to characterize the i-th first detection result, It is also used to represent the first set value corresponding to the i-th first detection result, and n is used to represent the number of model sets.

[0090] In another example, the electronic device receives and responds to the second setting value input by the user for each first detection result, determines the weight value of the first detection result based on the second setting value, and the sum of the weight values ​​corresponding to all the first detection results is 1. Based on this, the target detection result, the first detection result and the weight value of the first detection result satisfy the following relationship:

[0091] ;

[0092] in, Used to characterize target detection results, Used to characterize the i-th first detection result, The weight value used to characterize the i-th first detection result, is used to represent the second set value corresponding to the i-th first detection result, and n is used to represent the number of model sets.

[0093] In some embodiments, the user can determine the second setting value to be input by determining the weight value of the first detection result. For example, there are 3 model sets in total, and the user determines that the weight value of the first detection result of model set 1 is 0.5, the weight value of the first detection result of model set 2 is 0.3, and the weight value of the first detection result of model set 3 is 0.2. Then the user can enter the second setting value corresponding to the first detection result of model set 1 as 5, the second setting value corresponding to the first detection result of model set 2 as 3, and the second setting value corresponding to the first detection result of model set 3 as 2.

[0094] In some embodiments, the sum of the weight values ​​of all first detection results is 1. There is a positive correlation between the weight value of the first detection result and the first detection result; or, the weight value of the first detection result greater than or equal to the first threshold is the first weight value, the weight value of the first detection result less than the first threshold is the second weight value, and the first weight value is greater than the second weight value; or, the weight value of the first detection result greater than or equal to the first threshold is the third weight value, the weight value of the first detection result less than the first threshold is the fourth weight value, the fourth weight value is the quotient of the first value and the first quantity, the first value is the difference between 1 and the second value, the second value is the sum of all third weight values, and the first quantity is the number of all first detection results less than the first threshold. Wherein, the typical value of the first threshold is 0.5.

[0095] In an example, there are 3 model sets, the first detection result of model set 1 is 0.7, the first detection result of model set 2 is 0.3, and the first detection result of model set 3 is 0.4. There is a positive correlation between the first detection result and the weight value of the first detection result. The weight value of the first detection result of model set 1 should be greater than the weight value of the first detection result of model set 3, and greater than the weight value of the first detection result of model set 2. The user can input a first set value corresponding to the first detection result of model set 1 as 0.6, a first set value corresponding to the first detection result of model set 2 as 0.15, and a first set value corresponding to the first detection result of model set 3 as 0.25. Finally, the electronic device determines the weight value of the first detection result of model set 1 as 0.6, the weight value of the first detection result of model set 2 as 0.15, and the weight value of the first detection result of model set 3 as 0.25.

[0096] As another example, assuming the first threshold is 0.5, there are 6 model sets, of which the first detection results of 5 model sets are less than 0.5, and the first detection result of 1 model set is greater than 0.5, the first weight value is 0.25, and the second weight value is 0.15, then the electronic device can determine that the weight values ​​of the first detection results of the 5 model sets whose first detection results are less than 0.5 are each 0.15, and the weight value of the first detection result of the 1 model set whose first detection result is greater than 0.5 is 0.25.

[0097] As another example, assuming that the first threshold is 0.5, there are 6 model sets, of which the first detection results of 2 model sets are greater than 0.5, and the first detection results of the other 4 model sets are less than 0.5. Assuming that the electronic device determines that the weight values ​​of the first detection results of the 2 model sets whose first detection results are greater than 0.5 are both 0.3, that is, the third weight value is 0.3, the second value is 0.6, the first value is 0.4, and the first number is 4, then the fourth weight value is 0.1, and the weight values ​​of the first detection results of the 4 model sets whose first detection results are less than 0.5 are all 0.4÷4=0.1.

[0098] In some embodiments, the electronic device determines the type of forged image corresponding to the model set that outputs the first detection result greater than or equal to the first threshold as the type of forged image to which the image to be detected belongs.

[0099] For example, assuming there are 6 model sets, only the first detection result of 1 model set is greater than 0.5, the type of forged image corresponding to this model set is a copy-type forged image, and the first detection results of the other 5 model sets are all less than 0.5, then the electronic device can determine that the image to be detected is a copy-type forged image.

[0100] For another example, assuming that there are 6 model sets, the first detection results of 3 of them are greater than 0.5, and the types of forged images corresponding to these model sets are moving forged images, copying forged images, and splicing forged images, respectively. The first detection results of the other 3 model sets are all less than 0.5. Then the electronic device can determine that the image to be detected belongs to moving forged images, copying forged images, and splicing forged images.

[0101] In this way, by determining the type of forged image to which the image to be detected belongs, operations such as forged area positioning and forged image restoration can be performed more specifically in the subsequent process, thereby shortening the processing time.

[0102] In some embodiments, when the target detection result is greater than or equal to the second threshold, the electronic device outputs a first prompt message, and the first prompt message is used to indicate that the probability that the image to be detected is a forged image is the target detection result; when the target detection result is less than the second threshold, the electronic device outputs a second prompt message, and the second prompt message is used to indicate that the probability that the image to be detected is not a forged image is the difference between 1 and the target detection result. The second threshold is usually set to 0.5.

[0103] For example, assuming that the first threshold is 0.5, there are 5 model sets, the sum of the second setting values ​​corresponding to all the first detection results is 10, the first detection results of 2 model sets are less than 0.5, assuming that the electronic device determines that the second setting values ​​corresponding to the first detection results of these 2 model sets are both 1, and the first detection results of the other 3 model sets are all greater than 0.5, assuming that the electronic device determines that the second setting values ​​corresponding to the first detection results of these 3 sets are greater than 0.5 are , so that based on the above formula The weight values ​​of the first detection results and the target detection results are obtained. The detection process of the forged image can be as follows: Figure 5 As shown, each model set outputs a first detection result for the image to be detected, the electronic device determines a second set value for each first detection result, and then determines a weight value for each first detection result, and then the electronic device performs weighted fusion on each first detection result to obtain a target detection result, and determines whether the image to be detected is a forged image. If it is determined that the image to be detected is a forged image, the probability that the image to be detected is a forged image is output; if it is determined that the image to be detected is not a forged image, the probability that the image to be detected is not a forged image is output.

[0104] As another example, assuming that the first threshold is 0.5, there are 5 model sets, of which the first detection results of 2 model sets are greater than 0.5, assuming that the electronic device determines that the second set values ​​corresponding to the first detection results of the 2 model sets are both 6, and the first detection results of the other 3 model sets are all less than 0.5, assuming that the electronic device determines that the second set values ​​corresponding to the first detection results of the 3 model sets are both 1, then the final target detection result is: The process of detecting forged images can be as follows Figure 6 As shown, each model set outputs a first detection result for the image to be detected, the electronic device determines a second set value for each first detection result, and then determines a weight value for each first detection result, and then the electronic device performs weighted fusion on each first detection result to obtain a target detection result, and determines whether the image to be detected is a forged image. If it is determined that the image to be detected is a forged image, the probability that the image to be detected is a forged image is output; if it is determined that the image to be detected is not a forged image, the probability that the image to be detected is not a forged image is output.

[0105] As another example, assuming that the target detection result is 0.6, the electronic device outputs a prompt message that the content is “the probability that the image to be detected is a forged image is 60%”. Assuming that the target detection result is 0.3, the electronic device outputs a prompt message that the content is “the probability that the image to be detected is not a forged image is 70%”.

[0106] By outputting the first prompt information or the second prompt information, the user can perceive the possibility that the image to be detected is a forged image, thereby satisfying the user's usage needs.

[0107] Figure 3 The technical solution shown brings at least the following beneficial effects: the image to be detected is input into multiple model sets, the second model in each model set is first used to detect the image to be detected, and then the first model in the model set fuses the second detection result of the second model to obtain the first detection result, which can avoid the detection deviation of a single model and improve the accuracy of the first detection result. The second models in different model sets are used to detect different types of forged images. By inputting the image to be detected into the second models of each model set, the first detection results of different model sets characterize the possibility that the forged type corresponding to the model set exists in the image to be detected. This solution is suitable for forged image detection of various forged types and has a wider range of applications. In addition, based on the first detection results of each of the multiple model sets, the target detection result of the image to be detected is obtained. The target detection result fully considers the first detection results of each model set, the false alarm rate is lower, and the detection result of the forged image is more accurate.

[0108] The following describes the forged image detection solution provided by this application from the perspective of the overall process:

[0109] like Figure 7 As shown, first, based on prior knowledge, various forged image detection models are tested using image data sets of different forged types, and the detection models are classified and summarized to construct multiple model sets for detecting specific forged image types; the image to be detected is input into each model set for detection; within the model set, an integrated learning algorithm is used to fuse each second detection result, and the first detection result of each model set is output; each first detection result is summarized, and a threshold judgment is performed on each first detection result. If the first detection result is greater than the first threshold, a larger weight value is assigned to the model set, otherwise, a smaller weight value is assigned; the first detection results of each model set are weighted and fused, and the final target detection result is output. Based on the size relationship between the target detection result and the preset second threshold used to characterize the image as a forged image, it is determined whether the image to be detected is a forged image. If the target detection result is greater than the second threshold, it is determined that the image to be detected is a forged image, and the probability that the image to be detected is a forged image is output. Otherwise, it is determined that the image to be detected is not a forged image, and the probability that the image to be detected is not a forged image is output.

[0110] The above mainly introduces the solution provided by the embodiment of the present application from the perspective of the method. In order to achieve the above functions, it includes hardware structures and / or software modules corresponding to the execution of each function. It should be easy to realize that the technical goals in this field are combined with the units and algorithm steps of each example described in the embodiments disclosed in this article, and the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional and technical goals can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0111] like Figure 8 As shown, the embodiment of the present application further provides a forged image detection device, which is used in the forged image detection method shown in the above method embodiment. The forged image detection device 400 includes: a transmission module 401 and a processing module 402.

[0112] Among them, the transmission module 401 is used to obtain the image to be detected; the processing module 402 is used to input the image to be detected into multiple model sets to obtain the first detection results of each of the multiple model sets; wherein the first detection result is used to characterize the probability of detecting that the image to be detected is a forged image based on the model set, each model set includes a first model and multiple second models for detecting the same type of forged images, the second model is used to detect the image to be detected to obtain the second detection result of the image to be detected, the second detection result is used to characterize the probability that the second model detects that the image to be detected is a forged image, the first model is used to fuse the respective second detection results in the model set to obtain the first detection result; the second models in different set models are used to detect different types of forged images; the processing module 402 is also used to obtain a target detection result of the image to be detected based on the first detection results of each of the multiple model sets, and the target detection result is used to characterize the probability that the image to be detected is a forged image.

[0113] In one possible implementation, the processing module 402 is specifically used to: determine the weight value of the first detection result of each of the multiple model sets; based on the weight value of the first detection result of each of the multiple model sets, perform weighted fusion on the first detection results of each of the multiple model sets to obtain the target detection result of the image to be detected.

[0114] In one possible implementation, the processing module 402 is specifically used to: the sum of the weight values ​​of all first detection results is 1; the weight value of the first detection result is positively correlated with the first detection result; or, the weight value of the first detection result greater than or equal to the first threshold is the first weight value, the weight value of the first detection result less than the first threshold is the second weight value, and the first weight value is greater than the second weight value; or, the weight value of the first detection result greater than or equal to the first threshold is the third weight value, the weight value of the first detection result less than the first threshold is the fourth weight value, the fourth weight value is the quotient of the first numerical value and the first quantity, the first numerical value is the difference between 1 and the second numerical value, the second numerical value is the sum of all third weight values, and the first quantity is the number of all first detection results less than the first threshold.

[0115] In a possible implementation, the processing module 402 is specifically configured to: receive and respond to a first setting value input by a user for each first detection result, determine the first setting value as a weight value of the first detection result, and the target detection result, the first detection result, and the weight value of the first detection result satisfy the following relationship: ;in, Used to characterize target detection results, Used to characterize the i-th first detection result, The weight value used to characterize the i-th first detection result, is also used to characterize the first setting value corresponding to the i-th first detection result, and n is used to characterize the number of model sets; or, receiving and responding to a second setting value input by a user for each first detection result, determining a weight value of the first detection result based on the second setting value, and the target detection result, the first detection result and the weight value of the first detection result satisfy the following relationship: ;in, Used to characterize target detection results, Used to characterize the i-th first detection result, The weight value used to characterize the i-th first detection result, is used to represent the second set value corresponding to the i-th first detection result, and n is used to represent the number of model sets.

[0116] In a possible implementation, the processing module 402 is further configured to determine the type of forged image corresponding to the model set that outputs the first detection result greater than or equal to the first threshold as the type of forged image to which the image to be detected belongs.

[0117] In a possible implementation, the transmission module 401 is also used to: when the target detection result is greater than or equal to a second threshold, output a first prompt information, the first prompt information being used to indicate that the probability that the image to be detected is a forged image is the target detection result; when the target detection result is less than the second threshold, output a second prompt information, the second prompt information being used to indicate that the probability that the image to be detected is not a forged image is the difference between 1 and the target detection result.

[0118] In a possible implementation manner, each model set further includes at least one third model, and the third model is used to detect multiple types of forged images.

[0119] In one possible implementation, the processing module 402 is also used to: for any model set, input a preset number of sample images into each second model in the model set to obtain sample detection results output by each second model, where each sample image corresponds to a sample true value; input the sample detection results output by each second model into the preset model, and train the preset model based on the sample true value corresponding to each sample image; and when the iteration end condition is met, determine that the preset model that has completed training is the first model.

[0120] It should be noted that Figure 8 The division of modules in the example is schematic and is only a logical function division. There may be other division methods in actual implementation. For example, two or more functions may be integrated into one processing module. The above integrated modules may be implemented in the form of hardware or software function modules.

[0121] In actual implementation, the transmission module 401 and the processing module 402 can be implemented by the processor of the electronic device calling the computer program code in the memory. The specific execution process can refer to the description of the forged image detection method above, which will not be repeated here.

[0122] Another embodiment of the present application further provides a computer-readable storage medium, which stores computer instructions. When the computer instructions are executed on a computer device, the computer device executes each step executed by the electronic device in the method flow shown in the above method embodiment.

[0123] In another embodiment of the present application, a computer program product is also provided, which includes computer instructions. When the computer instructions are run on a computer device, the computer device executes the various steps performed by the electronic device in the method flow shown in the above method embodiment. In the above embodiment, it can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using a software program, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer execution instructions are loaded and executed on a computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server, or data center to another website site, computer, server, or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode. Computer-readable storage media can be any available media that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media. Available media can be magnetic media (e.g., floppy disks, hard disks, tapes), optical media (e.g., DVDs), etc.

[0124] The above is only a specific implementation of the present application. Those skilled in the art may conceive of changes or substitutions based on the specific implementation provided by the present application, which should all be included in the protection scope of the present application.

Claims

1. A forged image detection method, characterized in that: include: Acquire the image to be detected; Input the image to be detected into multiple model sets to obtain first detection results of each of the multiple model sets; wherein the first detection result is used to characterize the probability of detecting that the image to be detected is a forged image based on the model set, each of the model sets includes a first model, multiple second models for detecting forged images of the same type, and at least one third model for detecting multiple types of forged images, the second model is used to detect the image to be detected to obtain a second detection result of the image to be detected, the second detection result is used to characterize the probability that the image to be detected is detected as a forged image by the second model, and the first model is used to fuse the detection results of each second detection result in the model set and each third model on the image to be detected to obtain the first detection result; second models in different model sets are used to detect different types of forged images, different second models in the same model set are constructed by different machine learning algorithms, and the first model is constructed by a machine learning algorithm; in the same model set, the machine learning algorithm for detecting the type of forged image corresponding to the model set among the machine learning algorithms constituting the third model is different from the machine learning algorithm corresponding to any one of the second models in the model set; Determining a weight value of the first detection result of each of the multiple model sets; Based on the weight values ​​of the first detection results of the multiple model sets, the first detection results of the multiple model sets are weightedly fused to obtain a target detection result of the image to be detected, wherein the target detection result is used to characterize the probability that the image to be detected is a forged image; The determining of the weight values ​​of the first detection results of the plurality of model sets respectively includes: Receiving and responding to a first setting value input by a user for each first detection result, determining the first setting value as a weight value of the first detection result, wherein the target detection result, the first detection result, and the weight value of the first detection result satisfy the following relationship: ； in, used to characterize the target detection result, Used to characterize the i-th first detection result, The weight value used to characterize the i-th first detection result, is also used to represent the first set value corresponding to the i-th first detection result, and n is used to represent the number of model sets; or, Receiving and responding to a second setting value input by a user for each first detection result, determining a weight value of the first detection result based on the second setting value, wherein the target detection result, the first detection result, and the weight value of the first detection result satisfy the following relationship: ； in, used to characterize the target detection result, Used to characterize the i-th first detection result, The weight value used to characterize the i-th first detection result, is used to represent the second set value corresponding to the i-th first detection result, and n is used to represent the number of model sets.

2. The method according to claim 1, characterized in that: The sum of the weight values ​​of all the first detection results is 1; There is a positive correlation between the weight value of the first detection result and the first detection result; or, The weight value of the first detection result that is greater than or equal to the first threshold is a first weight value, the weight value of the first detection result that is less than the first threshold is a second weight value, and the first weight value is greater than the second weight value; or, The weight value of the first detection result greater than or equal to the first threshold is a third weight value, the weight value of the first detection result less than the first threshold is a fourth weight value, the fourth weight value is the quotient of the first numerical value and the first quantity, the first numerical value is the difference between 1 and the second numerical value, the second numerical value is the sum of all third weight values, and the first quantity is the number of all first detection results less than the first threshold.

3. The method according to claim 1, characterized in that The method further comprises: The type of forged image corresponding to the model set that outputs the first detection result greater than or equal to the first threshold is determined as the type of forged image to which the image to be detected belongs.

4. The method according to claim 1, characterized in that: The method further comprises: When the target detection result is greater than or equal to the second threshold, outputting first prompt information, wherein the first prompt information is used to indicate that the probability that the image to be detected is a forged image is the target detection result; When the target detection result is less than the second threshold, second prompt information is output, where the second prompt information is used to indicate that the probability that the image to be detected is not a forged image is the difference between 1 and the target detection result.

5. The method according to any one of claims 1 to 4, characterized in that: The method further comprises: For any model set, a preset number of sample images are input into each of the second models in the model set to obtain sample detection results output by each of the second models, where each of the sample images corresponds to a sample true value; Inputting the sample detection results output by each of the second models into a preset model, and training the preset model based on the sample true value corresponding to each of the sample images; When the iteration end condition is met, the preset model that has completed training is determined to be the first model.

6. A forged image detection device, characterized in that: include: A transmission module, used for acquiring an image to be detected; A processing module, configured to input the image to be detected into a plurality of model sets to obtain first detection results of each of the plurality of model sets; wherein the first detection result is used to characterize the probability of detecting the image to be detected as a forged image based on the model set, each of the model sets includes a first model, a plurality of second models for detecting forged images of the same type, and at least one third model for detecting multiple types of forged images, the second model is used to detect the image to be detected to obtain a second detection result of the image to be detected, the second detection result is used to characterize the probability of the image to be detected as a forged image by the second model, and the first model is used to fuse the detection results of each of the second detection results in the model set and each of the third models on the image to be detected to obtain the first detection result; second models in different model sets are used to detect different types of forged images, different second models in the same model set are constructed by different machine learning algorithms, and the first model is constructed by a machine learning algorithm; in the same model set, the machine learning algorithm for detecting the type of forged image corresponding to the model set among the machine learning algorithms constituting the third model is different from the machine learning algorithm corresponding to any of the second models in the model set; The processing module is further used to determine the weight value of the first detection result of each of the multiple model sets; The processing module is further used to perform weighted fusion on the first detection results of each of the multiple model sets based on the weight values ​​of the first detection results of each of the multiple model sets to obtain a target detection result of the image to be detected, wherein the target detection result is used to characterize the probability that the image to be detected is a forged image; The processing module is specifically used for: Receiving and responding to a first setting value input by a user for each first detection result, determining the first setting value as a weight value of the first detection result, wherein the target detection result, the first detection result, and the weight value of the first detection result satisfy the following relationship: ； in, used to characterize the target detection result, Used to characterize the i-th first detection result, The weight value used to characterize the i-th first detection result, is also used to represent the first set value corresponding to the i-th first detection result, and n is used to represent the number of model sets; or, Receiving and responding to a second setting value input by a user for each first detection result, determining a weight value of the first detection result based on the second setting value, wherein the target detection result, the first detection result, and the weight value of the first detection result satisfy the following relationship: ； in, used to characterize the target detection result, Used to characterize the i-th first detection result, The weight value used to characterize the i-th first detection result, is used to represent the second set value corresponding to the i-th first detection result, and n is used to represent the number of model sets.

7. The device according to claim 6, characterized in that The sum of the weight values ​​of all the first detection results is 1; there is a positive correlation between the weight value of the first detection result and the first detection result; or, the weight value of the first detection result greater than or equal to the first threshold is the first weight value, the weight value of the first detection result less than the first threshold is the second weight value, and the first weight value is greater than the second weight value; or, the weight value of the first detection result greater than or equal to the first threshold is the third weight value, the weight value of the first detection result less than the first threshold is the fourth weight value, the fourth weight value is the quotient of the first numerical value and the first quantity, the first numerical value is the difference between 1 and the second numerical value, the second numerical value is the sum of all third weight values, and the first quantity is the number of all first detection results less than the first threshold; The processing module is further used to determine the type of forged image corresponding to the model set that outputs the first detection result greater than or equal to the first threshold as the type of forged image to which the image to be detected belongs; The transmission module is further used to: when the target detection result is greater than or equal to a second threshold, output first prompt information, the first prompt information is used to indicate that the probability that the image to be detected is a forged image is the target detection result; when the target detection result is less than the second threshold, output second prompt information, the second prompt information is used to indicate that the probability that the image to be detected is not a forged image is the difference between 1 and the target detection result; The processing module is further used to: for any model set, input a preset number of sample images into each of the second models in the model set to obtain sample detection results output by each of the second models, each of the sample images corresponding to a sample true value; Inputting the sample detection results output by each of the second models into a preset model, and training the preset model based on the sample true value corresponding to each of the sample images; When the iteration end condition is met, the preset model that has completed training is determined to be the first model.

8. An electronic device comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 5.

9. A forged image detection system, characterized in that: Includes electronic devices and prompting equipment; The electronic device is used to perform the steps of the method according to any one of claims 1 to 5 above; The prompt device is used to display type information of the forged image.

10. A computer program product comprising computer instructions, characterized in that: When the computer instructions are executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images