Image copying detection method and device, computer equipment and storage medium

Abstract

The invention provides an image copying detection method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring a to-be-detected original image; preprocessing the original image to obtain a corresponding first image; performing moire elimination processing on the first image based on a preset moire elimination model to obtain a corresponding second image; performing spatial domain-frequency domain transformation on the first image and the second image to obtain a corresponding first frequency spectrum image and a corresponding second frequency spectrum image; performing merging processing on the first frequency spectrum image and the second frequency spectrum image to obtain a corresponding target dual-channel image; and performing copying judgment processing on the target dual-channel image based on a preset copying judgment model to obtain a copying detection result corresponding to the original image. According to the invention, the copying detection processing of the original image can be accurately realized, the processing efficiency of the image copying detection is improved, and the accuracy of the obtained copying detection result is ensured.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to an image copy detection method, device, computer equipment and storage medium. Background Art

[0002] In the current field of image processing and recognition technology, duplicate image detection plays a vital role as a key link in ensuring information security and authenticity verification. Duplicate image detection, that is, taking a photo of the target object displayed on an electronic screen instead of directly photographing the target object, has become a common physical attack method in sensitive scenarios such as face authentication and card recognition. This attack method attempts to bypass the direct verification mechanism through indirect photography to achieve the purpose of forgery or deception.

[0003] The emergence of re-photographed image detection technology is precisely to meet this challenge. This technology aims to accurately identify those digital images that are not directly taken from the real environment but obtained through re-photographing through a series of algorithms and analysis methods. In the field of face authentication, this technology is usually called liveness detection technology. Its core is to distinguish between real faces and face images re-photographed through media such as photos and videos, so as to effectively prevent fraud. In other natural scenes, re-photographed image detection technology may include a variety of technical means such as reflection detection. These technologies together constitute a multi-dimensional defense system for detecting re-photographing attacks.

[0004] Although the technology of duplicate image detection has made certain progress and has been widely used in multiple application scenarios, there are still some problems that need to be solved. Among them, the most prominent is the accuracy of image duplicate detection. Due to the diversity and complexity of duplicate means, as well as the differences in electronic screen display effects, the boundary between duplicate images and directly shot images becomes blurred, which poses a great challenge to accurate detection. In addition, factors such as different lighting conditions, shooting angles, and equipment performance may also have a significant impact on the detection effect, further increasing the difficulty of technical implementation and resulting in low accuracy of image duplicate detection. Summary of the invention

[0005] The main purpose of the present invention is to provide an image duplication detection method, device, computer equipment and storage medium, aiming to solve the technical problem of low accuracy of image duplication detection in the prior art.

[0006] To achieve the above object, the present invention provides an image copy detection method, the method comprising:

[0007] Obtain the original image to be detected;

[0008] Preprocessing the original image to obtain a corresponding first image;

[0009] Performing moiré removal processing on the first image based on a preset moiré removal model to obtain a corresponding second image;

[0010] Performing spatial-frequency transformation on the first image and the second image respectively to obtain a corresponding first spectrum image and a second spectrum image;

[0011] Merging the first spectrum image and the second spectrum image to obtain a corresponding target dual-channel image;

[0012] The target dual-channel image is subjected to copy discrimination processing based on a preset copy discrimination model to obtain a copy detection result corresponding to the original image.

[0013] Optionally, before performing copy discrimination processing on the target dual-channel image based on a preset copy discrimination model to obtain a copy detection result corresponding to the original image, the method further includes:

[0014] Get pre-built dual-channel image data;

[0015] Performing labeling processing on the dual-channel image data to obtain corresponding sample data;

[0016] Dividing the sample data into a training set and a test set;

[0017] Call the preset classification model;

[0018] Training the classification model based on the training set to obtain a corresponding first model;

[0019] Testing the first model based on the test set to obtain a second model that meets preset construction requirements;

[0020] The second model is used as the copy-make identification model.

[0021] Optionally, after using the second model as the copy discrimination model, the method further includes:

[0022] Determining a designated storage method corresponding to the copy discrimination model;

[0023] Performing storage processing on the copy discrimination model based on the specified storage method;

[0024] After the storage of the copy discrimination model is completed, the copy discrimination model is deployed.

[0025] Optionally, the merging the first spectrum image and the second spectrum image to obtain a corresponding target dual-channel image includes:

[0026] Placing the first spectrum image in a first channel of a preset initial dual-channel image, and placing the second spectrum image in a second channel of the initial dual-channel image, to obtain a corresponding first dual-channel image;

[0027] Performing a size adjustment process on the first dual-channel image to obtain a corresponding second dual-channel image;

[0028] Performing format adjustment processing on the second dual-channel image to obtain a corresponding third dual-channel image;

[0029] The third dual-channel image is used as the target dual-channel image.

[0030] Optionally, performing spatial-frequency transformation on the first image and the second image respectively to obtain a corresponding first spectrum image and a second spectrum image includes:

[0031] Obtaining a preset spatial-frequency domain transformation strategy;

[0032] Performing a space-domain-frequency transformation on the first image based on the space-domain-frequency transformation strategy to obtain a corresponding first initial spectrum image;

[0033] Performing image adjustment processing on the first initial spectrum image to obtain the first spectrum image;

[0034] Performing a spatial-frequency domain transformation on the second image based on the spatial-frequency domain transformation strategy to obtain a corresponding second initial spectrum image;

[0035] Performing image adjustment processing on the second initial spectrum image to obtain the second spectrum image.

[0036] Optionally, preprocessing the original image to obtain a corresponding first image includes:

[0037] Performing image resizing processing on the original image to obtain a corresponding first processed image;

[0038] Normalizing the first processed image to obtain a corresponding second processed image;

[0039] The second processed image is used as the first image.

[0040] Optionally, after performing copy discrimination processing on the target dual-channel image based on a preset copy discrimination model to obtain a copy detection result corresponding to the original image, the method further includes:

[0041] Obtaining an image identifier of the original image;

[0042] generating, based on the copy detection result and the image identifier, copy detection notification information corresponding to the original image;

[0043] Obtain communication information of relevant personnel;

[0044] Based on the communication information, the copy detection notification information is sent to the relevant personnel.

[0045] In addition, to achieve the above-mentioned purpose, the present invention further provides an image copy detection device, the image copy detection device comprising:

[0046] A first acquisition module, used for acquiring an original image to be detected;

[0047] A preprocessing module, used for preprocessing the original image to obtain a corresponding first image;

[0048] an elimination module, configured to perform moiré elimination processing on the first image based on a preset moiré elimination model to obtain a corresponding second image;

[0049] A transformation module, used to perform spatial-frequency transformation on the first image and the second image respectively to obtain a corresponding first spectrum image and a second spectrum image;

[0050] a merging module, configured to merge the first spectrum image with the second spectrum image to obtain a corresponding target dual-channel image;

[0051] The discrimination module is used to perform a copy discrimination process on the target dual-channel image based on a preset copy discrimination model to obtain a copy detection result corresponding to the original image.

[0052] In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical solution:

[0053] The computer device includes a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the steps of any one of the image copying detection methods proposed in the embodiments of the present application are implemented.

[0054] In order to solve the above technical problems, the embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solution:

[0055] The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the steps of any one of the image copying detection methods proposed in the embodiments of the present application are implemented.

[0056] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0057] The present invention provides an image duplication detection method, device, computer equipment and storage medium, the method comprising: first acquiring an original image to be detected; then preprocessing the original image to obtain a corresponding first image; then performing moiré removal processing on the first image based on a preset moiré removal model to obtain a corresponding second image; subsequently performing spatial-frequency domain transformation on the first image and the second image respectively to obtain a corresponding first spectrum image and a second spectrum image; further merging the first spectrum image and the second spectrum image to obtain a corresponding target dual-channel image; finally performing duplication discrimination processing on the target dual-channel image based on a preset duplication discrimination model to obtain a duplication detection result corresponding to the original image. The present invention obtains an original image to be detected, and preprocesses the original image to obtain a corresponding first image, then performs moiré removal processing on the first image based on a preset moiré removal model to obtain a second image, and then performs spatial-frequency domain transform on the first image and the second image respectively to obtain a corresponding first spectrum image and second spectrum image, and then merges the first spectrum image and the second spectrum image to obtain a target dual-channel image, and subsequently performs copy discrimination processing on the target dual-channel image based on the use of a preset copy discrimination model to obtain a copy detection result corresponding to the original image. In this way, the present invention can accurately implement copy detection processing for the original image through the combined use of the moiré removal model and the copy discrimination model, thereby improving the processing efficiency of image copy detection and ensuring the accuracy of the obtained copy detection result. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0059] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0060] Figure 2 is a flow chart of an image copy detection method provided by an embodiment of the present invention;

[0061] Figure 3 is a structural schematic diagram of an embodiment of an image copy detection device according to the present application;

[0062] Figure 4 This is a basic structural block diagram of the computer device in this embodiment. DETAILED DESCRIPTION

[0063] The image remake detection method provided by the embodiment of the present invention is applied to the image remake detection device. Unless otherwise defined, all technical and scientific terms used in this document have the same meaning as those generally understood by technicians in the technical field of this application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" in the specification and claims of this application and the above-mentioned figure descriptions and any variations thereof are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of this application or the above-mentioned figures are used to distinguish different objects, not to describe a specific order.

[0064] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0065] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0066] like Figure 1 As shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0067] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, 103, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social online platform software, etc.

[0068] Terminal devices 101, 102, 103 can be various electronic devices with display screens and supporting web browsing, including but not limited to smart phones, tablet computers, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III), MP4 (Moving Picture Experts Group Audio Layer IV) players, laptop computers, desktop computers, etc.

[0069] The server 105 may be a server that provides various services, such as a background server that provides support for web pages displayed on the terminal devices 101 , 102 , and 103 .

[0070] It should be noted that the image copying detection method provided in the embodiment of the present application is generally executed by a server / terminal device, and accordingly, the image copying detection device is generally arranged in the server / terminal device.

[0071] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to the implementation requirements.

[0072] In the current field of image processing and recognition technology, duplicate image detection plays a vital role as a key link in ensuring information security and authenticity verification. Duplicate image detection, that is, taking a photo of the target object displayed on an electronic screen instead of directly photographing the target object, has become a common physical attack method in sensitive scenarios such as face authentication and card recognition. This attack method attempts to bypass the direct verification mechanism through indirect photography to achieve the purpose of forgery or deception.

[0073] The emergence of re-photographed image detection technology is precisely to meet this challenge. This technology aims to accurately identify those digital images that are not directly taken from the real environment but obtained through re-photographing through a series of algorithms and analysis methods. In the field of face authentication, this technology is usually called liveness detection technology. Its core is to distinguish between real faces and face images re-photographed through media such as photos and videos, so as to effectively prevent fraud. In other natural scenes, re-photographed image detection technology may include a variety of technical means such as reflection detection. These technologies together constitute a multi-dimensional defense system for detecting re-photographing attacks.

[0074] Although the technology of duplicate image detection has made certain progress and has been widely used in multiple application scenarios, there are still some problems that need to be solved. Among them, the most prominent is the accuracy of image duplicate detection. Due to the diversity and complexity of duplicate means, as well as the differences in electronic screen display effects, the boundary between duplicate images and directly shot images becomes blurred, which poses a great challenge to accurate detection. In addition, factors such as different lighting conditions, shooting angles, and equipment performance may also have a significant impact on the detection effect, further increasing the difficulty of technical implementation and resulting in low accuracy of image duplicate detection.

[0075] Continue to refer Figure 2 , shows a flow chart of an embodiment of the image remake detection method proposed in the present application. The embodiment of the present application can acquire and process relevant data based on artificial intelligence technology.

[0076] The image copy detection method provided by the embodiment of the present invention comprises the following steps:

[0077] S210, obtaining an original image to be detected.

[0078] In this step, the execution subject of the present invention may be a copy detection system, referred to as a system. The present invention may be applied to image copy detection processing in various scenes such as faces, documents, and vehicles. The original image is an image to be copied and detected, and the original image is usually represented in the form of a pixel matrix, such as a two-dimensional array.

[0079] S220: Preprocess the original image to obtain a corresponding first image.

[0080] In this step, the specific implementation process of preprocessing the original image to obtain the corresponding first image will be further described in detail in the subsequent specific embodiments of the present invention, and will not be elaborated on here.

[0081] S230: Perform moiré removal processing on the first image based on a preset moiré removal model to obtain a corresponding second image.

[0082] In this step, a pre-trained moiré removal model is loaded, such as MBCNN-DDA or ESDNet. These moiré removal models are usually deep neural networks that have learned to identify and remove moiré through a large amount of data training. Specifically, by inputting the pre-processed first image into the moiré removal model, the moiré removal model will output a second image after the moiré is removed. This output image is also represented in the form of a pixel matrix.

[0083] S240 , performing space-domain-frequency domain transformation on the first image and the second image respectively to obtain a corresponding first spectrum image and a second spectrum image.

[0084] In this step, the specific implementation process of performing spatial-frequency domain transformation on the first image and the second image to obtain the corresponding first spectrum image and second spectrum image will be further described in detail in subsequent specific embodiments of the present invention and will not be elaborated on here.

[0085] S250: Merge the first spectrum image and the second spectrum image to obtain a corresponding target dual-channel image.

[0086] In this step, the specific implementation process of merging the first spectrum image and the second spectrum image to obtain the corresponding target dual-channel image will be further described in detail in subsequent specific embodiments of the present invention, and will not be elaborated on here.

[0087] S260, performing copy discrimination processing on the target dual-channel image based on a preset copy discrimination model to obtain a copy detection result corresponding to the original image.

[0088] In this step, by inputting the above-mentioned target dual-channel image into the copy discrimination model, the copy discrimination model will output a score representing the classification result, and then it can be judged whether the original image is a copy image based on the output score. Specifically, if the output score is higher than the preset threshold, the original image is judged to be a copy image (output 1), otherwise the original image is judged to be a non-copy image (output 0). Among them, the numerical setting of the above-mentioned threshold can be determined according to actual business needs, for example, it can be set to 0.7. In addition, the model construction process of the above-mentioned copy discrimination model will be further described in detail in the subsequent specific embodiments of the present invention, and will not be elaborated on here.

[0089] In an embodiment of the present invention, an original image to be detected is first acquired; then the original image is preprocessed to obtain a corresponding first image; then the first image is subjected to moiré removal processing based on a preset moiré removal model to obtain a corresponding second image; subsequently, the first image and the second image are subjected to spatial-frequency domain transformation respectively to obtain a corresponding first spectrum image and a second spectrum image; the first spectrum image and the second spectrum image are further merged to obtain a corresponding target dual-channel image; finally, the target dual-channel image is subjected to copy discrimination processing based on a preset copy discrimination model to obtain a copy detection result corresponding to the original image. The present invention obtains an original image to be detected, and preprocesses the original image to obtain a corresponding first image, then performs moiré removal processing on the first image based on a preset moiré removal model to obtain a second image, and then performs spatial-frequency domain transform on the first image and the second image respectively to obtain a corresponding first spectrum image and second spectrum image, and then merges the first spectrum image and the second spectrum image to obtain a target dual-channel image, and subsequently performs copy discrimination processing on the target dual-channel image based on the use of a preset copy discrimination model to obtain a copy detection result corresponding to the original image. In this way, the present invention can accurately implement copy detection processing for the original image through the combined use of the moiré removal model and the copy discrimination model, thereby improving the processing efficiency of image copy detection and ensuring the accuracy of the obtained copy detection result.

[0090] Optionally, before performing copy discrimination processing on the target dual-channel image based on a preset copy discrimination model to obtain a copy detection result corresponding to the original image, the method further includes:

[0091] Get pre-built two-channel image data.

[0092] In this step, the construction process of the dual-channel image data includes: collecting a certain number of non-reproduced images and collecting a certain number of reproduced images. There is no limitation on the value of the first number, which can be determined according to actual business needs, for example, it can be set to 1500. Then, for each of the non-reproduced images and the reproduced images, the spatial domain-frequency domain transformation step and the dual-channel image merging step are respectively performed to obtain the corresponding dual-channel image data.

[0093] The dual-channel image data is labeled to obtain corresponding sample data.

[0094] In this step, a classification label can be assigned to each dual-channel image in the above dual-channel image data: the dual-channel image label of the non-reproduced image is 0, and the dual-channel image label of the reproducible image is 1, and the dual-channel images and their corresponding labels are combined into samples to construct sample data containing 3000 samples.

[0095] The sample data is divided into a training set and a test set.

[0096] In this step, the sample data may be divided into corresponding training sets and test sets according to a preset division ratio, wherein the selection of the division ratio is not specifically limited, for example, 7:3 may be used.

[0097] Call the preset classification model.

[0098] In this step, a deep neural network structure suitable for the binary classification task of an image can be selected as the above classification model. This model can be a convolutional neural network (CNN), a residual network (ResNet), an Inception network, etc. The specific choice depends on the task requirements, computing resources, and available time. Among them, it is necessary to ensure that the input layer of the model can receive dual-channel images (that is, the number of input channels is 2). And set the output layer of the model to a single neuron, and use the sigmoid activation function to output a probability value between 0 and 1, indicating the probability that the image is a copy.

[0099] The classification model is trained based on the training set to obtain a corresponding first model.

[0100] In this step, the process of training the classification model based on the training set includes: the parameters of the classification model can be initialized by using random initialization or pre-trained weights. Then a loss function suitable for the binary classification task is selected, such as Binary Cross-Entropy Loss. And an optimizer is selected to update the parameters of the model, such as Adam, SGD, etc. Then the training set is divided into multiple batches, each batch contains a certain number of samples; forward propagation is performed on each batch to calculate the loss value; back propagation is performed using the optimizer to update the parameters of the model; the above process is repeated until the preset number of training rounds (epoch) is reached or the loss value no longer decreases significantly, thereby obtaining a trained first model.

[0101] The first model is tested based on the test set to obtain a second model that meets preset construction requirements.

[0102] In this step, the performance of the first model can be evaluated by using the test set to calculate the first model's accuracy, recall, F1 score and other indicators. Then, according to the evaluation results, the hyperparameters of the first model (such as learning rate, batch size, number of network layers, etc.) are adjusted, different model structures or optimizers are tried, and the training and evaluation process is repeated until the model performance reaches a satisfactory level, thereby obtaining a second model that meets the construction requirements.

[0103] The second model is used as the copy-make identification model.

[0104] In an embodiment of the present invention, by obtaining pre-constructed dual-channel image data; and annotating the dual-channel image data to obtain corresponding sample data; then dividing the sample data into a training set and a test set; then calling a preset classification model; and training the classification model based on the training set to obtain a corresponding first model; subsequently testing the first model based on the test set to obtain a second model that meets the preset construction requirements; and finally using the second model as the remake discrimination model. The present invention obtains sample data by annotating the acquired pre-constructed dual-channel image data, and divides the sample data into a training set and a test set, and then calling a preset classification model; then training the classification model based on the training set to obtain a corresponding first model; subsequently testing the first model based on the test set to obtain a second model that meets the preset construction requirements, and using the second model as the remake discrimination model, so that a remake discrimination model that can accurately distinguish between remake images and non-remake images can be efficiently and accurately constructed, thereby improving the construction efficiency of the remake discrimination model and ensuring the model effect of the remake discrimination model obtained.

[0105] Optionally, after using the second model as the copy discrimination model, the method further includes:

[0106] A designated storage method corresponding to the copy discrimination model is determined.

[0107] In this step, there is no limitation on the determination of the above-mentioned specified storage method, and it can be selected according to actual business needs. For example, local database storage, file system storage, disk storage, cloud server storage, blockchain storage, etc. can be used.

[0108] The copy discrimination model is stored based on the designated storage method.

[0109] In this step, the designated storage medium corresponding to the designated storage method can be determined, and then the copy discrimination model can be stored in the designated storage medium, thereby completing the storage process of the copy discrimination model.

[0110] After the storage of the copy discrimination model is completed, the copy discrimination model is deployed.

[0111] In this step, the trained copy discrimination model can be deployed to the system to handle complex copy detection tasks (copy discrimination of new images), such as copy detection tasks for various scenarios such as faces, documents, and vehicles.

[0112] In the embodiment of the present invention, by determining a designated storage method corresponding to the copy discrimination model; then storing the copy discrimination model based on the designated storage method; after completing the storage of the copy discrimination model, deploying the copy discrimination model. After completing the construction of the copy discrimination model, the present invention will also intelligently use the designated storage method to store the copy discrimination model, thereby ensuring the security of the generated copy discrimination model. In addition, the copy discrimination model will also be deployed so that the subsequent system can use the copy discrimination model to process the copy detection task, thereby providing efficient and high-quality copy detection services, which is conducive to improving the processing quality of copy detection, and further improving customer satisfaction.

[0113] Optionally, the merging the first spectrum image and the second spectrum image to obtain a corresponding target dual-channel image includes:

[0114] The first spectrum image is placed in a first channel of a preset initial dual-channel image, and the second spectrum image is placed in a second channel of the initial dual-channel image to obtain a corresponding first dual-channel image.

[0115] In this step, the corresponding first dual-channel image can be obtained by using the matrix of the first spectrum image as the first channel of the initial dual-channel image and using the matrix of the second spectrum image as the second channel of the initial dual-channel image.

[0116] The first dual-channel image is resized to obtain a corresponding second dual-channel image.

[0117] In this step, the size input requirement of the copy discrimination model regarding the input image can be obtained, and the first dual-channel image can be resized according to the size input requirement so that the generated second dual-channel image meets the input requirement of the copy discrimination model.

[0118] Performing format adjustment processing on the second dual-channel image to obtain a corresponding third dual-channel image.

[0119] In this step, the format requirements of the above-mentioned copy discrimination model regarding the input image can be obtained, and then the format of the second dual-channel image can be adjusted according to the format requirements, which may include adjusting the channel order or data type, so that the generated third dual-channel image meets the input requirements of the copy discrimination model.

[0120] The third dual-channel image is used as the target dual-channel image.

[0121] In an embodiment of the present invention, the first spectrum image is placed in the first channel of a preset initial dual-channel image, and the second spectrum image is placed in the second channel of the initial dual-channel image, to obtain a corresponding first dual-channel image; then the first dual-channel image is resized to obtain a corresponding second dual-channel image; then the second dual-channel image is formatted to obtain a corresponding third dual-channel image; and subsequently the third dual-channel image is used as the target dual-channel image. The present invention places the first spectrum image in the first channel of a preset initial dual-channel image, and places the second spectrum image in the second channel of the initial dual-channel image, to obtain a corresponding first dual-channel image, and then resizes the first dual-channel image to obtain a corresponding second dual-channel image, and subsequently formats the second dual-channel image to obtain a corresponding third dual-channel image as the desired target dual-channel image, thereby achieving intelligent and accurate completion of the merging process of the first spectrum image and the second spectrum image, and effectively ensuring the accuracy and standardization of the obtained target dual-channel image.

[0122] Optionally, performing spatial-frequency transformation on the first image and the second image respectively to obtain a corresponding first spectrum image and a second spectrum image includes:

[0123] Get the preset spatial-frequency transformation strategy.

[0124] In this step, an appropriate spatial-frequency domain transformation method, such as fast Fourier transform (FFT) or wavelet transform, can be selected according to the requirements. Fast Fourier transform is preferably used as the spatial-frequency domain transformation strategy because it can quickly convert the image from the spatial domain to the frequency domain.

[0125] Based on the spatial-frequency domain transformation strategy, the first image is transformed in the spatial-frequency domain to obtain a corresponding first initial spectrum image.

[0126] In this step, the spatial spectrum corresponding to the first image, i.e., the first initial spectrum image, can be obtained by applying the selected spatial-frequency domain transformation strategy to the first image. The spatial spectrum is the representation of the image in the frequency domain, usually a complex matrix, which contains information about the image at different frequencies.

[0127] Performing image adjustment processing on the first initial spectrum image to obtain the first spectrum image.

[0128] In this step, after the first initial spectrum image is obtained, image adjustment processing is performed on the first initial spectrum image, including taking absolute values, normalization and other processing, so as to facilitate subsequent analysis and comparison.

[0129] Based on the spatial-frequency domain transformation strategy, the second image is transformed in the spatial-frequency domain to obtain a corresponding second initial spectrum image.

[0130] In this step, the above-mentioned processing process of performing spatial-frequency domain transformation on the second image based on the spatial-frequency domain transformation strategy to obtain the corresponding second initial spectrum image can refer to the above-mentioned implementation process of performing spatial-frequency domain transformation on the first image based on the spatial-frequency domain transformation strategy to obtain the corresponding first initial spectrum image, and no further details are given here.

[0131] Performing image adjustment processing on the second initial spectrum image to obtain the second spectrum image.

[0132] In this step, the process of performing image adjustment processing on the second initial spectrum image to obtain the second spectrum image can refer to the implementation process of performing image adjustment processing on the first initial spectrum image to obtain the first spectrum image, and no further details are given here.

[0133] In an embodiment of the present invention, a preset spatial-frequency domain transformation strategy is obtained; then, based on the spatial-frequency domain transformation strategy, a spatial-frequency domain transformation is performed on the first image to obtain a corresponding first initial spectrum image; then, the first initial spectrum image is subjected to image adjustment processing to obtain the first spectrum image; subsequently, the second image is subjected to spatial-frequency domain transformation based on the spatial-frequency domain transformation strategy to obtain a corresponding second initial spectrum image; and finally, the second initial spectrum image is subjected to image adjustment processing to obtain the second spectrum image. The present invention obtains a preset spatial-frequency domain transformation strategy, then, based on the use of the spatial-frequency domain transformation strategy, a spatial-frequency domain transformation is performed on the first image to obtain a corresponding first initial spectrum image, and image adjustment processing is performed on the first initial spectrum image to obtain a first spectrum image, and based on the use of the spatial-frequency domain transformation strategy, a spatial-frequency domain transformation is performed on the second image to obtain a corresponding second initial spectrum image, and image adjustment processing is performed on the second initial spectrum image to obtain a second spectrum image, thereby achieving efficient and accurate completion of spatial-frequency domain transformation processing for the first image and the second image, and effectively ensuring the accuracy of the obtained first spectrum image and the second spectrum image.

[0134] Optionally, preprocessing the original image to obtain a corresponding first image includes:

[0135] The original image is subjected to image resizing processing to obtain a corresponding first processed image.

[0136] In this step, the input image size expected by the above-mentioned moiré removal model can be obtained. Then the current size of the above-mentioned original image is obtained, and the current size of the input original image and the input image size expected by the above-mentioned moiré removal model are compared to calculate the corresponding scaling ratio. Exemplarily, if the original image is 512x512 pixels, and the moiré removal model expects 256x256 pixels, then the scaling ratio is 0.5 (i.e., each dimension is reduced by half). Then, the above-mentioned original image is scaled to the expected input image size using an image scaling algorithm (such as bilinear interpolation, nearest neighbor interpolation, etc.). This step can be performed by recalculating each pixel position in the original image, and selecting or interpolating new pixel values ​​from the original image according to the scaling ratio.

[0137] The first processed image is normalized to obtain a corresponding second processed image.

[0138] In this step, the normalization process is to adjust the pixel values ​​of the image to a standard range (such as between 0 and 1) to eliminate the differences in brightness, contrast, etc. between different images, thereby improving the generalization ability of the model.

[0139] The second processed image is used as the first image.

[0140] In an embodiment of the present invention, the original image is resized to obtain a corresponding first processed image; the first processed image is then normalized to obtain a corresponding second processed image; and the second processed image is subsequently used as the first image. The present invention can efficiently and accurately complete the preprocessing of the original image by performing image resizing and normalization on the original image, thereby ensuring the accuracy and adaptability of the obtained first image and ensuring that the generated first image can meet the image input requirements of the moiré removal model.

[0141] Optionally, after performing copy discrimination processing on the target dual-channel image based on a preset copy discrimination model to obtain a copy detection result corresponding to the original image, the method further includes:

[0142] Obtain an image identifier of the original image.

[0143] In this step, the above-mentioned image identification refers to the image ID of the original image. There is a one-to-one correspondence between the image ID and the image, and the image ID is unique.

[0144] Generate copy detection notification information corresponding to the original image based on the copy detection result and the image identifier.

[0145] In this step, the copy detection result and the image identification can be filled into the corresponding positions in the preset notification information template to generate the copy detection notification information corresponding to the original image. The notification information template is a template constructed according to the actual notification requirements of the copy detection result. The content of the notification information template may include a copy detection result area and an image identification area.

[0146] Get the communication information of relevant personnel.

[0147] In this step, the above-mentioned relevant personnel may refer to business personnel or processing personnel related to the business processing of image remake detection. The above-mentioned communication information may include email addresses, mobile phone numbers, etc.

[0148] Based on the communication information, the copy detection notification information is sent to the relevant personnel.

[0149] In this step, the copy detection notification information can be automatically sent to the relevant personnel based on the communication information obtained from the relevant personnel, so that the relevant personnel can promptly know the copy detection results of the above-mentioned original pictures by referring to the copy detection notification information, and then can take corresponding processing measures according to the results of the copy detection results.

[0150] In an embodiment of the present invention, the image identifier of the original image is obtained; then based on the copy detection result and the image identifier, copy detection notification information corresponding to the original image is generated; then the communication information of the relevant personnel is obtained; and subsequently based on the communication information, the copy detection notification information is sent to the relevant personnel. After the present invention performs copy discrimination processing on the target dual-channel image based on the copy discrimination model and obtains the copy detection result corresponding to the original image, it will intelligently obtain the image identifier of the original image, and generate copy detection notification information corresponding to the original image based on the copy detection result and the image identifier, and then automatically send the copy detection notification information to the relevant personnel according to the acquired communication information of the relevant personnel, so that the relevant personnel can timely know the copy detection result of the above-mentioned original image by consulting the copy detection notification information, and then can take corresponding processing measures according to the result content of the copy detection result, which is conducive to improving the work efficiency and work experience of the relevant personnel.

[0151] In some optional implementations, the user information obtained is subject to the user's consent and complies with relevant laws and policies.

[0152] Further references Figure 3, as a response to the above Figure 2 In order to realize the method shown in the figure, the present application provides an embodiment of an image copy detection device 300, which is similar to Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0153] An embodiment of the present invention provides an image copy detection device 300, and the image copy detection device 300 includes:

[0154] A first acquisition module 310 is used to acquire an original image to be detected;

[0155] A preprocessing module 320, configured to preprocess the original image to obtain a corresponding first image;

[0156] The removal module 330 is used to perform moiré removal processing on the first image based on a preset moiré removal model to obtain a corresponding second image;

[0157] A transformation module 340, configured to perform spatial-frequency transformation on the first image and the second image respectively to obtain a corresponding first spectrum image and a second spectrum image;

[0158] A merging module 350 is used to merge the first spectrum image and the second spectrum image to obtain a corresponding target dual-channel image;

[0159] The discrimination module 360 ​​is used to perform a duplication discrimination process on the target dual-channel image based on a preset duplication discrimination model to obtain a duplication detection result corresponding to the original image.

[0160] Optionally, the image copy detection device 300 further includes:

[0161] A second acquisition module is used to acquire pre-constructed dual-channel image data;

[0162] A labeling module, used for labeling the dual-channel image data to obtain corresponding sample data;

[0163] A partitioning module, used for partitioning the sample data into a training set and a test set;

[0164] A calling module is used to call a preset classification model;

[0165] A training module, used for training the classification model based on the training set to obtain a corresponding first model;

[0166] A testing module, used for testing the first model based on the test set to obtain a second model that meets the preset construction requirements;

[0167] The first determining module is used to use the second model as the copy discrimination model.

[0168] Optionally, the image copy detection device 300 further includes:

[0169] A second determining module, used to determine a designated storage mode corresponding to the remake discrimination model;

[0170] A storage module, used for storing the copy discrimination model based on the specified storage method;

[0171] A deployment module is used to deploy the copy discrimination model after completing the storage of the copy discrimination model.

[0172] Optionally, the merging module 350 includes:

[0173] A first processing submodule, configured to place the first spectrum image in a first channel of a preset initial dual-channel image, and place the second spectrum image in a second channel of the initial dual-channel image, to obtain a corresponding first dual-channel image;

[0174] A second processing submodule, configured to perform a size adjustment process on the first dual-channel image to obtain a corresponding second dual-channel image;

[0175] A third processing submodule, configured to perform format adjustment processing on the second dual-channel image to obtain a corresponding third dual-channel image;

[0176] The first determining submodule is configured to use the third dual-channel image as the target dual-channel image.

[0177] Optionally, the transformation module 340 includes:

[0178] An acquisition submodule, used to acquire a preset spatial-frequency domain transformation strategy;

[0179] A first transformation submodule, configured to perform a space-domain-frequency domain transformation on the first image based on the space-domain-frequency domain transformation strategy to obtain a corresponding first initial spectrum image;

[0180] A first adjustment submodule, configured to perform image adjustment processing on the first initial spectrum image to obtain the first spectrum image;

[0181] A second transformation submodule, configured to perform a spatial-frequency domain transformation on the second image based on the spatial-frequency domain transformation strategy to obtain a corresponding second initial spectrum image;

[0182] The second adjustment submodule is used to perform image adjustment processing on the second initial spectrum image to obtain the second spectrum image.

[0183] Optionally, the preprocessing module 320 includes:

[0184] A fourth processing submodule, configured to perform image resizing processing on the original image to obtain a corresponding first processed image;

[0185] a fifth processing submodule, configured to perform normalization processing on the first processed image to obtain a corresponding second processed image;

[0186] The second determining submodule is configured to use the second processed image as the first image.

[0187] Optionally, the image copy detection device 300 further includes:

[0188] A third acquisition module, used to acquire the image identifier of the original image;

[0189] A generating module, configured to generate copy detection notification information corresponding to the original image based on the copy detection result and the image identifier;

[0190] The fourth acquisition module is used to obtain the communication information of relevant personnel;

[0191] A sending module is used to send the copy detection notification information to the relevant personnel based on the communication information.

[0192] To solve the above technical problems, the present application also provides a computer device. Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.

[0193] The computer device 4 includes a memory 41, a processor 42, and a network interface 43 that are interconnected through a system bus. It should be noted that the figure only shows a computer device 4 with components 41-43, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0194] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0195] The memory 41 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 41 can be an internal storage unit of the computer device 4, such as a hard disk or memory of the computer device 4. In other embodiments, the memory 41 can also be an external storage device of the computer device 4, such as a plug-in hard disk equipped on the computer device 4, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. Of course, the memory 41 can also include both the internal storage unit of the computer device 4 and its external storage device. In this embodiment, the memory 41 is generally used to store the operating system and various application software installed on the computer device 4, such as the program code of the image remake detection method, etc. In addition, the memory 41 can also be used to temporarily store various types of data that have been output or are to be output.

[0196] The processor 42 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments. The processor 42 is generally used to control the overall operation of the computer device 4. In this embodiment, the processor 42 is used to run the program code stored in the memory 41 or process data, such as running the program code of the image remake detection method.

[0197] The network interface 43 may include a wireless network interface or a wired network interface. The network interface 43 is generally used to establish a communication connection between the computer device 4 and other electronic devices.

[0198] The present application also provides another embodiment, namely, providing a computer-readable storage medium, wherein the computer-readable storage medium stores the application crash processing program, and the application crash processing program can be executed by at least one processor to enable the at least one processor to perform the steps of the image copy detection method as described above.

[0199] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware online platform, and of course, by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0200] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0201] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. A method for detecting image copying, characterized in that: include: Obtain the original image to be detected; Preprocessing the original image to obtain a corresponding first image; Performing moiré removal processing on the first image based on a preset moiré removal model to obtain a corresponding second image; Performing spatial-frequency transformation on the first image and the second image respectively to obtain a corresponding first spectrum image and a second spectrum image; Merging the first spectrum image and the second spectrum image to obtain a corresponding target dual-channel image; The target dual-channel image is subjected to copy discrimination processing based on a preset copy discrimination model to obtain a copy detection result corresponding to the original image.

2. The method according to claim 1, characterized in that Before performing copy discrimination processing on the target dual-channel image based on a preset copy discrimination model to obtain a copy detection result corresponding to the original image, the method further includes: Get pre-built dual-channel image data; Performing labeling processing on the dual-channel image data to obtain corresponding sample data; Dividing the sample data into a training set and a test set; Call the preset classification model; Training the classification model based on the training set to obtain a corresponding first model; Testing the first model based on the test set to obtain a second model that meets preset construction requirements; The second model is used as the copy-make identification model.

3. The method according to claim 2, characterized in that After using the second model as the copy discrimination model, the method further includes: Determining a designated storage method corresponding to the copy discrimination model; Performing storage processing on the copy discrimination model based on the specified storage method; After the storage of the copy discrimination model is completed, the copy discrimination model is deployed.

4. The method according to claim 1, characterized in that: The merging of the first spectrum image and the second spectrum image to obtain a corresponding target dual-channel image includes: Placing the first spectrum image in a first channel of a preset initial dual-channel image, and placing the second spectrum image in a second channel of the initial dual-channel image, to obtain a corresponding first dual-channel image; Performing a size adjustment process on the first dual-channel image to obtain a corresponding second dual-channel image; Performing format adjustment processing on the second dual-channel image to obtain a corresponding third dual-channel image; The third dual-channel image is used as the target dual-channel image.

5. The method according to claim 1, characterized in that The performing spatial-frequency transformation on the first image and the second image respectively to obtain a corresponding first spectrum image and a second spectrum image includes: Obtaining a preset spatial-frequency domain transformation strategy; Performing a space-domain-frequency transformation on the first image based on the space-domain-frequency transformation strategy to obtain a corresponding first initial spectrum image; Performing image adjustment processing on the first initial spectrum image to obtain the first spectrum image; Performing a spatial-frequency domain transformation on the second image based on the spatial-frequency domain transformation strategy to obtain a corresponding second initial spectrum image; Performing image adjustment processing on the second initial spectrum image to obtain the second spectrum image.

6. The method according to claim 1, characterized in that The preprocessing of the original image to obtain a corresponding first image includes: Performing image resizing processing on the original image to obtain a corresponding first processed image; Normalizing the first processed image to obtain a corresponding second processed image; The second processed image is used as the first image.

7. The method according to claim 1, characterized in that After performing copy discrimination processing on the target dual-channel image based on the preset copy discrimination model to obtain a copy detection result corresponding to the original image, the method further includes: Obtaining an image identifier of the original image; generating, based on the copy detection result and the image identifier, copy detection notification information corresponding to the original image; Obtain communication information of relevant personnel; Based on the communication information, the copy detection notification information is sent to the relevant personnel.

8. An image copy detection device, characterized in that: include: A first acquisition module, used for acquiring an original image to be detected; A preprocessing module, used for preprocessing the original image to obtain a corresponding first image; an elimination module, configured to perform moiré elimination processing on the first image based on a preset moiré elimination model to obtain a corresponding second image; A transformation module, used to perform spatial-frequency transformation on the first image and the second image respectively to obtain a corresponding first spectrum image and a second spectrum image; a merging module, configured to merge the first spectrum image with the second spectrum image to obtain a corresponding target dual-channel image; The discrimination module is used to perform a copy discrimination process on the target dual-channel image based on a preset copy discrimination model to obtain a copy detection result corresponding to the original image.

9. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the image copying detection method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the image copying detection method according to any one of claims 1 to 7 are implemented.

Images