Image deduplication method and device and computer equipment

Abstract

The invention relates to an image deduplication method and device and computer equipment. The method comprises the steps of obtaining at least two to-be-processed images introduced into a target development project, calling a pre-deployed project development editor, performing similarity recognition on the to-be-processed images, obtaining a similarity value between every two to-be-processed images in the to-be-processed images, and performing duplicate removal processing on the to-be-processed images according to the similarity values, wherein the project development editor is integrated with an image similarity identification function and an image duplicate removal function. According to the image de-duplication method and device, an extra editor does not need to be called in the project development process, image de-duplication can be achieved before project development only by calling the project development editor with the project development function, image de-duplication processing is more convenient, and the image de-duplication speed and efficiency can be improved.

Description

Technical Field

[0001] This application relates to the field of web development technology, and in particular to an image deduplication method, apparatus, and computer equipment. Background Technology

[0002] With the development of project development technologies, the scale of project development is becoming increasingly larger. Taking web development as an example, during the execution of a web development project, it is inevitable to introduce multiple images that are duplicated or highly similar. Therefore, deduplicating multiple different images to improve web development performance becomes particularly important.

[0003] In related technologies, the main approach is to manually traverse the images by using their names and sizes to distinguish the similarity between pairs of images, and then remove duplicate images based on the similarity between them.

[0004] However, related technologies suffer from low image deduplication efficiency during the image deduplication process. Summary of the Invention

[0005] Therefore, it is necessary to provide an image deduplication method, apparatus, and computer equipment to address the aforementioned technical problems and improve the efficiency of image deduplication.

[0006] In a first aspect, embodiments of this application provide an image deduplication method, the method comprising:

[0007] Obtain at least two images to be processed from the target development project;

[0008] The pre-deployed project development editor is invoked to perform similarity recognition on each image to be processed, obtain the similarity value between each pair of images to be processed, and perform deduplication processing on each image to be processed based on the similarity value.

[0009] The project development editor integrates image similarity recognition and image deduplication functions.

[0010] The technical solution in this application embodiment acquires at least two images to be processed introduced into the target development project, calls a pre-deployed project development editor, performs similarity recognition on each image to be processed, obtains the similarity value between each pair of images to be processed, and performs deduplication processing on each image to be processed based on the similarity value. The project development editor integrates image similarity recognition and image deduplication functions. This method does not require calling an additional editor during project development; image deduplication can be achieved before project development simply by calling a project development editor with project development functions, making image deduplication processing more convenient. This method improves the speed and efficiency of image deduplication. Furthermore, it eliminates the need for manual intervention, enabling image deduplication during project development by utilizing a project development editor. This not only significantly speeds up deduplication, making subsequent project development more efficient, but also reduces the manpower required for deduplication and improves its accuracy. Additionally, the project development editor integrates image similarity recognition and deduplication functions, enhancing its functionality and broad applicability.

[0011] In one embodiment, similarity recognition is performed on each image to be processed to obtain similarity values ​​between each pair of images to be processed, including:

[0012] Based on a deep learning model, feature extraction is performed on each image to be processed to obtain the feature vector of each image to be processed.

[0013] Similarity processing is performed on the feature vectors of each image to be processed to determine the similarity value between each pair of images to be processed.

[0014] The technical solution in this application embodiment extracts features from each image to be processed based on a deep learning model to obtain feature vectors for each image to be processed, and performs similarity processing based on the feature vectors of each image to be processed to determine the similarity value between each pair of images to be processed. The above method uses a deep learning model to identify the similarity of each image to be processed. Compared with the traditional technology of distinguishing image similarity by image name or size, it can minimize the discrimination error and greatly improve the accuracy of similarity recognition, thus providing reference information for improving the accuracy of image deduplication in the future.

[0015] In one embodiment, feature extraction is performed on each image to be processed based on a deep learning model to obtain the feature vector of each image to be processed, including:

[0016] Each image to be processed is preprocessed to obtain the corresponding preprocessed image; the preprocessing includes at least one of the following: size adjustment, grayscale conversion, and noise reduction.

[0017] Each preprocessed image is input into a deep learning model to obtain the feature vector of each image to be processed.

[0018] The technical solution in this application embodiment preprocesses each image to be processed to obtain a corresponding preprocessed image, and inputs each preprocessed image into a deep learning model to obtain the feature vector of each image to be processed. The preprocessing includes at least one of size adjustment processing, grayscale processing, and noise reduction processing. The above-mentioned method can first preprocess the image and then extract features from the preprocessed image, thereby improving the accuracy of the obtained feature vectors of each image to be processed, providing reference information for improving the accuracy of the similarity processing results in the future. At the same time, this process does not require complex algorithm implementation, thereby reducing the complexity of the image preprocessing process and improving the speed and efficiency of image preprocessing.

[0019] In one embodiment, deduplication is performed on each image to be processed based on its similarity value, including:

[0020] Based on each similarity value, determine the similarity comparison score between each pair of images to be processed in each image to be processed;

[0021] Based on the similarity comparison scores, at least one target image in each image to be processed is classified to determine at least one target classification image set;

[0022] Based on the image sets of each target category, deduplication is performed on each target image to complete the deduplication of each image to be processed.

[0023] The technical solution in this application embodiment determines the similarity comparison score between each pair of images to be processed based on each similarity value, and classifies at least one target image in each image to be processed based on each similarity comparison score to determine at least one target classification image set. Then, deduplication processing is performed on each target image based on each target classification image set to complete the deduplication processing of each image to be processed. The above method can obtain the similarity comparison score between each pair of images to be processed, and then obtain the target classification image set containing the target image based on the similarity comparison score between each pair of images to be processed, and perform image deduplication processing on the target classification image set. This can uniformly perform deduplication processing on newly introduced images and already introduced images, minimize the introduction of multiple identical or similar images during the development of the target project, improve the accuracy of image deduplication during the development of the target project, and prepare for further improving the accuracy of the development results of the target project. At the same time, the above method does not require the participation of complex algorithms, and the processing process is relatively simple, thereby improving the speed, efficiency and accuracy of image deduplication.

[0024] In one embodiment, each image to be processed is a newly introduced image; based on each similarity value, a similarity comparison score is determined between each pair of images to be processed, including:

[0025] The similarity values ​​are searched in a pre-built mapping table, and the scores that match each similarity value in the mapping table are determined as the similarity comparison scores between each pair of newly introduced images. The mapping table includes the association between different similarity values ​​and different scores.

[0026] The technical solution in this application embodiment searches for each similarity value in a pre-constructed mapping table, and determines the score that matches each similarity value in the mapping table as the similarity comparison score between each pair of newly introduced images. The above method can quickly obtain the similarity comparison score between different images by searching through a pre-constructed mapping table. The processing is relatively simple and can also improve the speed and efficiency of information retrieval.

[0027] In one embodiment, each image to be processed includes a newly introduced image and multiple introduced images from at least one classification image set; based on each similarity value, a similarity comparison score is determined between each pair of images to be processed, including:

[0028] For any newly introduced image, the similarity value between the newly introduced image and each of the previously introduced images is found in the pre-built mapping table, and the score that matches each similarity value in the mapping table is determined as the corresponding similarity score between the newly introduced image and each of the previously introduced images.

[0029] Based on the similarity score between the newly introduced image and each introduced image in each category image set, the similarity comparison score between the newly introduced image and each category image set is determined.

[0030] The technical solution in this application embodiment, for any newly introduced image, searches for the similarity value between the newly introduced image and each already introduced image in a pre-constructed mapping table, and determines the score that matches each similarity value in the mapping table as the corresponding similarity score between the newly introduced image and each already introduced image. Based on the similarity score between the newly introduced image and each already introduced image in each category image set, the similarity comparison score between the newly introduced image and each category image set is determined. The above method can determine the similarity comparison score between the newly introduced image and the category image set in each image to be processed according to whether all images to be processed are newly introduced images or already introduced images, thereby improving the accuracy of the similarity comparison score between the newly introduced image and the category image set obtained in this scenario. At the same time, this process does not require the participation of complex algorithms and the processing process is relatively simple, thereby improving the speed, efficiency and accuracy of the obtained similarity comparison score between the newly introduced image and the category image set.

[0031] In one embodiment, based on the similarity comparison scores, at least one target image in each image to be processed is classified to determine at least one target classification image set, including:

[0032] Based on the similarity comparison scores, multiple similarity comparison scores are determined for each target image;

[0033] For any target image, obtain the maximum similarity comparison score from the similarity comparison scores corresponding to the target image;

[0034] Based on the maximum similarity comparison score, the target image is assigned to the target classification image set.

[0035] The technical solution in this application embodiment determines multiple similarity comparison scores corresponding to each target image based on each similarity comparison score. For any target image, the maximum similarity comparison score is obtained from the similarity comparison scores corresponding to the target image, and the target image is assigned to the target classification image set based on the maximum similarity comparison score. The above method can classify all newly introduced images into target classification image sets based on the similarity comparison scores between pairs of images to be processed, thus preparing for subsequent deduplication of images with high similarity, improving the accuracy of image deduplication. At the same time, this process does not require complex algorithms and is relatively simple, thereby improving the speed, efficiency and accuracy of similar image classification.

[0036] In one embodiment, each target image includes at least one newly introduced image; deduplication is performed on each target image according to each target classification image set, including:

[0037] For any target classification image set, if the target classification image set is any one of multiple classified image sets, then the original position order of each image is adjusted according to the attribute information of each image in the target classification image set and the requirement information of the target development project to obtain the adjusted image set; each image in the target classification image set includes at least one newly introduced image.

[0038] Based on the current position order of each image in the adjusted image set, deduplication is performed on each image in the adjusted image set to complete the deduplication of each image to be processed.

[0039] The technical solution in this application embodiment, for any target classification image set, adjusts the original position order of each image according to the attribute information of each image in the target classification image set and the requirement information of the target development project to obtain an adjusted image set, and performs deduplication processing on each image in the adjusted image set according to the current position order of each image in the adjusted image set to complete the deduplication processing of each target image. The target classification image set includes at least one target image. The above method can perform deduplication processing on newly introduced images according to image attribute information and actual project development requirements, so that the application scenario of image deduplication processing fits the actual development project, thereby improving the accuracy of image deduplication during the execution of the target development project.

[0040] In one embodiment, the method further includes:

[0041] In the case of deduplicating each image to be processed, the target development project is executed based on the multiple valid images obtained after image deduplication.

[0042] The technical solution in this application embodiment, after deduplicating each image to be processed, executes the target development project based on the multiple valid images obtained after image deduplication. The above method can execute the target development project based on the multiple valid images obtained after image deduplication, which can reduce unnecessary errors introduced during project development and thus improve the correctness of project development results. At the same time, the above method can reduce the size of the project, speed up project development, improve project performance, and enhance user experience based on image deduplication.

[0043] Secondly, embodiments of this application provide an image deduplication device, the device comprising:

[0044] The image acquisition module is used to acquire at least two images to be processed introduced into the target development project;

[0045] The deduplication module is used to call the pre-deployed project development editor to perform similarity recognition on each image to be processed, obtain the similarity value between each pair of images to be processed, and perform deduplication on each image to be processed based on the similarity value.

[0046] The project development editor integrates image similarity recognition and image deduplication functions.

[0047] Thirdly, embodiments of this application also provide a computer device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the method in any of the embodiments of the first aspect described above.

[0048] Fourthly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the method in any of the embodiments of the first aspect described above.

[0049] Fifthly, embodiments of this application also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the method in any of the embodiments of the first aspect described above.

[0050] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0051] Figure 1 This is a flowchart illustrating an image deduplication method in one embodiment;

[0052] Figure 2 This is a flowchart illustrating the image deduplication method in another embodiment;

[0053] Figure 3 This is a flowchart illustrating the image deduplication method in another embodiment;

[0054] Figure 4 This is a flowchart illustrating the image deduplication method in another embodiment;

[0055] Figure 5 This is a flowchart illustrating the image deduplication method in another embodiment;

[0056] Figure 6 This is a flowchart illustrating the image deduplication method in another embodiment;

[0057] Figure 7This is a flowchart illustrating the image deduplication method in another embodiment;

[0058] Figure 8 This is a structural block diagram of an image deduplication device in one embodiment;

[0059] Figure 9 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0060] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0061] In mobile application development, website development, front-end development, desktop application development, game development, and other fields, projects inevitably introduce multiple images that are duplicated or highly similar. Therefore, deduplication of multiple different images to improve project development performance becomes particularly important. Related technologies primarily rely on manually traversing the images by name and size to distinguish the similarity between pairs of images, and then deduplicating images based on this similarity. However, these technologies suffer from low image deduplication efficiency. Therefore, this application provides an image deduplication method that can improve the speed of image deduplication.

[0062] The image deduplication method provided in this application can be applied to computer devices, such as personal computers, laptops, or desktop computers, and can also be electronic devices with software development capabilities. This embodiment does not limit the specific form of the computer device. The following embodiments of this application use a computer device as the executing subject of the image deduplication method to describe the specific process of the image deduplication method.

[0063] like Figure 1 The diagram shown is a flowchart of an image deduplication method provided in an embodiment of this application. This method can be implemented through the following steps:

[0064] S100: Obtain at least two images to be processed from the target development project.

[0065] In practical applications, the aforementioned target development project can be a mobile application development project, a website development project, a desktop application development project, a game development project, etc. In this embodiment of the application, the target development project is a web page development project (i.e., a Web development project) as an example for illustration.

[0066] The at least two images to be processed can both be newly introduced images, or they can be at least one newly introduced image and at least one previously introduced image. Optionally, the different images to be processed can have the same size or different sizes.

[0067] In the embodiments of this application, the computer device can obtain at least two images to be processed that are introduced into the target development project at one time from the image library, or it can obtain at least two images to be processed that are introduced into the target development project multiple times from the image library.

[0068] S200: Invoke the pre-deployed project development editor to perform similarity recognition on each image to be processed, obtain the similarity value between each pair of images to be processed, and perform deduplication processing on each image to be processed based on the similarity value. The project development editor integrates image similarity recognition and image deduplication functions.

[0069] Prior to performing the steps in S200 above, a project development editor may be deployed (or installed) on the computer device. Optionally, the project development editor can be understood as a tool or application for executing the target development project, and in addition to having project development functions, it also extends with several other functions.

[0070] Specifically, the aforementioned project development editor can be understood as an integrated editor that integrates multiple functional units, such as image detection, image enhancement, and image segmentation functions. These functional units can be integrated into the project development editor as plugins. In this embodiment, the multiple functional units integrated into the project development editor may include an image similarity recognition unit and an image deduplication unit. The image similarity recognition unit can implement image similarity recognition, and the image deduplication unit can implement image deduplication. It is understood that the aforementioned project development editor not only has project development functions but also image similarity recognition and image deduplication functions.

[0071] Specifically, the computer device can call a pre-deployed project development editor and use the project development editor to perform similarity recognition on each image to be processed, obtain the similarity value between each pair of images to be processed, and then perform deduplication processing on each image to be processed based on each similarity value.

[0072] In one embodiment, the similarity recognition of each image to be processed can be performed by pre-training an algorithm model, and then inputting each image to be processed into the algorithm model, which outputs the similarity value between each pair of images to be processed.

[0073] In another embodiment, the similarity recognition of each image to be processed can also be performed by using a similarity recognition algorithm. Optionally, the similarity recognition algorithm can be a histogram algorithm, a grayscale image algorithm, a hash algorithm, a cosine similarity algorithm, and / or an Euclidean distance algorithm, etc.

[0074] One method for deduplicating each image based on its similarity value is to obtain multiple similarity values ​​with larger values ​​from all similarity values ​​and delete any one of the pairs of images with these similarity values ​​to complete the deduplication process.

[0075] It should be noted that during the execution of the target development project, the steps in S100-S200 above can be followed to remove duplicate or highly similar images introduced in the target development project, so that the output results after the target development project is executed are more accurate.

[0076] The technical solution in this application embodiment acquires at least two images to be processed introduced into the target development project, calls a pre-deployed project development editor, performs similarity recognition on each image to be processed, obtains the similarity value between each pair of images to be processed, and performs deduplication processing on each image to be processed based on the similarity value. The project development editor integrates image similarity recognition and image deduplication functions. This method does not require calling an additional editor during project development; image deduplication can be achieved before project development simply by calling a project development editor with project development functions, making image deduplication processing more convenient. This method improves the speed and efficiency of image deduplication. Furthermore, it eliminates the need for manual intervention, enabling image deduplication during project development by utilizing a project development editor. This not only significantly speeds up deduplication, making subsequent project development more efficient, but also reduces the manpower required for deduplication and improves its accuracy. Additionally, the project development editor integrates image similarity recognition and deduplication functions, enhancing its functionality and broad applicability.

[0077] The process of performing similarity recognition on each image to obtain the similarity value between each pair of images to be processed will be described below. In one embodiment, as... Figure 2 As shown, the steps in S200 above can be implemented in the following ways:

[0078] S210. Based on a deep learning model, feature extraction is performed on each image to be processed to obtain the feature vector of each image to be processed.

[0079] In practical applications, computer devices can input each image to be processed into a deep learning model. This deep learning model extracts features from each image and outputs a feature vector for each image. Optionally, the feature extraction can include processes for extracting color features, texture features, shape features, etc.

[0080] Optionally, the aforementioned deep learning model may be composed of at least one of the following: deep neural network model, deep belief network model, variational autoencoder, residual neural network model, convolutional neural network model, bidirectional long short-term memory network model, recurrent neural network model, and long short-term memory network model.

[0081] The network structure of the deep learning model may include at least one VGG series network structure, but in this embodiment, the network structure of the deep learning model may include VGG16.

[0082] S220. Perform similarity processing based on the feature vectors of each image to be processed, and determine the similarity value between each pair of images to be processed.

[0083] Specifically, the computer device can employ a similarity algorithm to perform similarity processing based on the feature vectors of each pair of images to be processed, i.e., similarity calculation, to obtain the similarity value between each pair of images to be processed. Optionally, the aforementioned similarity algorithm can be a cosine similarity algorithm, a correlation coefficient method, a Manhattan distance method, etc.

[0084] The technical solution in this application embodiment extracts features from each image to be processed based on a deep learning model to obtain feature vectors for each image to be processed, and performs similarity processing based on the feature vectors of each image to be processed to determine the similarity value between each pair of images to be processed. The above method uses a deep learning model to identify the similarity of each image to be processed. Compared with the traditional technology of distinguishing image similarity by image name or size, it can minimize the discrimination error and greatly improve the accuracy of similarity recognition, thus providing reference information for improving the accuracy of image deduplication in the future.

[0085] In some scenarios, to significantly improve the accuracy of image feature extraction, the image can be preprocessed before feature extraction. The following describes the feature extraction process for each image to be processed. In one embodiment, such as... Figure 3 As shown, the step in S210 above, which extracts features from each image to be processed based on a deep learning model to obtain the feature vector of each image to be processed, can be implemented in the following way:

[0086] S211. Preprocess each image to be processed to obtain the corresponding preprocessed image. The preprocessing includes at least one of the following: size adjustment, grayscale conversion, and noise reduction.

[0087] For any image to be processed, the computer device can use an open-source image processing library to preprocess the image to obtain the corresponding preprocessed image. In the embodiments of this application, the preprocessing may include at least one of resizing, grayscale conversion, and noise reduction.

[0088] The aforementioned size adjustment processing can be achieved through image segmentation or image cropping algorithms; the aforementioned grayscale processing can be achieved through the maximum value method, the average value method, or the weighted average method; the aforementioned denoising processing can be achieved through mean filters, geometric mean filters, harmonic mean filters, inverse harmonic mean filters, adaptive Wiener filters, median filters, nonlocal mean algorithms, block matching-based three-dimensional filtering algorithms, etc.

[0089] S212. Input each preprocessed image into the deep learning model to obtain the feature vector of each image to be processed.

[0090] Furthermore, based on the preprocessed images corresponding to each image to be processed obtained in the previous steps, each preprocessed image can be input into a deep learning model to obtain the feature vector of each image to be processed.

[0091] The technical solution in this application embodiment preprocesses each image to be processed to obtain a corresponding preprocessed image, and inputs each preprocessed image into a deep learning model to obtain the feature vector of each image to be processed. The preprocessing includes at least one of size adjustment processing, grayscale processing, and noise reduction processing. The above-mentioned method can first preprocess the image and then extract features from the preprocessed image, thereby improving the accuracy of the obtained feature vectors of each image to be processed, providing reference information for improving the accuracy of the similarity processing results in the future. At the same time, this process does not require complex algorithm implementation, thereby reducing the complexity of the image preprocessing process and improving the speed and efficiency of image preprocessing.

[0092] The process of deduplicating each image based on its similarity value is described below. In one embodiment, as... Figure 4 As shown, the steps in S300 above can be implemented in the following ways:

[0093] S310. Based on each similarity value, determine the similarity comparison score between each pair of images to be processed in each image to be processed.

[0094] Specifically, the computer device can pre-train an algorithm model. Then, for any two images to be processed, the similarity value between the two images is input into the algorithm model, which outputs a similarity comparison score between the two images. Here, the similarity comparison score can be understood as a quantified value of similarity. Optionally, the similarity comparison score can be any value between 0 and 100.

[0095] In some scenarios, all images to be processed are newly introduced images. The specific process for obtaining the similarity comparison scores between pairs of images to be processed in such scenarios is described below. In one embodiment, all images to be processed are newly introduced images; the step in S310 above, which determines the similarity comparison scores between pairs of images to be processed based on each similarity value, may include: searching for each similarity value in a pre-constructed mapping table, and determining the scores in the mapping table that match each similarity value as the similarity comparison scores between pairs of newly introduced images. The mapping table includes the association between different similarity values ​​and different scores.

[0096] In practical applications, if each image to be processed is a newly introduced image, the computer device can look up the similarity value between each pair of newly introduced images in a pre-built mapping table, and determine the score that matches each similarity value in the mapping table as the similarity comparison score between each pair of newly introduced images.

[0097] This embodiment can look up each similarity value in a pre-built mapping table, and determine the score that matches each similarity value in the mapping table as the similarity comparison score between each pair of newly introduced images. The above method can quickly obtain the similarity comparison score between different images by using a search method through a pre-built mapping table. The processing is relatively simple and can also improve the speed and efficiency of information retrieval.

[0098] S320. Based on the similarity comparison scores, classify at least one target image in each image to be processed, and determine at least one target classification image set.

[0099] Specifically, the computer device can, based on all acquired similarity comparison scores, identify at least one image to be processed whose similarity comparison score falls within a preset score threshold range. These images are then used as target images. Next, similarity comparison scores that are identical or approximately equal among the target images are grouped into the same category. Finally, the target images corresponding to similarity comparison scores within the same category are grouped into the same category, thus completing the classification of the target images and obtaining at least one target classification image set. Optionally, the aforementioned preset score threshold range can be flexibly determined according to actual project requirements, broadening the application scope of the image deduplication scheme.

[0100] Optionally, each target classification image set may include at least one target image. Images within the same target classification image set exhibit high similarity, while images within different target classification image sets show low similarity.

[0101] S330. Based on the image sets of each target classification, perform deduplication processing on each target image to complete the deduplication processing of each image to be processed.

[0102] Based on the target classification image sets obtained in the preceding steps, duplicate images can be removed from each target image set. It's important to note that removing duplicate images from each target classification image set can be understood as removing or deleting images from each target classification image set. In practical applications, for any target classification image set, either the target images within the set can be deleted, or other images outside the target classification image set can be deleted.

[0103] The technical solution in this application embodiment determines the similarity comparison score between each pair of images to be processed based on each similarity value, and classifies at least one target image in each image to be processed based on each similarity comparison score to determine at least one target classification image set. Then, deduplication processing is performed on each target image based on each target classification image set to complete the deduplication processing of each image to be processed. The above method can obtain the similarity comparison score between each pair of images to be processed, and then obtain the target classification image set containing the target image based on the similarity comparison score between each pair of images to be processed, and perform image deduplication processing on the target classification image set. This can uniformly perform deduplication processing on newly introduced images and already introduced images, minimize the introduction of multiple identical or similar images during the development of the target project, improve the accuracy of image deduplication during the development of the target project, and prepare for further improving the accuracy of the development results of the target project. At the same time, the above method does not require the participation of complex algorithms, and the processing process is relatively simple, thereby improving the speed, efficiency and accuracy of image deduplication.

[0104] In other scenarios, each image to be processed includes both newly introduced images and previously introduced images. In this case, an alternative method can be used to obtain the similarity comparison scores between each pair of images to be processed. The specific process of this method is described below. In one embodiment, each image to be processed includes one newly introduced image and at least one set of previously introduced images from a classification image set; such as... Figure 5 As shown, the step in S310 above, which determines the similarity comparison score between each pair of images to be processed based on each similarity value, can be implemented in the following way:

[0105] S311. For any newly introduced image, find the similarity value between the newly introduced image and each of the previously introduced images in the pre-constructed mapping table, and determine the score that matches each similarity value in the mapping table as the corresponding similarity score between the newly introduced image and each of the previously introduced images.

[0106] In practical applications, if each image to be processed includes a newly introduced image and multiple introduced images from at least one classification image set, then for any newly introduced image, the computer device can look up the similarity value between the newly introduced image and each introduced image in a pre-built mapping table, and determine the score in the mapping table that matches these similarity values ​​as the corresponding similarity score between the newly introduced image and each of the corresponding introduced images.

[0107] It should be noted that the similarity value between the newly introduced image and each of the previously introduced images can be selected from the similarity values ​​between each pair of images to be processed.

[0108] S312. Based on the similarity score between the newly introduced image and each introduced image in each category image set, determine the similarity comparison score between the newly introduced image and each category image set.

[0109] It should be noted that the aforementioned image sets for classification can be image sets generated before the current time. Specifically, if there is only one generated image set, the computer device can execute an attention mechanism once, that is, perform arithmetic operations on the similarity scores between the newly introduced image and each already introduced image in the image set to obtain a similarity comparison score between the newly introduced image and the image set. Optionally, the arithmetic operations can include at least one implementation of addition, subtraction, division, multiplication, logarithmic operations, exponential operations, etc.

[0110] In addition, the computer device can also obtain the maximum, minimum, mode, or median value from the similarity scores between the newly introduced image and each introduced image in the classification image set, and determine the maximum, minimum, mode, or median value as the similarity comparison score between the newly introduced image and the classification image set.

[0111] Meanwhile, if multiple classification image sets have been generated, the computer device can execute two attention mechanisms. The first attention mechanism may include: for any classification image set, performing arithmetic operations on the similarity scores between the newly introduced image and each already introduced image in that classification image set to obtain a similarity comparison score between the newly introduced image and the classification image set; the second attention mechanism may include: performing arithmetic operations on the similarity comparison scores between the newly introduced image and the classification image set obtained in the previous steps to obtain a similarity comparison score between the newly introduced image and each classification image set.

[0112] In practical applications, the computer can also respond to interface display commands and output a display interface. This output interface can display information such as the similarity comparison scores between each newly introduced image and the categorized image set, the generated categorized image sets, and image deduplication controls, thereby improving the user experience. Specifically, based on the similarity comparison scores displayed on the output interface, it can be determined whether deduplication processing is needed for each newly introduced image and the images within the categorized image sets. If so, the computer device can receive and respond to the image deduplication command input by the user based on the image deduplication controls, and perform deduplication on each newly introduced image and the images within the categorized image sets.

[0113] The technical solution in this application embodiment, for any newly introduced image, searches for the similarity value between the newly introduced image and each already introduced image in a pre-constructed mapping table, and determines the score that matches each similarity value in the mapping table as the corresponding similarity score between the newly introduced image and each already introduced image. Based on the similarity score between the newly introduced image and each already introduced image in each category image set, the similarity comparison score between the newly introduced image and each category image set is determined. The above method can determine the similarity comparison score between the newly introduced image and the category image set in each image to be processed according to whether all images to be processed are newly introduced images or already introduced images, thereby improving the accuracy of the similarity comparison score between the newly introduced image and the category image set obtained in this scenario. At the same time, this process does not require the participation of complex algorithms and the processing process is relatively simple, thereby improving the speed, efficiency and accuracy of the obtained similarity comparison score between the newly introduced image and the category image set.

[0114] The process of classifying at least one target image set from each image to be processed based on the similarity comparison scores is described below. In one embodiment, as... Figure 6 As shown, the steps in S320 above can be implemented in the following ways:

[0115] S321. Based on the similarity comparison scores, determine the multiple similarity comparison scores corresponding to each target image.

[0116] It should be noted that each of the target images mentioned above can be a newly introduced image within the image to be processed. In practical applications, the similarity score between each newly introduced image and other newly introduced or already introduced images can be obtained from all similarity comparison scores.

[0117] S322. For any target image, obtain the maximum similarity comparison score from the similarity comparison scores corresponding to the target image.

[0118] In practical applications, each target image can correspond to multiple similarity comparison scores. That is, for any target image, it can include the similarity comparison scores between the target image and other newly introduced or already introduced images.

[0119] Specifically, for any target image, the similarity comparison scores of the target image can be sorted according to their size, and the maximum similarity comparison score of the target image can be obtained from the sorting results.

[0120] Alternatively, the maximum similarity score of the target image can be obtained by maximizing the similarity comparison score of the target image.

[0121] S323. Based on the maximum similarity comparison score, classify the target image into the target classification image set.

[0122] In practical applications, computer equipment can pre-train an algorithm model, and then input the maximum similarity comparison score corresponding to the target image into the algorithm model to classify the target image into a target classification image set. Optionally, the aforementioned target classification image set can be any of all classification image sets generated before the current time, or it can be a new classification image set other than all classification image sets generated before the current time.

[0123] It should be noted that different image sets are configured with corresponding preset score threshold ranges. In this embodiment, the computer device can also determine whether the maximum similarity comparison score of the target image belongs to any range of the preset score threshold range of each image set. If it does, the image set corresponding to the range of the maximum similarity comparison score is determined as the target image set, and the target image is assigned to the target image set.

[0124] In addition, if the maximum similarity comparison score corresponding to the target image does not fall within the preset score threshold range of each category image set, a new category image set is constructed and the new category image set is determined as the target category image set. Then the target image is assigned to the target category image set.

[0125] The technical solution in this application embodiment determines multiple similarity comparison scores corresponding to each target image based on each similarity comparison score. For any target image, the maximum similarity comparison score is obtained from the similarity comparison scores corresponding to the target image, and the target image is assigned to the target classification image set based on the maximum similarity comparison score. The above method can classify all newly introduced images into target classification image sets based on the similarity comparison scores between pairs of images to be processed, thus preparing for subsequent deduplication of images with high similarity, improving the accuracy of image deduplication. At the same time, this process does not require complex algorithms and is relatively simple, thereby improving the speed, efficiency and accuracy of similar image classification.

[0126] The process of deduplicating each target image based on the target classification image set described above is explained below. In one embodiment, as... Figure 7 As shown, the steps in S330 above can be implemented in the following ways:

[0127] S331. For any target classification image set, based on the attribute information of each image in the target classification image set and the requirement information of the target development project, the original positional order of each image is adjusted to obtain an adjusted image set. Each image in the target classification image set includes at least one target image.

[0128] It should be noted that the aforementioned multiple classified image sets can be understood as all classified image sets generated before the current moment. Optionally, the attribute information of the aforementioned images may include information such as image resolution, content, color, saturation, and brightness; the requirement information of the aforementioned target development project may include the attributes of images required at different locations in the target development project and the number of different types of images required in the target development project, and these attributes may include information such as resolution, content, color, saturation, and brightness. In this scenario, the aforementioned target image can be a newly introduced image.

[0129] In this embodiment, for any target classification image set, the computer device can adjust the position order of images in the target classification image set that satisfy the requirements of the target development project (i.e., adjust the position order of images that satisfy the requirements of the target development project from their original position order to their current position order) based on the attribute information of each image in the target classification image set, and adjust the position order of images that do not satisfy the requirements of the target development project to their reverse order. Optionally, after the position order of the current images in the target classification image set that satisfy the requirements of the target development project is adjusted, the corresponding current position order can be a first order, a second order, etc., but the current position order of the current image is greater than the original position order of the current image in the target classification image set.

[0130] In addition, the computer device can pre-train a positional order adjustment model, and then input the attribute information of each image in the target classification image set and the requirement information of the target development project into the positional order adjustment model. The positional order adjustment model adjusts the original positional order of each image and outputs the adjusted image set. Optionally, the above-mentioned positional order adjustment model can be implemented by at least one of the following: convolutional neural network model, fully connected neural network model, residual neural network model, long short-term memory neural network model, recurrent recurrent neural network model, etc.

[0131] S332. Based on the current position order of each image in the adjusted image set, perform deduplication processing on each image in the adjusted image set to complete the deduplication processing of each target image.

[0132] Based on the adjusted image sets obtained in the previous steps, for any adjusted image set, the images in the adjusted image set can be deduplicated according to the current position order of each image in the adjusted image set, so as to complete the deduplication of each target image.

[0133] One method for deduplicating images in the adjusted image set is to retain only the image corresponding to the first position in the adjusted image set according to the current position order of each image, and delete all other images after the first position, so as to complete the deduplication of the adjusted image set, that is, to complete the deduplication of each target image.

[0134] Meanwhile, another method for deduplicating images in the adjusted image set is to delete at least one image from the adjusted image set based on the current position order of the images and the number of images of the corresponding type required in the target development project, thereby completing the deduplication of each target image. Optionally, the number of images remaining after deleting images in the adjusted image set can be equal to the number of images of the same type required in the target development project.

[0135] The technical solution in this application embodiment, for any target classification image set, adjusts the original position order of each image according to the attribute information of each image in the target classification image set and the requirement information of the target development project to obtain an adjusted image set, and performs deduplication processing on each image in the adjusted image set according to the current position order of each image in the adjusted image set to complete the deduplication processing of each target image. The target classification image set includes at least one target image. The above method can perform deduplication processing on newly introduced images according to image attribute information and actual project development requirements, so that the application scenario of image deduplication processing fits the actual development project, thereby improving the accuracy of image deduplication during the execution of the target development project.

[0136] In practical applications, after image deduplication is completed, a target development project can be executed based on the multiple valid images obtained after image deduplication. The process is described below. In one embodiment, after performing the steps in S200 above, the method may further include: in the case of deduplicating each image to be processed, executing the target development project based on the multiple valid images obtained after image deduplication.

[0137] Specifically, after deduplicating at least two images to be processed in the target development project, the computer equipment can obtain multiple valid images, and then execute the target development project based on the multiple valid images using the project development method.

[0138] Optionally, the number of valid images can be less than or equal to the number of images to be processed introduced into the target development project. Optionally, the project development method can be a waterfall model, iterative model, spiral model, incremental model, prototyping model, etc.

[0139] The technical solution in this application embodiment, after deduplicating each image to be processed, executes the target development project based on the multiple valid images obtained after image deduplication. The above method can execute the target development project based on the multiple valid images obtained after image deduplication, which can reduce unnecessary errors introduced during project development and thus improve the correctness of project development results. At the same time, the above method can reduce the size of the project, speed up project development, improve project performance, and enhance user experience based on image deduplication.

[0140] In one embodiment, this application also provides an image deduplication method applied to a computer device, the method comprising the following steps:

[0141] (1) Obtain at least two images to be processed from the target development project;

[0142] (2) Call the pre-deployed project development editor to perform similarity recognition on each image to be processed, obtain the similarity value between each pair of images to be processed, and perform deduplication processing on each image to be processed according to each similarity value; wherein, the project development editor integrates image similarity recognition function and image deduplication function.

[0143] The process of performing similarity recognition on each image to be processed in step (2) above to obtain the similarity value between each pair of images to be processed can be achieved through the following steps:

[0144] (21) Preprocess each image to be processed to obtain the corresponding preprocessed image; the preprocessing includes at least one of size adjustment processing, grayscale processing, and noise reduction processing;

[0145] (22) Input each preprocessed image into the deep learning model to obtain the feature vector of each image to be processed;

[0146] (23) Perform similarity processing based on the feature vectors of each image to be processed, and determine the similarity value between each pair of images to be processed in each image to be processed;

[0147] The process of deduplicating each image based on its similarity value in step (2) above can be achieved through the following steps:

[0148] (24) Based on each similarity value, determine the similarity comparison score between each pair of images to be processed in each image to be processed;

[0149] The above step (24) can be implemented in any of the following ways:

[0150] The first method includes:

[0151] (241) If each image to be processed is a newly introduced image, then find each similarity value in the pre-constructed mapping table, and determine the score that matches each similarity value in the mapping table as the similarity comparison score between each pair of newly introduced images in each newly introduced image; the mapping table includes the association between different similarity values ​​and different scores;

[0152] The second method includes:

[0153] (242) If each image to be processed includes a newly introduced image and multiple introduced images from at least one classification image set, then for any newly introduced image, the similarity value between the newly introduced image and each introduced image is found in the pre-constructed mapping table, and the score that matches each similarity value in the mapping table is determined as the corresponding similarity score between the newly introduced image and each introduced image.

[0154] (243) For any classification image set, determine the similarity comparison score between the newly introduced image and the classification image set based on the similarity score between the newly introduced image and each introduced image in the classification image set;

[0155] (25) Based on the similarity comparison scores, determine the multiple similarity comparison scores corresponding to each target image;

[0156] (26) For any target image, obtain the maximum similarity comparison score from the similarity comparison scores corresponding to the target image;

[0157] (27) Based on the maximum similarity comparison score, classify the target image into the target classification image set;

[0158] (28) For any target classification image set, the original position order of each image is adjusted according to the attribute information of each image in the target classification image set and the requirement information of the target development project to obtain the adjusted image set; each image in the target classification image set includes at least one target image;

[0159] (29) Based on the current position order of each image in the adjusted image set, perform deduplication processing on each image in the adjusted image set to complete the deduplication processing of each target image;

[0160] (3) In the case of deduplication of each image to be processed, the target development project is executed based on the multiple valid images obtained after image deduplication.

[0161] For details of the execution process of (1) to (3) above, please refer to the description of the above embodiments. The implementation principle and technical effect are similar, and will not be repeated here.

[0162] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0163] Based on the same inventive concept, this application also provides an image deduplication apparatus for implementing the image deduplication method described above. The solution provided by this apparatus is similar to the implementation described in the above method; therefore, the specific limitations in one or more image deduplication apparatus embodiments provided below can be found in the limitations of the image deduplication method described above, and will not be repeated here.

[0164] In one embodiment, Figure 8 This is a schematic diagram of the structure of an image deduplication device in one embodiment of this application. The image deduplication device provided in this embodiment can be applied to computer equipment. Figure 8 As shown, the image deduplication device of this application embodiment may include: an image acquisition module 11 and a deduplication processing module 12, wherein:

[0165] Image acquisition module 11 is used to acquire at least two images to be processed introduced in the target development project;

[0166] The deduplication module 12 is used to call the pre-deployed project development editor to perform similarity recognition on each image to be processed, obtain the similarity value between each pair of images to be processed, and perform deduplication on each image to be processed based on the similarity value.

[0167] The project development editor integrates image similarity recognition and image deduplication functions.

[0168] The image deduplication device provided in this application embodiment can be used to execute the technical solutions in the above-described image deduplication method embodiments of this application. Its implementation principle and technical effect are similar, and will not be repeated here.

[0169] In one embodiment, the deduplication module 12 includes: a feature extraction unit and a similarity processing unit, wherein:

[0170] The feature extraction unit is used to extract features from each image to be processed based on a deep learning model, and obtain the feature vector of each image to be processed.

[0171] The similarity processing unit is used to perform similarity processing based on the feature vectors of each image to be processed, and to determine the similarity value between each pair of images to be processed.

[0172] The image deduplication device provided in this application embodiment can be used to execute the technical solutions in the above-described image deduplication method embodiments of this application. Its implementation principle and technical effect are similar, and will not be repeated here.

[0173] In one embodiment, the feature extraction unit is specifically used for:

[0174] Each image to be processed is preprocessed to obtain the corresponding preprocessed image; the preprocessing includes at least one of the following: size adjustment, grayscale conversion, and noise reduction.

[0175] Each preprocessed image is input into a deep learning model to obtain the feature vector of each image to be processed.

[0176] The image deduplication device provided in this application embodiment can be used to execute the technical solutions in the above-described image deduplication method embodiments of this application. Its implementation principle and technical effect are similar, and will not be repeated here.

[0177] In one embodiment, the deduplication module 12 includes: a determination unit, a classification unit, and a deduplication processing unit, wherein:

[0178] The determining unit is used to determine the similarity comparison score between each pair of images to be processed based on each similarity value;

[0179] The classification unit is used to classify at least one target image in each image to be processed based on the similarity comparison scores, and to determine at least one target classification image set.

[0180] The deduplication unit is used to perform deduplication processing on each target image according to each target classification image set, so as to complete the deduplication processing of each image to be processed.

[0181] The image deduplication device provided in this application embodiment can be used to execute the technical solutions in the above-described image deduplication method embodiments of this application. Its implementation principle and technical effect are similar, and will not be repeated here.

[0182] In one embodiment, each image to be processed is a newly introduced image; the determining unit includes a first determining subunit, wherein:

[0183] The first determining subunit is specifically used to: find each similarity value in a pre-constructed mapping table, and determine the score that matches each similarity value in the mapping table as the similarity comparison score between each pair of newly introduced images in each newly introduced image; the mapping table includes the association between different similarity values ​​and different scores.

[0184] The image deduplication device provided in this application embodiment can be used to execute the technical solutions in the above-described image deduplication method embodiments of this application. Its implementation principle and technical effect are similar, and will not be repeated here.

[0185] In one embodiment, each image to be processed includes a newly introduced image and at least a plurality of previously introduced images from a classification image set; the determining unit includes a second determining subunit, wherein:

[0186] The second determined subunit is specifically used for:

[0187] For any newly introduced image, the similarity value between the newly introduced image and each of the previously introduced images is found in the pre-built mapping table, and the score that matches each similarity value in the mapping table is determined as the corresponding similarity score between the newly introduced image and each of the previously introduced images.

[0188] For any set of classified images, the similarity score between the newly introduced image and each previously introduced image in the set of classified images is determined based on the similarity score between the newly introduced image and each previously introduced image in the set of classified images.

[0189] The image deduplication device provided in this application embodiment can be used to execute the technical solutions in the above-described image deduplication method embodiments of this application. Its implementation principle and technical effect are similar, and will not be repeated here.

[0190] In one embodiment, the classification unit is specifically used for:

[0191] Based on the similarity comparison scores, multiple similarity comparison scores are determined for each target image;

[0192] For any target image, obtain the maximum similarity comparison score from the similarity comparison scores corresponding to the target image;

[0193] Based on the maximum similarity comparison score, the target image is assigned to the target classification image set.

[0194] The image deduplication device provided in this application embodiment can be used to execute the technical solutions in the above-described image deduplication method embodiments of this application. Its implementation principle and technical effect are similar, and will not be repeated here.

[0195] In one embodiment, the deduplication unit is specifically used for:

[0196] For any target classification image set, the original positional order of each image is adjusted according to the attribute information of each image in the target classification image set and the requirement information of the target development project to obtain the adjusted image set; each image in the target classification image set includes at least one target image;

[0197] Based on the current position order of each image in the adjusted image set, deduplication is performed on each image in the adjusted image set to complete the deduplication of each target image.

[0198] The image deduplication device provided in this application embodiment can be used to execute the technical solutions in the above-described image deduplication method embodiments of this application. Its implementation principle and technical effect are similar, and will not be repeated here.

[0199] In one embodiment, the image deduplication device further includes a project execution module, wherein:

[0200] The project execution module is used to develop a target project based on multiple valid images obtained after deduplication, provided that each image to be processed is deduplicated.

[0201] The image deduplication device provided in this application embodiment can be used to execute the technical solutions in the above-described image deduplication method embodiments of this application. Its implementation principle and technical effect are similar, and will not be repeated here.

[0202] For specific limitations regarding the image deduplication device, please refer to the limitations of the image deduplication method above, which will not be repeated here. Each module in the aforementioned image deduplication device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0203] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 9 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides processing power. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores various images to be processed introduced into the target development project. The network interface communicates with external endpoints via a network connection. When executed by the processor, the computer program implements an image deduplication method.

[0204] Those skilled in the art will understand that Figure 9 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0205] In one embodiment, a computer device is also provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the technical solution in the above-described image deduplication method embodiment of this application. The implementation principle and technical effect are similar, and will not be repeated here.

[0206] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, it implements the technical solution of the image deduplication method described above in this application. Its implementation principle and technical effect are similar, and will not be repeated here.

[0207] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the technical solution of the image deduplication method described above in this application. The implementation principle and technical effect are similar and will not be repeated here.

[0208] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0209] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0210] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. An image deduplication method, characterized by, The method comprises: acquiring at least two to-be-processed images introduced in a target development project; calling a pre-deployed project development editor to perform similarity recognition on each of the to-be-processed images, to obtain a similarity value between each pair of to-be-processed images in the to-be-processed images, and to perform deduplication processing on each of the to-be-processed images according to each of the similarity values. The project development editor is integrated with an image similarity recognition function and an image deduplication function.

2. The method of claim 1, wherein, The similarity recognition on each of the to-be-processed images to obtain a similarity value between each pair of to-be-processed images in the to-be-processed images comprises: performing feature extraction on each of the to-be-processed images based on a deep learning model to obtain a feature vector of each of the to-be-processed images; performing similarity processing according to the feature vector of each of the to-be-processed images to determine a similarity value between each pair of to-be-processed images in the to-be-processed images.

3. The method of claim 2, wherein, The feature extraction on each of the to-be-processed images based on the deep learning model to obtain a feature vector of each of the to-be-processed images comprises: performing preprocessing on each of the to-be-processed images to obtain a corresponding preprocessed image; the preprocessing comprises at least one of size adjustment processing, grayscale processing, and denoising processing; inputting each of the preprocessed images into the deep learning model to obtain a feature vector of each of the to-be-processed images.

4. The method according to any one of claims 1 to 3, characterized in that, The deduplication processing on each of the to-be-processed images according to each of the similarity values comprises: determining a similarity comparison score between each pair of to-be-processed images in the to-be-processed images according to each of the similarity values; classifying at least one target image in each of the to-be-processed images according to each of the similarity comparison scores to determine at least one target classification image set; performing deduplication processing on each of the target images according to each of the target classification image sets to complete the deduplication processing on each of the to-be-processed images.

5. The method of claim 4, wherein, Each of the to-be-processed images is a newly introduced image. The determination of a similarity comparison score between each pair of to-be-processed images in the to-be-processed images according to each of the similarity values comprises: finding each of the similarity values in a pre-constructed mapping relationship table, and determining a score matched with each of the similarity values in the mapping relationship table as a similarity comparison score between each pair of newly introduced images in each of the newly introduced images; the mapping relationship table comprises an association relationship between different similarity values and different scores.

6. The method of claim 4, wherein, Each of the to-be-processed images comprises one newly introduced image and a plurality of introduced images in at least one classification image set. The determination of a similarity comparison score between each pair of to-be-processed images in the to-be-processed images according to each of the similarity values comprises: for any newly introduced image, finding a similarity value between the newly introduced image and each of the introduced images in a pre-constructed mapping relationship table, and determining a score matched with each of the similarity values in the mapping relationship table as a corresponding similarity score between the newly introduced image and each of the introduced images; determining a similarity comparison score between the newly introduced image and each of the classification image sets according to the similarity score between the newly introduced image and each of the introduced images in each of the classification image sets.

7. The method of claim 4, wherein, The method further comprises: In the case of deduplication of each of the to-be-processed images, performing a target development project according to a plurality of effective images obtained after image deduplication. The device comprises: An image acquisition module is configured to acquire at least two to-be-processed images introduced in a target development project.

8. The method of claim 4, wherein, A deduplication processing module is configured to call a pre-deployed project development editor, perform similarity recognition on each of the to-be-processed images, obtain a similarity value between each pair of to-be-processed images in each of the to-be-processed images, and perform deduplication processing on each of the to-be-processed images according to each of the similarity values. The project development editor is integrated with an image similarity recognition function and an image deduplication function. The processor executes the computer program to implement the steps of the method in any one of claims 1-9.

9. The method of any one of claims 1-3, wherein, The processor executes the computer program to implement the steps of the method in any one of claims 1-9. ​ 10. An image deduplication apparatus, comprising: ​ ​ ​ ​ 11.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-10. ​

Images