Method, system, device and medium for obtaining duplicate content between multiple pictures

Abstract

The present disclosure relates to the field of image processing technology, and in particular to a method, system, device and storage medium for obtaining duplicate content between multiple images. The method constructs a training data set containing multiple training images, and uses the training data set to train a model that can detect and distinguish between marked areas in multiple training images and whether the marked areas are the same, and then uses the trained model to distinguish whether the marked areas detected in the target image are the same, and finally quickly finds the duplicate content between the target images based on the detected marked areas and the distinction results.

Description

Technical Field

[0001] The present disclosure relates to the field of image processing technology, and in particular to a method, system, device and storage medium for acquiring repeated content between multiple pictures. Background Art

[0002] The imaging range of a camera can achieve different effects through different lenses. However, when the viewing area facing the camera is large, or the camera is close to the viewing area, no matter which lens is used, the imaging range of the camera is always difficult to completely cover the viewing area. In order to completely cover the viewing area, it is necessary to deploy multiple groups of cameras or use the same camera to take multiple photos.

[0003] Whether multiple cameras are deployed or the same camera is used to take multiple photos, the final result is that multiple pictures will be generated. For a complete framing area, the complete framing area can only be obtained by splicing the multiple pictures generated. Obviously, when splicing these pictures, it is necessary to consider the overlapping of repeated content between pictures in order to achieve correct splicing of pictures.

[0004] In the retail industry, the method of using cameras to take pictures of goods on containers and using artificial intelligence technology to identify different areas of goods in the image, and finally forming a report on the sales of goods is gradually being applied. Due to the limitations of objective requirements such as clarity, taking multiple container images has become a necessary step. If you want to use these images to stitch together a complete container image, you also need to manually identify the repeated content between multiple images in order to achieve accurate stitching. Summary of the invention

[0005] One aspect of the present disclosure provides a method for obtaining repeated content between multiple pictures. The method comprises the following steps:

[0006] Marking regions of a plurality of training images to obtain a training data set including marked regions encircled by a plurality of marking frames; wherein an association relationship is established between the marked regions, and the association relationship between different marked regions is different from the association relationship between the same marked regions;

[0007] Inputting the training data set into the model for training, so as to obtain a detection classification model capable of detecting the marked areas and determining whether any two marked areas are identical;

[0008] Inputting the target image into the detection classification model to detect the marked areas in the target image and outputting a determination result after determining whether every two marked areas are the same;

[0009] The repeated content in the target image is obtained according to the detected marked area in the target image and the discrimination result.

[0010] In one embodiment, the step of marking regions of a plurality of training images to obtain a training data set including marked regions encircled by a plurality of marked frames; wherein an association relationship is established between the marked regions, and the association relationship between the marked regions of different contents is different from the association relationship between the marked regions of the same content comprises:

[0011] Marking a plurality of marking frames that can cover the training images on a plurality of the training images to form a training set with marking frames; wherein the plurality of marking frames are distributed at intervals on the training images;

[0012] Obtain the marked areas in the training set and establish associations between the marked areas, wherein

[0013] The association relationship between the marked areas of different contents is different from the association relationship between the marked areas of the same content.

[0014] In one embodiment, the step of marking a plurality of marked boxes on a plurality of training pictures that can cover the training pictures to form a training set with marked boxes includes:

[0015] Obtaining a plurality of the training images;

[0016] The training set is obtained by manually marking a plurality of marking frames on the training image at intervals.

[0017] In one embodiment, the step of obtaining the marked regions in the training set and establishing association relationships between the marked regions includes:

[0018] Manually determine the same marked area in the training set;

[0019] Establishing association relationships between marked areas;

[0020] The training data set is formed by using the marked regions and the association relationships in the training set.

[0021] In one embodiment, the step of inputting the training data set into the model for training to obtain a detection classification model capable of detecting the marked areas and determining whether two marked areas are identical comprises:

[0022] Inputting the training set into a model for training to obtain a marked area detection model capable of detecting the marked area;

[0023] The training data set is input into another model for training to obtain a marked region classification model capable of determining whether the marked regions are the same.

[0024] In one embodiment, the step of inputting the target image into the detection classification model to detect the marked areas in the target image and outputting the determination result after determining whether every two marked areas are the same comprises:

[0025] Inputting the target image into the marked area detection model to detect the marked area in the target image;

[0026] The marked areas of the target image are input into the marked area classification model to obtain a determination result of whether or not there is a match between every two marked areas.

[0027] In one embodiment, the training model is a Mask R-CNN model, a Faster R-CNN model, a YOLO model or an SSD model.

[0028] Another aspect of the present disclosure is to provide a system for acquiring repeated content between multiple pictures. The system for acquiring repeated regions between multiple pictures is used to implement the steps of the method for acquiring repeated content between multiple pictures as described above. The system for acquiring repeated content between multiple pictures includes:

[0029] A data acquisition module, used for marking regions of a plurality of training images to obtain a training data set including marked regions encircled by a plurality of marking frames; wherein an association relationship is established between the marked regions, and the association relationship between the marked regions of different contents is different from the association relationship between the marked regions of the same content;

[0030] A model training module, used for inputting the training data set into the model for training, so as to obtain a detection classification model capable of detecting the marked areas and determining whether any two marked areas are identical;

[0031] An image recognition module is used to input the target image into the detection classification model to detect the marked areas in the target image and output a determination result after determining whether every two marked areas are the same; and

[0032] The same marked area in the target image is obtained according to the detected marked area in the target image and the discrimination result.

[0033] Another aspect of the present disclosure further provides a device for obtaining repeated content between multiple pictures. The device for obtaining repeated content between multiple pictures includes:

[0034] Memory for storing computer programs;

[0035] A processor is used to implement the steps of the method for obtaining repeated content between multiple pictures as described in the previous aspect of the present disclosure when executing the computer program.

[0036] The last aspect of the present disclosure provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the method for obtaining repeated content between multiple pictures as described above are implemented.

[0037] In the method, system, device and storage medium for obtaining repeated content between multiple images provided by the present invention, a training data set including multiple training images is constructed, and a model that can detect and distinguish the marked areas in multiple training images and whether the marked areas are the same is trained using the training data set. Then, the trained model is used to distinguish whether the marked areas detected in the target image are the same, and finally, the repeated content between the target images is quickly found based on the detected marked areas and the distinction results.

[0038] On the other hand, the present disclosure can also utilize marking boxes to mark as many repeated contents as possible, thereby ensuring that the largest range of identical contents can be identified.

[0039] On the other hand, the present disclosure also uses manual labeling to produce training data sets, thereby greatly improving the accuracy of training results.

[0040] Finally, the present disclosure also utilizes different training models for training respectively, and then uses two trained models in combination to ultimately achieve the purpose of identifying repeated content in the target image. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings exemplarily illustrate the embodiments and constitute a part of the specification, and together with the text description of the specification, are used to explain the exemplary implementation of the embodiments. The embodiments shown are for illustrative purposes only and do not limit the scope of the claims. In all drawings, the same reference numerals refer to similar but not necessarily identical elements.

[0042] Figure 1 is a flowchart of the steps of a method for obtaining repeated content between multiple pictures shown in an embodiment of the present disclosure;

[0043] Figure 2 It is a flowchart of the steps of a preferred method for obtaining repeated content between multiple pictures provided by an embodiment of the present disclosure;

[0044] Figure 3 is a schematic diagram of direction setting of a marking frame provided by an embodiment of the present disclosure;

[0045] Figure 4 It is a flowchart of the steps of a marking method for a preferred marking frame provided by an embodiment of the present disclosure;

[0046] Figure 5 is a flowchart of the preferred steps for forming a training data set provided by an embodiment of the present disclosure;

[0047] Figure 6 This is a schematic diagram of module connections of a system for acquiring duplicate content between multiple pictures provided by an embodiment of the present disclosure;

[0048] Figure 7 is a structural schematic diagram of a device for acquiring duplicate content between multiple pictures provided by an embodiment of the present disclosure;

[0049] Figure 8 It is a schematic diagram of the structure of a computer-readable storage medium provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0050] From the above description, it can be seen that finding duplicate content in multiple images is a very important step in the field of artificial intelligence. Especially for the retail field, quickly finding duplicate content in multiple images is conducive to using artificial intelligence technology to obtain more information about retail faster.

[0051] In supermarkets, due to the limitation of clarity requirements and the space between shelves, it is almost impossible for a camera to obtain all the product information on a shelf through one picture. However, it is possible to completely cover the entire shelf area by taking multiple pictures of the same shelf. The shelf pictures obtained by taking different pictures at different positions are difficult to meet the prerequisite of the image stitching method of feature point matching, that is, the content of the stitched pictures is the same approximate plane or the picture projection point is in the same position in space. The traditional automatic stitching method is difficult to meet the requirements. However, these pictures can be stitched and aligned manually. In this process, the repeated content between each picture and other pictures is first found out by manual recognition, which is conducive to the subsequent stitching work. In fact, the goods on the shelves are updated every day, and the number of shelves is also large, so the pictures generated by taking pictures of the shelves are very large. Obviously, using manual recognition to find the repeated content between multiple pictures requires huge manpower, material resources and time. How to find the repeated content between multiple pictures in a more efficient way, so as to replace manual recognition of repeated content, will be conducive to the rapid development of the retail industry.

[0052] In order to solve the problems existing in the prior art, improve the speed of obtaining repeated content between multiple pictures, and reduce the labor input cost and work intensity, the inventors have proposed a method, system, device and storage medium for obtaining repeated content between multiple pictures through creative labor. It is worth noting that the method for obtaining repeated content between multiple pictures provided by the present disclosure is to facilitate finding repeated content or repeated areas between multiple pictures. Obviously, the found repeated content can not only be applied to picture splicing in the retail field, but also can be applied to other aspects of work that require obtaining repeated content between multiple pictures.

[0053] The following is a further detailed description of the method, system, device and storage medium for obtaining repeated content between multiple pictures proposed in the present disclosure in combination with the accompanying drawings and specific embodiments. According to the following description, the advantages and features of the present disclosure will become clearer. It should be noted that the accompanying drawings are all in a very simplified form and are not in precise proportions, which are only used to conveniently and clearly assist in explaining the purpose of the embodiments of the present disclosure.

[0054] It should be understood that the terms used in the specification are only used to describe specific embodiments and are not intended to limit the present disclosure. All terms (including technical terms and scientific terms) used in the specification have the meanings commonly understood by those skilled in the art unless otherwise defined. For the sake of brevity and / or clarity, well-known functions or structures are not described in detail.

[0055] An example of how to obtain duplicate content between multiple images

[0056] See also Figure 1 , which shows a flowchart of the steps of a method for obtaining repeated content between multiple pictures in one embodiment of the present disclosure.

[0057] In step S001 of this embodiment, multiple training images are region-marked to obtain a training data set including marked areas encircled by multiple marking frames; wherein an association relationship is established between the marked areas, and the association relationship between the marked areas of different contents is different from the association relationship between the marked areas of the same content.

[0058] The training images are labeled here so that the labeled areas in these images can be recognized by the model to be trained, so as to obtain the corresponding areas on the target image that can identify the characteristics of these labeled areas. The association between the labeled areas is to "tell" the trained model what kind of labeled areas are the same or different.

[0059] In addition, the region marking work involved in step S001 and the association relationship between the marked regions can be completed by manual marking and setting, or other reasonable methods understood by those skilled in the art can be used to complete the work.

[0060] In addition, it is necessary to further explain the "same marked area" or "identical marked area" and "different marked area" or "different marked area" involved in the present disclosure. The same marked area or different marked area involved in the present disclosure refers to the same or different image content in the marked area. In addition, it is easy for those skilled in the art to understand that in the field of computer vision recognition, it can be achieved through some network models to determine whether the content of two pictures or pictures in two areas is the same. However, the "same" does not necessarily mean that the two pictures (content) or two areas are completely consistent, but when the similarity between the picture contents in two pictures or two areas reaches a set threshold according to certain judgment rules, the two pictures or two areas can be considered to be "same". This is common in machine recognition and image labeling applications.

[0061] In step S002, the training data set is input into a training model for training to obtain a detection classification model that can detect the marked areas and can determine whether every two marked areas are the same.

[0062] For step S002, by inputting the training data set into the model for training, the characteristics of the model itself can be used to automatically find the features in the training data set (circled by the marked box), and these features can be classified so that the model can quickly make a judgment the next time it encounters similar features. Of course, the specific process will vary depending on the network model, but the ultimate goal is to train the detection model so that similar features in the target image can be effectively and accurately identified. Of course, in order to detect the accuracy of the trained model and evaluate the recognition effect of the detection model, some detection data sets can also be made, and the trained model can be tested using the detection data sets. When an undesirable state occurs, improvements can be made more targeted and timely.

[0063] In step S003, the target image is input into the detection and classification model to detect the marked areas in the target image, and output a determination result after determining whether every two marked areas are the same.

[0064] In step S004, the repeated content in the target image is obtained according to the detected marked area in the target image and the discrimination result.

[0065] It can be seen that the method for obtaining repeated content between multiple images implemented by the above-mentioned embodiments of the present invention constructs a training data set including multiple training images, and uses the training data set to train a model that can detect and distinguish the marked areas in multiple training images and whether the marked areas are the same. Then, the trained model is used to distinguish whether the detected marked areas in the target image are the same, and finally the repeated content between the target images is quickly found based on the detected marked areas and the distinction results.

[0066] In one embodiment of the present disclosure, a preferred implementation of a method for obtaining repeated content between multiple pictures is also described. Figure 2 In the figure, a flow chart of the steps of the method for obtaining repeated content between multiple pictures provided by the embodiment is shown. The preferred scheme involved in this embodiment can be described by taking a picture of goods placed in a vertical refrigerator as an example. Under existing conditions, it is difficult for a camera installed on the refrigerator door to capture all the goods in the refrigerator within the opening and closing range of the refrigerator door. Therefore, in order to capture all the goods in the refrigerator, two cameras will be installed at the upper and lower positions on the refrigerator door, and all the goods will be captured by the combination of the two cameras. However, the two cameras will inevitably produce two pictures, and due to objective problems such as the installation angle and the shooting sequence, some identical content (repeated content) will appear in the two pictures. If two pictures need to be spliced ​​or other processing operations are required, obtaining (or identifying) the repeated content in the two pictures becomes a prerequisite for subsequent operations. Since the two pictures here are taken by cameras at the upper and lower positions. Therefore, overlapping content will appear in the horizontal direction of the goods (this is because the width of the refrigerator is small and the height is large. Usually, the shooting area of ​​a camera can cover the width range of the refrigerator, but it is difficult to cover the height range of the refrigerator). Similarly, if two pictures are taken in a horizontal order, the repeated content between the two pictures should appear in the vertical direction of the goods. Based on this, the relationship between direction and repeated content can be used to limit the arrangement of the marking boxes in the training pictures to identify as many repeated contents as possible between the two pictures. If other shooting orders appear, the arrangement of the marking boxes can also be adjusted accordingly. In short, the arrangement of the marking boxes should allow the marking boxes to cover as much repeated content as possible.

[0067] In step S011 of this embodiment, a plurality of marking frames that can cover the training images are marked on the plurality of training images to form a training set with marking frames; wherein the plurality of marking frames are spaced apart on the training images. If the images are images taken by the upper and lower cameras on the refrigerator, then according to the usual marking method, when the training images are marked with rectangular marking frames, reference can be made to the following. Figure 3As shown in . The length direction A of the rectangular marking frame is perpendicular to the first direction B.

[0068] In step S012, the marked areas in the training set are obtained, and association relationships are established between the marked areas, wherein the association relationship between the marked areas of different contents is different from the association relationship between the marked areas of the same content.

[0069] Through step S012, an association relationship can be established between the marked areas in the training image, so as to inform the training model that needs to be trained subsequently which marked areas are the same and which marked areas are different.

[0070] In step S013, the training set is input into a model for training to obtain a marked area detection model capable of detecting the marked area. The model used for training here can be a Mask R-CNN model, a Faster R-CNN model, a YOLO model or an SSD model.

[0071] In step S014, the training data set is input into another model for training to obtain a marked region classification model capable of distinguishing whether the marked regions are the same. Of course, the model used for training in this step can be a twin neural network model.

[0072] It can be seen from steps S013 and S014 that using a single detection model can reduce the difficulty of model training and improve the recognition accuracy of the model.

[0073] In step S015, the target image is input into the marked area detection model to detect the marked area in the target image.

[0074] In step S016, the marked areas of the target image are input into the marked area classification model to obtain a determination result of whether every two marked areas are identical.

[0075] Based on the above steps, it can be seen that the repeated content between multiple images can be quickly and accurately obtained after the processes of marking, detection, recognition (determination), etc.

[0076] In one embodiment of the present disclosure, a preferred solution for "marking a marking box on a training image" is further described. Figure 4 In the figure, a flowchart of the steps of the marking method for a preferred marking frame is shown.

[0077] In step S021, a plurality of training images are obtained.

[0078] In step S022, a plurality of marking frames are marked on the training image at intervals by manual marking to obtain the training set.

[0079] The marking box marked by the above method can better cover the repeated content, thereby helping to improve the accuracy of model recognition.

[0080] exist Figure 5 In FIG. 1 , a flowchart of a preferred method for forming a training data set provided by an embodiment of the present disclosure is shown.

[0081] In step S031, identical marked regions in the training set are manually determined.

[0082] In step S032, an association relationship between the marked areas is established.

[0083] In step S033, the training data set is formed using the training set and the association relationship.

[0084] Based on the above-mentioned disclosed content, it can be known that in the method for obtaining repeated content between multiple images provided by the present disclosure, a training data set containing multiple training images is constructed in a preferred manner, and two models that can detect and distinguish the marked areas in multiple training images and whether the marked areas are the same are trained using the training data set. Then, the two trained models are used to detect the marked areas and distinguish whether the detected marked areas in the target images are the same. Finally, the repeated content between the target images is quickly found based on the detected marked areas and the distinction results, which greatly improves the speed and accuracy of obtaining repeated content.

[0085] Example of a system for obtaining duplicate content between multiple images

[0086] In one embodiment of the present disclosure, a system for obtaining duplicate content between multiple pictures is also provided. Figure 6 , a schematic diagram of module connection of a system for obtaining duplicate content between multiple pictures provided by an embodiment of the present disclosure is shown. The system can implement the method for obtaining duplicate content between multiple pictures described in the present disclosure. In order to implement the method for obtaining duplicate content between multiple pictures described in the present disclosure, the system includes:

[0087] The data acquisition module 501 is used to perform region marking on a plurality of training images to obtain a training data set including marked regions encircled by a plurality of marking frames; wherein an association relationship is established between the marked regions, and the association relationship between the marked regions of different contents is different from the association relationship between the marked regions of the same content;

[0088] A model training module 502 is used to input the training data set into the model for training, so as to obtain a detection classification model that can detect the marked areas and can determine whether every two marked areas are the same;

[0089] The image recognition module 503 is used to input the target image into the detection and classification model to output a discrimination result after detecting the marked area in the target image and judging whether each two marked areas are the same; and obtain the same marked area in the target image based on the detected marked area in the target image and the discrimination result.

[0090] Example description of the device for obtaining duplicate content between multiple pictures

[0091] In one embodiment of the present disclosure, a device for obtaining repeated content between multiple pictures is also provided. The device for obtaining repeated content between multiple pictures includes:

[0092] Memory for storing computer programs;

[0093] A processor is used to implement the steps of the method for obtaining repeated content between multiple pictures described in the present disclosure when executing the computer program.

[0094] Various aspects of the present disclosure may be implemented as a system, method or program product. Therefore, various aspects of the present disclosure may be specifically implemented in the following forms, i.e., a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or an implementation combining hardware and software aspects, which may be collectively referred to herein as a "circuit", "module" or "platform".

[0095] Figure 7 1 is a schematic diagram of a device for obtaining duplicate content between multiple pictures provided by an embodiment of the present disclosure. Figure 7 The electronic device 600 implemented according to the implementation in this embodiment is described in detail. Figure 7 The electronic device 600 shown is merely an example and should not bring any limitation to the functions and scope of use of any embodiment of the present disclosure.

[0096] like Figure 7 As shown, the electronic device 600 is in the form of a general computing device. The components of the electronic device 600 may include but are not limited to: at least one processing unit 610, at least one storage unit 620, a bus 630 connecting different platform components (including the storage unit 620 and the processing unit 610), a display unit 640, etc.

[0097] The storage unit stores program codes, which can be executed by the processing unit 610, so that the processing unit 610 executes the implementation steps in this embodiment described in the above-mentioned method for obtaining duplicate content. For example, the processing unit 610 can execute the following steps: Figure 1 , Figure 2 , Figure 4 and Figure 5 Follow the steps shown in .

[0098] The storage unit 620 may include a readable medium in the form of a volatile storage unit, such as a random access unit (RAM) and / or a cache storage unit, and may further include a read-only storage unit (ROM) 6203 .

[0099] The storage unit 620 may also include a program / utility 6204 having a set (at least one) of program modules 6205, such program modules 6205 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0100] Bus 630 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0101] The electronic device 600 may also communicate with one or more external devices 700 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may communicate with one or more devices that enable a user to interact with the electronic device 600, and / or communicate with any device that enables the electronic device to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed through an input / output (I / O) interface 650. Furthermore, the electronic device 600 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through a network adapter 660. The network adapter 660 may communicate with other modules of the electronic device 600 through the bus 630. It should be understood that although Figure 7 Not shown, other hardware and / or software modules may be used in conjunction with electronic device 600, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage platforms.

[0102] Example description of readable storage medium

[0103] One embodiment of the present disclosure further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the method for obtaining duplicate content between multiple pictures in the above disclosure can be implemented. Although this embodiment does not list other specific implementations in detail, in some possible implementations, various aspects of the disclosure can also be implemented in the form of a program product, which includes program code. When the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps described in the method for obtaining duplicate content between multiple pictures in the present disclosure according to the implementation methods in various embodiments of the present disclosure.

[0104] As described above, when the computer program stored in the computer-readable storage medium provided in this embodiment is executed, a training data set is constructed by using the provided method for obtaining duplicate content between multiple images, and then a detection model is trained using these training data sets. Finally, this detection model is used to determine whether there is duplicate content in the target image.

[0105] Figure 8 Schematic diagram of the structure of a computer-readable storage medium provided by an embodiment of the present disclosure. Figure 8 As shown, a program product 800 for implementing the above method in an embodiment of the present disclosure is described, which can adopt a portable compact disk read-only memory (CD-ROM) and include program code, and can be run on a terminal device, such as a personal computer. Of course, the program product generated according to this embodiment is not limited thereto. In the present disclosure, a readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, an apparatus or a device.

[0106] The program product may use any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0107] Computer readable storage media may include data signals propagated in baseband or as part of a carrier wave, wherein readable program codes are carried. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The readable storage medium may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, device, or device. The program codes contained on the readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination thereof.

[0108] Program code for performing the operations of the present disclosure may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as C or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., through the Internet using an Internet service provider).

[0109] In summary, in the method, system, device and storage medium for obtaining repeated content between multiple images provided by the present invention, in the method for obtaining repeated content between multiple images provided by the present invention, a training data set containing multiple training images is constructed in an optimal manner, and two models that can detect and distinguish the marked areas in multiple training images and whether the marked areas are the same are trained using the training data set. Then, the two trained models are used to detect the marked areas and distinguish whether the detected marked areas in the target images are the same. Finally, the repeated content between the target images is quickly found based on the detected marked areas and the distinction results, which greatly improves the speed and accuracy of obtaining repeated content.

[0110] On the other hand, the present disclosure can also utilize marking boxes to mark as many repeated contents as possible, thereby ensuring that the largest range of identical areas can be identified.

[0111] On the other hand, the present disclosure also uses manual labeling to produce training data sets, thereby greatly improving the accuracy of training results.

[0112] Finally, the present disclosure also utilizes different training models for training respectively, and then uses two trained models in combination to ultimately achieve the purpose of identifying repeated content in the target image.

[0113] The above description is only a description of the preferred embodiments of the present disclosure, and is not intended to limit the scope of the present disclosure. Any changes or modifications made by a person of ordinary skill in the art based on the above disclosure shall fall within the scope of protection of the claims.

Claims

1. A method for obtaining repeated content between multiple pictures, characterized in that: The steps include: Marking regions of a plurality of training images to obtain a training data set including marked regions encircled by a plurality of marking frames; wherein an association relationship is established between the marked regions, and the association relationship between different marked regions is different from the association relationship between the same marked regions; Inputting the training data set into the model for training, so as to obtain a detection classification model capable of detecting the marked areas and determining whether two marked areas are identical; Inputting the target image into the detection classification model to detect the marked areas in the target image and outputting a determination result after determining whether every two marked areas are the same; The repeated content in the target image is obtained according to the detected marked area in the target image and the discrimination result.

2. The method for obtaining repeated content between multiple pictures according to claim 1, characterized in that: The multiple training images are region-marked to obtain a training data set including a marked region encircled by a plurality of marking frames; The step of establishing an association relationship between the marked areas, and the association relationship between the marked areas of different contents is different from the association relationship between the marked areas of the same content comprises: Marking a plurality of marking frames that can cover the training images on a plurality of the training images to form a training set with marking frames; wherein the plurality of marking frames are distributed at intervals on the training images; Obtain the marked areas in the training set and establish associations between the marked areas, wherein The association relationship between the marked areas of different contents is different from the association relationship between the marked areas of the same content.

3. The method for obtaining repeated content between multiple pictures as claimed in claim 2, characterized in that: The step of marking a plurality of marked frames that can cover the training images on the plurality of training images to form a training set with marked frames includes: Obtaining a plurality of the training images; The training set is obtained by manually marking a plurality of marking frames on the training image at intervals.

4. The method for obtaining repeated content between multiple pictures as claimed in claim 2, characterized in that: The step of obtaining the marked areas in the training set and establishing association relationships between the marked areas includes: Manually determine the same marked area in the training set; Establishing association relationships between marked areas; The training data set is formed by using the marked regions and the association relationships in the training set.

5. The method for obtaining repeated content between multiple pictures as claimed in claim 2, characterized in that: The step of inputting the training data set into the model for training to obtain a detection classification model capable of detecting the marked areas and determining whether two marked areas are identical comprises: Inputting the training set into a model for training to obtain a marked area detection model capable of detecting the marked area; The training data set is input into another model for training to obtain a marked region classification model capable of determining whether the marked regions are the same.

6. The method for obtaining repeated content between multiple pictures as claimed in claim 5, characterized in that: The step of inputting the target image into the detection classification model to detect the marked areas in the target image and outputting the determination result after determining whether every two marked areas are the same comprises: Inputting the target image into the marked area detection model to detect the marked area in the target image; The marked areas of the target image are input into the marked area classification model to obtain a determination result of whether or not there is a match between every two marked areas.

7. The method for obtaining repeated content between multiple pictures as claimed in claim 2, characterized in that: The training model is a Mask R-CNN model, a Faster R-CNN model, a YOLO model or an SSD model.

8. A system for acquiring repeated content between multiple pictures, used to implement the steps of the method for acquiring repeated areas between multiple pictures as described in any one of claims 1 to 7, characterized in that: The system comprises: A data acquisition module, used for marking regions of a plurality of training images to obtain a training data set including marked regions encircled by a plurality of marking frames; wherein an association relationship is established between the marked regions, and the association relationship between the marked regions of different contents is different from the association relationship between the marked regions of the same content; A model training module, used for inputting the training data set into the model for training, so as to obtain a detection classification model capable of detecting the marked areas and determining whether any two marked areas are identical; An image recognition module is used to input the target image into the detection classification model to detect the marked areas in the target image and output a determination result after determining whether every two marked areas are the same; and The same marked area in the target image is obtained according to the detected marked area in the target image and the discrimination result.

9. A device for acquiring repeated content between multiple pictures, characterized in that: include: Memory for storing computer programs; A processor, configured to implement the steps of the method for obtaining repeated content between multiple pictures as described in 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 method for obtaining repeated content between multiple pictures as described in any one of claims 1 to 7 are implemented.

Images