Image reordering method and device and medium

Abstract

The invention relates to the field of computer data processing, in particular to an image reordering method, which comprises the following steps: acquiring an image to be queried, the image to be queried comprising a trademark, characters and description content; based on the to-be-queried image, retrieving in a database by using a feature model to obtain a retrieval result picture; calculating a brand matching degree between the to-be-queried image and each retrieval result picture in the retrieval result pictures, wherein the brand matching degree comprises a trademark matching degree, a character matching degree and a description content matching degree; and sorting the retrieval result pictures according to the brand matching degree to obtain a sorting result.

Description

Technical Field

[0001] The present application relates to the field of computer data processing, and in particular to an image reordering method, device and medium. Background Art

[0002] Image re-ranking is to use the feature information in the image to re-rank the results returned by the image search engine to obtain search results that are more satisfactory to users. When the target of the search is a product image, it is necessary to pay attention to whether the search result is consistent with the brand of the searched image.

[0003] In related technologies, multi-task learning solutions are usually introduced to achieve brand matching of search results. However, these methods of adding category information are only optimizations at the feature level, which cannot meet the consistency requirements of brands or texts, and are prone to inaccurate sorting results. Summary of the invention

[0004] The purpose of this application is to provide an image reordering method, device, medium and program product, which can improve the accuracy of recall sorting.

[0005] An embodiment of the present application discloses an image reordering method, comprising: obtaining an image to be queried, wherein the image to be queried includes a trademark, text, and description content; based on the image to be queried, searching in a database using a feature model to obtain a retrieval result image; calculating the brand matching degree between the image to be queried and each retrieval result image in the retrieval result images, wherein the brand matching degree includes a trademark matching degree, a text matching degree, and a description content matching degree; and sorting the retrieval result images according to the brand matching degree to obtain a sorting result.

[0006] Optionally, calculating the trademark matching degree includes detecting the trademark contained in the image to be queried based on a trademark detection model.

[0007] Optionally, calculating the text matching degree includes identifying the text in the image to be queried and the text in the search result image; and calculating the similarity between the text in the image to be queried and the text in the search result image.

[0008] Optionally, calculating the description content matching degree includes obtaining the description content in the image to be queried and the description content of the retrieval result image; calculating the loss matching matrix of the description content in the image to be queried and the description content of the retrieval result image; and calculating the minimum loss matching using a bipartite graph matching algorithm.

[0009] Optionally, the brand matching degree between the image to be queried and each search result in the search result picture is calculated, including: when the trademark and the text are matched successfully, the higher matching result between the trademark matching degree and the text matching degree is used as the brand matching degree; when the trademark and the text are matched unsuccessfully, the description content matching degree is used as the brand matching degree.

[0010] Optionally, the brand matching degree between the image to be queried and each of the search result images is calculated, including setting priorities for the trademark matching degree, the text matching degree and the description content matching degree in turn; according to the priority, if one of the trademark matching degree, the text matching degree and the description content matching degree matches successfully, then one of the trademark matching degree, the text matching degree and the description content matching degree is determined as the brand matching degree.

[0011] Optionally, based on the image to be queried, searching in the database using a feature model to obtain the retrieval result picture includes obtaining the brand information of each picture in the database; classifying according to the brand information of each picture, and dividing the pictures of the same brand into the same database sub-library; and searching in the database sub-library based on the image to be queried to obtain the retrieval result picture.

[0012] An embodiment of the present application discloses an electronic device, which includes a processor and a memory storing computer-executable instructions, wherein the processor is configured to execute the instructions to implement the above-mentioned image reordering method.

[0013] An embodiment of the present application discloses a computer-readable storage medium, wherein at least one computer instruction is stored in the computer-readable storage medium. The at least one instruction is loaded and executed by a processor to implement the above-mentioned image reordering method.

[0014] The embodiments of the present application disclose a computer program product, which includes computer instructions. When the computer instructions are executed, the above-mentioned image reordering method is implemented.

[0015] The present application provides an image reordering method, device and medium, which obtains an image to be queried including a trademark, text and description content, and retrieves a search result image from a database according to a feature model. The brand matching degree including the trademark matching degree, text matching degree and description content matching degree between the image to be queried and each search result image in the search result image is calculated, so as to reorder the search result images according to the brand matching degree. The present application combines the multimodal information of trademarks, texts and description content into the recall reordering of image search, which can obtain more accurate recall ranking results. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a schematic diagram of a scene of an image reordering method according to an embodiment of the present application;

[0017] Figure 2 is a flowchart of an image reordering method 200 according to an embodiment of the present application;

[0018] Figure 3 is a flowchart of a method 300 for calculating brand matching according to an embodiment of the present application;

[0019] Figure 4 is a flowchart of a search process 400 according to an embodiment of the present application;

[0020] Figure 5 is a hardware structure block diagram of an electronic device 500 according to an embodiment of the present application. DETAILED DESCRIPTION

[0021] The present application will be further described below in conjunction with specific embodiments and accompanying drawings. It will be appreciated that the specific embodiments described herein are only intended to explain the present application and are not intended to limit the present application. In addition, for ease of description, only some but not all structures or processes related to the present application are shown in the accompanying drawings. It should be noted that in this specification, similar reference numerals and letters represent similar items in the following accompanying drawings.

[0022] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below in conjunction with the accompanying drawings.

[0023] Figure 1 It is a scene diagram of an image reordering method according to an embodiment of the present application.

[0024] like Figure 1 As shown, the application scenario of the present application includes a client 100 and a server 200, and communication between the client 100 and the server 200 is achieved through a wired network or a wireless network connection.

[0025] It is worth noting that those skilled in the art will understand that there can be one or more clients 100, and the number of clients 100 should not be used as a limiting condition of the present application. The number of clients 100 is subject to the actual needs in the application scenario. For example, in an image retrieval system, the data source comes from different clients 100, and at least one client 100 sends data to the image retrieval system. The client 100 can be a computer device such as a smart phone, a tablet computer, a user-generated content computer, a desktop computer, etc., which has computing power, input and output functions, and network connection functions. The server 200 can be a single server device, or it can be a computer device such as a server cluster, a cloud server, etc. that provides game application services to the outside.

[0026] In the image retrieval scenario, a user usually inputs an image on the client 100 to initiate a search, and the image retrieval system returns multiple recall results that are similar in image features. For example, if the image input by the user contains an object, such as clothing, shoes, or bags, the image retrieval system retrieves and returns the most similar image in the database.

[0027] When the search target is product images, in addition to paying attention to whether similar results are ranked in the front position in the search results, we should also pay attention to whether the search results are consistent with the brand to which the searched image belongs. For example, when searching for pictures of brand A, pictures of search results of other brands should not appear in the search results. In related technologies, in order to achieve brand matching of search results, multi-task learning solutions are usually introduced. That is, when training the retrieval model, the category label of each picture is added to the training, so that the image features of the same category are clustered together, and the distance between image features of different categories is pulled apart. These methods of adding category information to learn better features are optimizations at the feature level, but for explicit consistency requirements such as brand and text consistency, feature-based methods cannot guarantee high-accuracy recall results.

[0028] In view of the above problems, an embodiment of the present application provides an image reordering method 200, which can use multi-modal data to ensure the extraction of entity information to obtain accurate recall results. Figure 2 As shown, the method 200 includes the following steps.

[0029] In S210, an image to be queried is obtained, where the image to be queried includes a trademark, text, and description content.

[0030] Obtain the image to be queried provided by the user. The image to be queried is, for example, a published note, including a trademark logo of an object, text in the image and a corresponding object description. The object in the image may include one or more commodities, such as one or more of clothing, shoes, bags, and jewelry.

[0031] In S220, based on the image to be queried, the feature model is used to search in the database to obtain a search result image.

[0032] Input the image to be queried into the feature model, and retrieve the result images similar to the image to be queried from the database. The feature model can be any one of VGG, ResNet, MobileNet, SqueezeNet.

[0033] In S230, the brand matching degree between the image to be queried and each of the search result images is calculated, wherein the brand matching degree includes a trademark matching degree, a text matching degree and a description content matching degree.

[0034] For all the results of the query image recall, the brand matching degree between the query image and each result is calculated. For the trademark, text and description content contained in the query image, the trademark matching degree, text matching degree and description content matching degree between the query image and each search result are calculated respectively.

[0035] In some embodiments, calculating the trademark matching degree includes detecting the trademark contained in the image to be queried based on a trademark detection model. The trademark detection model may be a trained logo detection model, which is used to detect whether the image to be queried and the retrieval result image contain a logo of a certain brand. In some examples, the trademark detection model may be a model obtained based on a convolutional neural network (CNN). Typically, these models are two-stage, and candidate regions need to be generated first, and then classification and regression are performed on the candidate regions. In some other examples, the trademark detection model may be a YOLO or SSD model. These models are single-stage, and only one CNN network is used to directly predict the categories and positions of different targets.

[0036] In some embodiments, calculating the text matching degree includes identifying the text in the image to be queried and the text in the search result image; calculating the similarity between the text in the image to be queried and the text in the search result image. An OCR model can be used to identify all text information in the image. After obtaining the text information in the image, the similarity between the text in the image to be queried and the text in each search result image can be calculated. In some examples, the similarity between the texts can be obtained based on any one of cosine similarity, Levenshtein distance, Levenshtein ratio, Hamming distance, Jaro distance, and Jaro-Winkler distance. For example, Levenshtein distance refers to the minimum number of editing operations required to convert one string into another between two strings. The smaller the distance, the greater the similarity between the two strings. For another example, the Hamming distance between two equal-length strings is the number of different characters in the corresponding positions of the two strings. The smaller the distance, the greater the similarity between the two strings. For another example, the Levenshtein ratio is a method of obtaining the similarity between strings by calculating the edit distance between two strings, and the higher the Levenshtein ratio means the more similar the two strings are.

[0037] In some embodiments, calculating the description content matching degree includes obtaining the description content in the image to be queried and the description content of the retrieval result image; calculating the loss matching matrix of the description content in the image to be queried and the description content of the retrieval result image; and calculating the minimum loss matching using a bipartite graph matching algorithm. The NER (Named Entity Recognition) content in the image to be queried and the retrieval result image can be extracted. The NER results of the two are lists NER_query and NER_doc, respectively, and their original similarity is s. The matching loss matrix is ​​calculated for the NER_query list and the NER_doc list respectively. Among them, the loss between each string can also use the similarity calculation method between the characters mentioned above. After obtaining the loss matching matrix, the minimum loss matching can be calculated using a bipartite graph matching algorithm to obtain the matching score between the NER in the queried image and the NER in the retrieval result image.

[0038] In S240, the search result images are sorted according to the brand matching degree to obtain a sorting result.

[0039] All search result images are reordered according to the brand matching degree in S230 to obtain a reordered result.

[0040] In method 200, an image to be queried including a trademark, text and description content is obtained, and a search result image is retrieved from a database according to a feature model. The brand matching degree including the trademark matching degree, text matching degree and description content matching degree between the image to be queried and each search result image in the search result image is calculated, so as to reorder the search result images according to the brand matching degree. Method 200 combines the multimodal information of trademark, text and description content into the recall reordering of image search, and can obtain more accurate recall ranking results.

[0041] In some embodiments, calculating the brand matching degree between the image to be queried and each search result image includes, when the trademark and the text match successfully, taking the higher matching result between the trademark matching degree and the text matching degree as the brand matching degree. Otherwise, when the trademark and the text match unsuccessfully, the description content matching degree is taken as the brand matching degree. That is, if the logo or the OCR-recognized content between the image to be queried and each search result image matches successfully, the result with the greater similarity between the two is taken as the fused matching result. If neither the logo nor the OCR-recognized content matches successfully, the NER matching result is directly taken as the fused matching result. If none of the logo, the OCR-recognized content, and the NER match, the original recall order is not changed.

[0042] In some embodiments, calculating the brand match between the image to be queried and each search result image includes setting priorities for trademark matching, text matching, and description content matching in turn; according to the priority, if one of the trademark matching, text matching, and description content matching is successfully matched, then one of the trademark matching, text matching, and description content matching is determined as the brand matching.

[0043] For example, Figure 3 The flowchart of a method 300 for calculating brand matching of an embodiment of the present application is shown. In method 300, priorities are set according to text matching, trademark matching, and description content matching. In S310, first determine whether the text matching is successful, that is, determine whether the OCR text matching in the query image and the search result image is successful. If so, the OCR text matching is used as the final brand matching (S320); otherwise, enter S330. In S330, determine whether the logo matching in the query image and the search result image is successful. If so, the logo matching is used as the final brand matching (S320); otherwise, enter S340. In S340, determine whether the NER matching in the query image and the search result image is successful. If so, the NER matching is used as the final brand matching (S320); otherwise, no processing is performed.

[0044] In image retrieval, a single base database is usually used, which stores the visual features of images or picture frames. After extracting the visual features of the image requested by the user, an ANN (Approximate Nearest Neighbor) search is performed in the single base database to obtain several images or picture frames that are visually closest to the requested image in the base database, and the search results are returned. However, this method is not suitable for the situation where there are similar styles of objects in different brands. For visually similar objects, there may be multiple brands corresponding to them. Using a single base database and relying only on visual similarity to recall solutions cannot meet the user's needs for searching for the same brand. Figure 4 FIG. 4 is a flowchart of a search process 400 according to an embodiment of the present application.

[0045] In S410, brand information of each image in the database is obtained.

[0046] The brand information corresponding to each picture frame object in the database can be extracted through OCR, NER and logo recognition technology.

[0047] In some implementations, priorities may be set for trademark matching, text matching, and description content matching in turn; according to the priorities, if one of the trademark matching, text matching, and description content matching is successfully matched, the brand information corresponding to each image is determined based on one of the trademark matching, text matching, and description content matching.

[0048] In S420, classification is performed according to the brand information of each image, and images of the same brand are divided into the same database sub-library.

[0049] After obtaining the brand information of the picture frames in the database, the database is divided by brand according to the brand information of each picture frame, that is, objects of each brand are divided into a sub-library.

[0050] In S430, based on the image to be queried, a search is performed in the database sub-library to obtain a search result image.

[0051] During the recall process, according to the brand information of the object in the picture frame, the sub-library corresponding to the brand is searched to obtain the search result image. This ensures that the brand information of the object in the search result is consistent with the image initiated by the user.

[0052] Reference now Figure 5, which is a block diagram of an electronic device 500 according to an embodiment of the present application. The electronic device 500 may include one or more processors 502, a system control logic 508 connected to at least one of the processors 502, a system memory 504 connected to the system control logic 508, a non-volatile memory (NVM) 506 connected to the system control logic 508, and a network interface 510 connected to the system control logic 508.

[0053] The processor 502 may include one or more single-core or multi-core processors. The processor 502 may include any combination of general-purpose processors and special-purpose processors (e.g., graphics processors, application processors, baseband processors, etc.). In the embodiments of the present invention, the processor 502 may be configured to execute the following operations: Figure 2-4 One or more of the various embodiments shown.

[0054] In some embodiments, system control logic 508 may include any suitable interface controller to provide any suitable interface to at least one of processors 502 and / or any suitable device or component in communication with system control logic 508 .

[0055] In some embodiments, the system control logic 508 may include one or more memory controllers to provide an interface to the system memory 504. The system memory 504 may be used to load and store data and / or instructions. In some embodiments, the memory 504 of the device 500 may include any suitable volatile memory, such as a suitable dynamic random access memory (DRAM).

[0056] NVM / memory 506 may include one or more tangible, non-transitory computer-readable media for storing data and / or instructions. In some embodiments, NVM / memory 506 may include any suitable non-volatile memory such as flash memory and / or any suitable non-volatile storage device, such as at least one of a HDD (Hard Disk Drive), a CD (Compact Disc) drive, and a DVD (Digital Versatile Disc) drive.

[0057] NVM / storage 506 may include a portion of storage resources installed on a device of apparatus 500, or it may be accessible to the apparatus but not necessarily a part of the apparatus. For example, NVM / storage 506 may be accessed over a network via network interface 510.

[0058] In particular, system memory 504 and NVM / storage 506 may include, respectively, a temporary copy and a permanent copy of instructions 520. Instructions 520 may include instructions that, when executed by at least one of processors 502, cause device 500 to perform the following operations: Figure 2-4 In some embodiments, instructions 520 , hardware, firmware, and / or software components thereof may additionally / alternatively be placed in system control logic 508 , network interface 510 , and / or processor 502 .

[0059] The network interface 510 may include a transceiver for providing a radio interface for the device 500, thereby communicating with any other suitable device (such as a front-end module, an antenna, etc.) through one or more networks. In some embodiments, the network interface 510 may be integrated with other components of the device 500. For example, the network interface 510 may be integrated with at least one of the processor 502, the system memory 504, the NVM / storage 506, and a firmware device (not shown) having instructions. When at least one of the processors 502 executes the instructions, the device 500 implements Figure 2-4 One or more of the various embodiments shown.

[0060] The network interface 510 may further include any suitable hardware and / or firmware to provide a multiple-input multiple-output radio interface. For example, the network interface 510 may be a network adapter, a wireless network adapter, a telephone modem and / or a wireless modem.

[0061] In one embodiment, at least one of the processors 502 may be packaged together with logic for one or more controllers of the system control logic 508 to form a system in package (SiP). In one embodiment, at least one of the processors 502 may be integrated on the same die with logic for one or more controllers of the system control logic 508 to form a system on a chip (SoC).

[0062] The device 500 may further include an input / output (I / O) device 512. The I / O device 512 may include a user interface, which enables a user to interact with the device 500; and a peripheral component interface design enables peripheral components to interact with the device 500.

[0063] In some embodiments, the user interface may include, but is not limited to, a display (e.g., an LCD display, a touch screen display, etc.), a speaker, a microphone, one or more cameras (e.g., a still image camera and / or a video camera), a flashlight (e.g., an LED flash), and a keyboard.

[0064] In some embodiments, the peripheral component interface may include, but is not limited to, a non-volatile memory port, an audio jack, and a power interface.

[0065] Each method implementation mode of the present application can be implemented in the form of software, magnetic components, firmware, etc.

[0066] Program code can be applied to input instructions to perform the functions described herein and generate output information. The output information can be applied to one or more output devices in a known manner. For purposes of this application, a processing system includes any system having a processor such as, for example, a digital signal processor (DSP), a microcontroller, an application specific integrated circuit (ASIC), or a microprocessor.

[0067] Program code can be implemented with high-level programming language or object-oriented programming language to communicate with the processing system. When necessary, program code can also be implemented with assembly language or machine language. In fact, the mechanism described herein is not limited to the scope of any specific programming language. In either case, the language can be a compiled language or an interpreted language.

[0068] One or more aspects of at least one embodiment may be implemented by representative instructions stored on a computer-readable storage medium, which represent various logic in a processor, and when the instructions are read by a machine, cause the machine to fabricate logic for performing the techniques described herein. These representations, known as "IP cores," may be stored on a tangible computer-readable storage medium and supplied to multiple customers or manufacturing facilities to load into a manufacturing machine that actually manufactures the logic or processor.

[0069] According to one embodiment of the present application, a computer-readable storage medium is further proposed, wherein at least one computer instruction is stored in the computer-readable storage medium, and the at least one instruction is loaded and executed by a processor to implement the aforementioned image reordering method.

[0070] According to an embodiment of the present application, a computer program product is also proposed. The computer program product includes computer instructions. When the computer instructions are executed, the aforementioned image reordering method is implemented.

[0071] Illustrative embodiments of the present application include, but are not limited to, an image reordering method, apparatus, medium, and program product.

[0072] The various aspects of the illustrative embodiments will be described using terms commonly used by those skilled in the art to convey the essence of their work to other persons skilled in the art. However, it is apparent to those skilled in the art that some alternative embodiments may be implemented using the features described in part. For purposes of explanation, specific numbers and configurations are set forth to provide a more thorough understanding of the illustrative embodiments. However, it is apparent to those skilled in the art that alternative embodiments may be implemented without the specific details. In some other cases, some well-known features are omitted or simplified herein to avoid obscuring the illustrative embodiments of the present application.

[0073] In addition, various operations will be described as multiple separate operations in a manner that is most helpful for understanding the illustrative embodiments; however, the order of description should not be interpreted as implying that these operations must rely on the order of description, and many of the operations can be implemented in parallel, concurrently, or simultaneously. In addition, the order of the operations can also be rearranged. When the described operations are completed, the process can be terminated, but can also have additional steps not included in the drawings. The process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0074] References in the specification to "an example", "an example", "an embodiment", "an implementation", etc. indicate that the described embodiment may include a particular feature, structure, or property, but each embodiment may or may not necessarily include the particular feature, structure, or property. Moreover, these phrases are not necessarily directed to the same embodiment. In addition, when particular features are described in conjunction with a specific embodiment, the knowledge of those skilled in the art can affect the combination of these features with other embodiments, whether or not these embodiments are explicitly described.

[0075] Unless the context dictates otherwise, the terms "comprising," "having," and "including" are synonymous. The phrase "A and / or B" means "(A), (B), or (A and B)."

[0076] As used herein, the term "module" may refer to, be part of, or include: memory (shared, dedicated, or group) for running one or more software or firmware programs, application-specific integrated circuits (ASICs), electronic circuits and / or processors (shared, dedicated, or group), combinational logic circuits, and / or other suitable components that provide the described functionality.

[0077] In the accompanying drawings, some structural or method features may be shown in a specific arrangement and / or order. However, it should be understood that such a specific arrangement and / or order is not necessary. Instead, in some embodiments, these features may be described in a manner and / or order different from that shown in the illustrative drawings. In addition, the structural or method features included in a specific figure do not mean that all embodiments need to include such features. In some embodiments, these features may not be included or may be combined with other features.

[0078] In the description provided herein, a large number of specific details are described. However, it is understood that the embodiments of the present application can be practiced without these specific details. In some instances, well-known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this description.

[0079] Similarly, it should be understood that in order to streamline the present application and aid in understanding one or more of the various inventive aspects, in the above description of the exemplary embodiments of the present application, the various features of the present application are sometimes grouped together into a single embodiment, figure, or description thereof. However, this disclosed method should not be interpreted as reflecting the following intention: the claimed application requires more features than the features explicitly recited in each claim. More specifically, as reflected in the claims, the inventive aspects lie in less than all the features of the single embodiment disclosed above. Therefore, the claims that follow the specific embodiment are hereby expressly incorporated into the specific embodiment, with each claim itself serving as a separate embodiment of the present application.

[0080] Those skilled in the art will appreciate that the modules in the devices in the embodiments may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments may be combined into one module or unit or component, and in addition they may be divided into a plurality of submodules or subunits or subcomponents. Except that at least some of such features and / or processes or units are mutually exclusive, all features disclosed in this specification (including the accompanying claims, abstracts and drawings) and all processes or units of any method or device disclosed in this manner may be combined in any combination. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstracts and drawings) may be replaced by an alternative feature providing the same, equivalent or similar purpose.

[0081] In addition, those skilled in the art will appreciate that, although some embodiments described herein include certain features included in other embodiments but not other features, the combination of features of different embodiments is meant to be within the scope of the present application and form different embodiments. For example, in the claims, any one of the claimed embodiments can be used in any combination.

Claims

1. An image reordering method, characterized in that: include, Acquire an image to be queried, wherein the image to be queried includes a trademark, text, and description content; Based on the image to be queried, searching in a database using a feature model to obtain a search result image; Calculating the brand matching degree between the image to be queried and each of the search result images, wherein the brand matching degree includes a trademark matching degree, a text matching degree and a description content matching degree; The search result images are sorted according to the brand matching degree to obtain a sorting result.

2. The image reordering method according to claim 1, characterized in that: Calculating the trademark matching degree includes detecting the trademark contained in the image to be queried based on a trademark detection model.

3. The image reordering method according to claim 1, characterized in that: Calculating the text matching degree includes identifying the text in the image to be queried and the text in the search result image; The similarity between the text in the image to be queried and the text in the search result image is calculated.

4. The image reordering method according to claim 1, characterized in that: Calculating the description content matching degree includes obtaining the description content in the image to be queried and the description content of the search result image; Calculating a loss matching matrix between the description content in the image to be queried and the description content in the retrieval result image; The minimum loss matching is calculated using a bipartite graph matching algorithm.

5. The image reordering method according to any one of claims 2 to 4, characterized in that: Calculating the brand matching degree between the image to be queried and each search result in the search result picture, including: When the trademark and the text are successfully matched, the higher matching result between the trademark matching degree and the text matching degree is used as the brand matching degree; When the trademark and the text are not matched successfully, the description content matching degree is used as the brand matching degree.

6. The image reordering method according to any one of claims 2 to 4, characterized in that: Calculating the brand matching degree between the image to be queried and each of the search result images, including: Prioritize the trademark matching degree, the text matching degree, and the description content matching degree in order; According to the priority, if one of the trademark matching degree, the text matching degree and the description content matching degree matches successfully, then one of the trademark matching degree, the text matching degree and the description content matching degree is determined as the brand matching degree.

7. The image reordering method according to claim 1, characterized in that: Based on the image to be queried, the feature model is used to search in the database, and the search result images obtained include: Obtaining brand information of each image in the database; Classifying the pictures according to the brand information of each picture, and dividing the pictures of the same brand into the same database sub-library; as well as Based on the image to be queried, a search is performed in the database sub-library to obtain the search result image.

8. An electronic device, characterized in that: The device comprises a memory storing computer executable instructions and a processor, wherein the processor is configured to execute the instructions to implement the method according to any one of claims 1-7.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores at least one computer instruction, and the at least one instruction is loaded and executed by a processor to implement the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that The computer program product comprises computer instructions, and when the computer instructions are executed, the method according to any one of claims 1 to 7 is implemented.

Images