Object recognition method and device, equipment, storage medium and computer program product

Abstract

The embodiment of the invention provides an object recognition method and device, equipment, a storage medium and a computer program product, and can be applied to the fields or scenes of cloud technology, artificial intelligence and the like, the method comprises the steps that to-be-processed information of a to-be-recognized object is acquired, and the to-be-processed information comprises program name information and interface text information of an application program, the application program and the to-be-recognized object have an association relationship, and the interface text information comprises text content in an application program interface of the application program; the to-be-recognized object is classified and recognized according to the to-be-processed information, a classification recognition result is obtained, and the classification recognition result comprises a first probability that the to-be-recognized object belongs to a target category and a second probability that the to-be-recognized object belongs to a reference category; and determining an object recognition result of the to-be-recognized object according to the first probability and the second probability, wherein the object recognition result is used for indicating whether the to-be-recognized object belongs to the target category or not. According to the embodiment of the invention, the object recognition efficiency can be effectively improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to an object recognition method, an object recognition device, a computer device, a computer-readable storage medium, and a computer program product. Background Art

[0002] In daily life, object recognition is a very common and important operation. Object recognition includes classification recognition, that is, identifying the classification category to which the object belongs, for example: identifying whether the object belongs to the abnormal category or the normal category. At present, it is usually done by obtaining the feature data of the object, and then the relevant staff manually determines the classification category to which the object belongs based on the acquired feature data. However, the manual judgment method not only consumes manpower costs, but also has the problem of low efficiency. Therefore, how to improve the efficiency of object recognition is a problem that needs to be solved urgently. Summary of the invention

[0003] The embodiments of the present application provide an object recognition method, apparatus, device, storage medium and computer program product, which can realize automation and intelligence of object recognition and effectively improve the efficiency of object recognition.

[0004] In one aspect, an embodiment of the present application provides an object recognition method, the method comprising:

[0005] Acquire information to be processed of the object to be identified, the information to be processed includes program name information and interface text information of an application, the application has an association relationship with the object to be identified, and the interface text information includes text content in an application interface of the application;

[0006] Performing classification and recognition processing on the object to be recognized according to the information to be processed to obtain a classification and recognition result, wherein the classification and recognition result includes a first probability that the object to be recognized belongs to a target category and a second probability that the object to be recognized belongs to a reference category;

[0007] An object recognition result of the object to be recognized is determined according to the first probability and the second probability, and the object recognition result is used to indicate whether the object to be recognized belongs to the target category.

[0008] In one aspect, an embodiment of the present application provides an object recognition device, the device comprising:

[0009] an acquisition unit, configured to acquire information to be processed of the object to be identified, wherein the information to be processed includes program name information and interface text information of an application program, wherein the application program has an association relationship with the object to be identified, and the interface text information includes text content in an application program interface of the application program;

[0010] a processing unit, configured to perform classification and recognition processing on the object to be recognized according to the information to be processed, and obtain a classification and recognition result, wherein the classification and recognition result includes a first probability that the object to be recognized belongs to a target category and a second probability that the object to be recognized belongs to a reference category;

[0011] The processing unit is used to determine an object recognition result of the object to be recognized according to the first probability and the second probability, and the object recognition result is used to indicate whether the object to be recognized belongs to the target category.

[0012] On the one hand, an embodiment of the present application provides a computer device, comprising: a processor, a communication interface and a memory, wherein the processor, the communication interface and the memory are interconnected, wherein the memory stores an executable program code, and the processor is used to call the executable program code to implement the object recognition method provided in the embodiment of the present application.

[0013] Accordingly, an embodiment of the present application further provides a computer-readable storage medium, in which instructions are stored. When the computer-readable storage medium is executed on a computer, the computer implements the object recognition method provided by the embodiment of the present application.

[0014] Accordingly, an embodiment of the present application further provides a computer program product, which includes a computer program or computer instructions, and when the computer program or computer instructions are executed by a processor, the object recognition method provided by the embodiment of the present application is implemented.

[0015] In an embodiment of the present application, information to be processed of an object to be identified is obtained, the information to be processed includes program name information and interface text information of an application, the application has an association relationship with the object to be identified, and the interface text information includes text content in the application interface of the application; the object to be identified is classified and identified according to the information to be processed to obtain a classification and identification result, the classification and identification result includes a first probability that the object to be identified belongs to a target category and a second probability that the object to be identified belongs to a reference category; the object recognition result of the object to be identified is determined according to the first probability and the second probability, and the object recognition result is used to indicate whether the object to be identified belongs to the target category. By adopting the above method, the automation and intelligence of object recognition can be realized, and the efficiency of object recognition can be effectively improved; in addition, by first determining the probability that the object to be identified belongs to the target category and the reference category respectively, and then finally determining whether the object to be identified belongs to the target category according to the probability that the object to be identified belongs to the target category and the reference category respectively, the method based on multi-stage recognition can effectively improve the accuracy of object recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the background technology, the drawings required for use in the embodiments of the present invention or the background technology will be described below.

[0017] Figure 1 is a schematic diagram of the architecture of an object recognition system provided by an exemplary embodiment of the present application;

[0018] Figure 2 is a flowchart of an object recognition method provided by an exemplary embodiment of the present application;

[0019] Figure 3 is a schematic diagram of an application program interface image provided by an exemplary embodiment of the present application;

[0020] Figure 4 is a schematic diagram of a flow chart of object recognition using a model provided by an exemplary embodiment of the present application;

[0021] Figure 5 is a schematic diagram of a network structure of a classification model provided by an exemplary embodiment of the present application;

[0022] Figure 6 is a schematic diagram of obtaining an object recognition result provided by an exemplary embodiment of the present application;

[0023] Figure 7 It is a flowchart of a model training method provided by an exemplary embodiment of the present application;

[0024] Figure 8 is a schematic block diagram of an object recognition device provided by an exemplary embodiment of the present application;

[0025] Fig. 9 It is a schematic block diagram of a computer device provided by an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0027] The embodiment of the present application provides an object recognition method to realize the automation and intelligence of object recognition, and effectively improve the efficiency and accuracy of object recognition. The object recognition method provided in the embodiment of the present application can be implemented based on AI (Artificial Intelligence) technology. AI refers to the theory, method, technology and application system of using digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. AI technology is an interdisciplinary subject, and the fields it involves are relatively wide; and the object recognition method provided in the embodiment of the present application mainly involves machine learning (ML) technology in AI technology. Machine learning generally includes artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and teaching learning.

[0028] In a feasible embodiment, the object recognition method provided in the embodiment of the present application can also be implemented based on cloud technology and / or blockchain technology. Specifically, it may involve one or more of cloud storage, cloud database, and big data in cloud technology. For example, the data generated by executing the object recognition method (such as object recognition results, etc.) is stored in a cloud database. For another example, the data generated by executing the object recognition method (such as object recognition results, classification recognition results, etc.) is stored in the form of blocks on the blockchain; in addition, the object recognition device that executes the object recognition method can be a node device in the blockchain network.

[0029] The object recognition method provided in the embodiment of the present application can be applied to Figure 1The object recognition system shown in the figure includes: one or more clients 11 and an object recognition device 12. The client 11 is a terminal device configured with an application, and the object recognition device 12 is a device for performing object recognition processing. In one embodiment, when there is only one client, the object recognition device 12 and the client 11 may belong to the same device, such as the object recognition device 12 is configured on the client 11. In another embodiment, the object recognition device 12 and the client 11 may be independent devices, the object recognition device 12 may be a server, and the client 11 is connected to the object recognition device through a network. In a feasible implementation, the server may be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The terminal device may be a smart phone, a tablet computer, a laptop computer, a desktop computer, an intelligent voice interaction device, a smart home appliance, a car terminal, etc., but is not limited thereto. The object recognition method provided in the embodiment of the present application includes but is not limited to the following steps:

[0030] The object recognition device 12 obtains the information to be processed of the object to be recognized, and there are one or more applications associated with the object to be recognized. The information to be processed includes the program name information and interface text information of each application in the one or more applications, and the interface text information includes the text content of each application interface in the one or more application interfaces of the application. In a feasible implementation, the application can be set on the client 11, and the program name information and interface text information of the application can be sent by the client 11 to the object recognition device. The client 11 obtains the application interface image of the application, obtains the interface text information from the application interface image, and sends the interface text information to the object recognition device 12. Alternatively, the client 11 obtains the application interface image of the application, sends the application interface image and the program name information of the application to the object recognition device 12, and the object recognition device 12 obtains the interface text information from the application interface image.

[0031] The object recognition device 12 performs classification and recognition processing on the object to be recognized based on the information to be processed to obtain a classification and recognition result, which includes a first probability that the object to be recognized belongs to the target category and a second probability that the object to be recognized belongs to each reference category in one or more reference categories. In a feasible implementation, the target category may be an abnormal category, and the reference category may be a normal category. The object recognition device 12 determines an object recognition result of the object to be recognized based on the first probability and the second probability, and the object recognition result is used to indicate whether the object to be recognized belongs to the target category.

[0032] The object recognition method provided in the embodiment of the present application can realize the automation and intelligence of object recognition, and effectively improve the efficiency of object recognition; in addition, by first determining the probability that the object to be recognized belongs to the target category and the reference category respectively, and then finally determining whether the object to be recognized belongs to the target category according to the probability that the object to be recognized belongs to the target category and the reference category respectively, this multi-stage recognition-based method can effectively improve the accuracy of object recognition.

[0033] It should be noted that in the specific implementation of this application, related data such as the program name information, application interface images, interface text information, etc. of the application are involved. When the embodiments of this application apply the above data to specific products or technologies, user permission or consent is required, and the collection, use and processing of relevant data need to comply with relevant laws, regulations and standards of the relevant regions.

[0034] The above briefly introduces the object recognition method provided in the embodiment of the present application. The following is a detailed description of the specific implementation of the object recognition method.

[0035] See also Figure 2 , Figure 2 A flowchart of an object recognition method provided in an embodiment of the present application. The object recognition method described in the embodiment of the present application can be applied to an object recognition device in the above-mentioned object recognition system, and the method includes but is not limited to the following steps:

[0036] S201, obtaining information to be processed of an object to be identified.

[0037] In an embodiment of the present application, the information to be processed includes the program name information and interface text information of the application of the object to be identified. The interface text information includes the text content in the application interface of the application. The application can be a small program or an App (Application), etc. The application has an association relationship with the object to be identified. For example, the object to be identified can be the owner of the application. The application that has an association relationship with the object to be identified can be one or more. When there are multiple applications that have an association relationship with the object to be identified, the information to be processed includes the program name information and interface text information of each application of the object to be identified. For example, the application that has an association relationship with the object to be identified includes a first application and a second application, then the processing information includes the program name information and interface text information of the first application, and includes the program name information and interface text information of the second application.

[0038] In one embodiment, the application is running on the client, and the application interface image can be obtained by taking a screenshot of the application interface of the object to be identified displayed on the client. The application interface image can be one or more. For example, screenshots are taken of the interfaces of different menus of the application to obtain multiple application interface images. The client obtains the program name information of the application, and the program name information of the application can also be one or more. In one implementation, the client can perform text recognition on the application interface image, obtain the text content contained in the application interface image, and determine the interface text information of the application based on the text content. Please refer to Figure 3 , Figure 3 A schematic diagram of an application program interface image provided in an embodiment of the present application, for example Figure 3 The interface text information obtained by the program interface image is shown as "XX applet food catering leisure entertainment return food catering dessert beverage supermarket convenience meal preferred supper home cooking breakfast snack". The client sends the obtained interface text information and program name information of the application to the object recognition device. Correspondingly, the object recognition device receives the interface text information and program name information of the application sent by the client.

[0039] In another embodiment, the client may send the acquired program name information and application interface image of the application to an object recognition device, and the object recognition device may obtain interface text information from the application interface image. The device may perform text recognition on the application interface image, obtain text content contained in the application interface image, and determine the interface text information of the application based on the text content.

[0040] The interface text information of the application includes text content obtained from each application interface image of one or more application interface images of the application.

[0041] When there are multiple applications associated with the object to be identified, the multiple applications may be running on multiple clients, and the object identification device may receive program name information and interface text information of the applications sent by each client, or receive program name information and application interface images of the applications sent by each client.

[0042] In a feasible implementation, OCR technology may be used to perform text recognition on the application program interface image. OCR is the abbreviation of Optical Character Recognition, which refers to optical character recognition.

[0043] S202 , classify and identify the object to be identified according to the information to be processed to obtain a classification and identification result, wherein the classification and identification result includes a first probability that the object to be identified belongs to a target category and a second probability that the object to be identified belongs to a reference category.

[0044] In an embodiment of the present application, when there are one or more applications associated with the object to be identified, the object identification device respectively obtains first representation information of the program name information of each of the one or more applications, and respectively obtains second representation information of the interface text information of each of the one or more applications, the first representation information and the second representation information are respectively used to represent the semantics of the program name information and the interface text information, and the first representation information and the second representation information can be in the form of vectors; first fused representation information is determined based on the first representation information, and second fused representation information is determined based on the second representation information; the first fused representation information and the second fused representation information are spliced ​​to obtain spliced ​​representation information; and the object to be identified is classified and identified based on the spliced ​​representation information to obtain a classification recognition result.

[0045] In one embodiment, the object recognition device determines the first fused representation information based on the first representation information in the following manner: the object recognition device obtains resource feature information of the target application, which may be a resource transfer amount, performs weighted processing on the first representation information of the program name information of the target application using the resource feature information of the target application, obtains weighted representation information of the program name information of the target application, and determines the first fused representation information based on the weighted representation information of the program name information of each of the one or more applications; the target application is any one of the one or more applications, for example, the applications associated with the object to be recognized are the first application and the second application, then the object recognition device performs weighted processing on the first representation information of the program name information of the first application using the resource feature information of the first application, obtains weighted representation information of the program name information of the first application, the object recognition device performs weighted processing on the first representation information of the program name information of the second application using the resource feature information of the second application, obtains weighted representation information of the program name information of the second application, and determines the first fused representation information based on the weighted representation information of the program name information of the first application and the second application.

[0046] In one embodiment, the object recognition device determines the second fused representation information according to the second representation information in a manner as follows: the object recognition device obtains resource feature information of the target application, the resource feature information may be a resource transfer amount, and uses the resource feature information of the target application to perform weighted processing on the second representation information of the interface text information of the target application to obtain weighted representation information of the interface text information of the target application, and determines the second fused representation information according to the weighted representation information of the interface text information of each application of the one or more applications; the target application is any one of the one or more applications, for example, the applications associated with the object to be recognized are the first application and the second application, then the object recognition device performs weighted processing on the second representation information of the interface text information of the first application using the resource feature information of the first application to obtain weighted representation information of the interface text information of the first application, and the object recognition device performs weighted processing on the second representation information of the interface text information of the second application using the resource feature information of the second application to obtain weighted representation information of the interface text information of the second application. The second fused representation information is determined according to the weighted representation information of the interface text information of the first application and the second application.

[0047] In one embodiment, the first fused representation information is determined based on the weighted representation information of the program name information of each application program. When there is only one application program associated with the object to be identified, the object recognition device determines the weighted representation information of the program name information of the application program as the first fused representation information. If there are multiple applications program associated with the object to be identified, the object recognition device fuses the weighted representation information of the program name information of the multiple applications to obtain the first fused representation information. For example, if the applications program associated with the object to be identified are the first application program and the second application program, the object recognition device fuses the weighted representation information of the program name information of the first application program with the weighted representation information of the program name information of the second application program to obtain the first fused representation information.

[0048] In one embodiment, the second fused representation information is determined based on the weighted representation information of the interface text information of each application program. When there is only one application program associated with the object to be identified, the object recognition device determines the weighted representation information of the interface text information of the application program as the second fused representation information. If there are multiple applications associated with the object to be identified, the object recognition device fuses the weighted representation information of the interface text information of the multiple applications to obtain the second fused representation information. For example, if the applications associated with the object to be identified are the first application program and the second application program, the object recognition device fuses the weighted representation information of the interface text information of the first application program with the weighted representation information of the interface text information of the second application program to obtain the second fused representation information.

[0049] Among them, when there are one or more applications that have an association relationship with the object to be identified, the weighted processing of the first characterization information and the second characterization information is selective. For example, one of the first characterization information and the second characterization information can be weighted, neither of the first characterization information and the second characterization information can be weighted, or both the first characterization information and the second characterization information can be weighted. Different weighted processing schemes can be selected in different situations. For example, when it is necessary to reduce the amount of data operations, it is possible to choose not to weight the first characterization information and the second characterization information. When it is necessary to improve the accuracy of object recognition, both the first characterization information and the second characterization information can be weighted.

[0050] In one embodiment, see Figure 4 , Figure 4 FIG. 1 is a flow chart of an exemplary embodiment of the present application providing a method of using a model to identify an object. Figure 4As shown, the object recognition device classifies and recognizes the object to be recognized based on the information to be processed, and the classification recognition result can be obtained by inputting the program name information and interface text information of the application of the object to be recognized into the target classification model for processing to obtain the classification recognition result.

[0051] See also Figure 5 , Figure 5 is a schematic diagram of a network structure of a classification model provided by an exemplary embodiment of the present application, such as Figure 5 As shown, the classification model may include a representation extraction network, a representation fusion network, a representation splicing network, and a classification recognition network. The output information of the representation extraction network is input into the representation fusion network, the output information of the representation fusion network is input into the representation splicing network, and the output information of the representation splicing network is input into the classification recognition network.

[0052] In one embodiment, the object recognition device obtains first representation information of program name information of each application in one or more applications, and obtains second representation information of interface text information of each application in one or more applications. The method may be as follows: the object recognition device puts the program name information and interface text information of each application in one or more applications into a first representation extraction network and a second representation extraction network, respectively, to obtain first representation information of each program name information and second representation information of each interface text information in one or more applications. For example, the applications associated with the object to be recognized include a first application, a second application, ..., and an nth application. Figure 5 As shown, the object recognition device puts the program name information of the first application into the first representation extraction network to obtain the first representation information of the program name information of the first application, puts the program name information of the second application into the first representation extraction network to obtain the first representation information of the program name information of the second application, ..., puts the program name information of the nth application into the first representation extraction network to obtain the first representation information of the program name information of the nth application, puts the interface text information of the first application into the second representation extraction network to obtain the second representation information of the interface text information of the first application, puts the interface text information of the second application into the second representation extraction network to obtain the second representation information of the interface text information of the second application, ..., puts the interface text information of the nth application into the second representation extraction network to obtain the second representation information of the interface text information of the nth application, and the first representation extraction network and the second representation extraction network can be the same representation extraction network or different representation extraction networks.

[0053] In one embodiment, the first representation information of the program name information of the target application is weighted by using the resource feature information of the target application, and the weighted representation information of the program name information of the target application can be obtained by: the object recognition device puts the resource feature information of the target application and the first representation information of the program name information of the target application into a representation fusion network to obtain the weighted representation information of the program name information of the target application. For example, the applications associated with the object to be identified are the first application, the second application, ..., the nth application, then Figure 5 As shown, the object recognition device puts the resource feature information v1 of the first application and the first representation information of the program name information of the first application into the representation fusion network to obtain the weighted representation information of the program name information of the first application, the object recognition device puts the resource feature information v2 of the second application and the first representation information of the program name information of the second application into the representation fusion network to obtain the weighted representation information of the program name information of the second application, ..., the object recognition device puts the resource feature information vn of the nth application and the second representation information of the program name information of the nth application into the representation fusion network to obtain the weighted representation information of the program name information of the nth application.

[0054] In one embodiment, the second representation information of the interface text information of the target application is weighted by using the resource feature information of the target application, and the weighted representation information of the interface text information of the target application can be obtained by placing the resource feature information of the target application and the second representation information of the interface text information of the target application into a representation fusion network to obtain the weighted representation information of the interface text information of the target application. For example, if the application having an association relationship with the object to be identified is the first application, the second application, ..., the nth application, then Figure 5 As shown, the object recognition device puts the resource feature information v1 of the first application and the second representation information of the interface text information of the first application into the representation fusion network to obtain the weighted representation information of the interface text information of the first application, the object recognition device puts the resource feature information v2 of the second application and the second representation information of the interface text information of the second application into the representation fusion network to obtain the weighted representation information of the interface text information of the second application, ..., the object recognition device puts the resource feature information vn of the nth application and the second representation information of the interface text information of the nth application into the representation fusion network to obtain the weighted representation information of the interface text information of the nth application.

[0055] In one embodiment, when there are multiple applications associated with the object to be identified, the object identification device fuses the weighted representation information of the program name information of the multiple applications to obtain the first fused representation information: the object identification device puts the weighted representation information of the program name information of the multiple applications into the representation fusion network to obtain the first fused representation information. For example, if the applications associated with the object to be identified are the first application, the second application, ..., the nth application, then Figure 5 As shown, the object recognition device puts the weighted representation information of the program name information of the first application and the weighted representation information of the program name information of the second application, ..., the weighted representation information of the program name information of the nth application into a representation fusion network to obtain first fused representation information.

[0056] In one embodiment, when there are multiple applications associated with the object to be identified, the object identification device fuses the weighted representation information of the interface text information of the multiple applications to obtain the second fused representation information. The object identification device puts the weighted representation information of the interface text information of the multiple applications into the representation fusion network to obtain the second fused representation information. For example, if the applications associated with the object to be identified are the first application, the second application, ..., the nth application, then Figure 5 As shown, the object recognition device puts the weighted representation information of the interface text information of the first application, the weighted representation information of the interface text information of the second application, ..., the weighted representation information of the interface text information of the nth application into a representation fusion network to obtain second fused representation information.

[0057] In one embodiment, the first fused representation information and the second fused representation information are spliced ​​to obtain the spliced ​​representation information: when there are one or more applications associated with the object to be identified, the object recognition device puts the first fused representation information and the second fused representation information into a representation splicing network to obtain the spliced ​​representation information.

[0058] In one embodiment, the object to be identified is classified and identified based on the spliced ​​representation information, and the classification and identification result can be: the object recognition device puts the spliced ​​representation information into the classification and identification network to determine the classification and identification result, which includes a first probability that the object to be identified belongs to the target category and a second probability that the object to be identified belongs to the reference category.

[0059] S203: Determine an object recognition result of the object to be recognized according to the first probability and the second probability, where the object recognition result is used to indicate whether the object to be recognized belongs to the target category.

[0060] In an embodiment of the present application, the object recognition device determines the object recognition result of the object to be recognized based on the first probability and each second probability included in the classification recognition result. In one embodiment, the object recognition device can process the first probability and each second probability included in the classification recognition result by calling the target decision tree model to obtain the object recognition result of the object to be recognized.

[0061] By first determining the probability that the object to be identified belongs to the target category and the reference category respectively, and then finally determining whether the object to be identified belongs to the target category according to the probability that the object to be identified belongs to the target category and the reference category respectively, the accuracy of object identification can be improved. The reference category can be one or more. In one embodiment, the reference category can be a normal category, and the target category can be an abnormal category.

[0062] In one embodiment, see Figure 6 , Figure 6 is a schematic diagram of obtaining an object recognition result provided by an exemplary embodiment of the present application. The object recognition result of the object to be recognized can be determined according to the first probability and the second probability: the object recognition device obtains the resource feature information of the object to be recognized; the object recognition device inputs the probability that the object to be recognized belongs to the target category and the reference category respectively and the resource feature information of the object to be recognized into the target decision tree model, and obtains the object recognition result of the object to be recognized, such as Figure 6 As shown, when there are k reference categories, the object recognition device inputs the probability that the object to be recognized belongs to the target category, the probability that the object to be recognized belongs to reference category 1, the probability that the object to be recognized belongs to reference category 2, ..., the probability that the object to be recognized belongs to reference category k and the resource feature information of the object to be recognized into the target decision tree model to obtain the object recognition result of the object to be recognized, and further improves the accuracy and coverage of object recognition by integrating more clues.

[0063] Among them, the target decision tree model is obtained by training the initial decision tree model according to the labeled category information, the predicted classification results of the object to be identified and the resource feature information.

[0064] The object recognition method is implemented based on the method provided in the embodiment of the present application, which can realize the automation and intelligence of object recognition, reduce the amount of manual review, and improve the accuracy and coverage of object recognition by integrating a large amount of information.

[0065] Based on this description, an example can be used to briefly describe this application. In the product transaction application scenario, it is determined whether the product operator is an abnormal category. The abnormality is the second clearing. The second clearing means that the product operator performs virtual resource clearing in substance without obtaining a payment license. In this scenario, the object to be identified is the product operator to be identified. The information to be processed of the object to be identified is the program name information and interface text information of the product transaction application. The interface text information of the product transaction application includes the text content in the application interface of the product transaction application. The product transaction application has an association with the product operator to be identified. For example, the product operator to be identified can be the owner of the product transaction application. The object recognition device classifies and identifies the product operator to be identified according to the information to be processed, and obtains the classification recognition result. The classification recognition result includes: the probability that the product operator to be identified is an abnormal category, and the probability that the product operator to be identified is a reference category. The reference category is a given category with reference significance, which can be one or more. For example, the reference category is clothing accessories and sports and outdoor. The object recognition device determines the object recognition result according to the probability that the product operator to be identified is of an abnormal category and the probability that the product operator to be identified is of a reference category. The object recognition result is the probability that the product operator is of an abnormal category.

[0066] Based on this description, another example can be used to simply describe this application. In a video viewing scenario, it is determined whether the video is a target category, and the target is animation. In this scenario, the object to be identified is a video to be identified, and the information to be processed of the object to be identified is the program name information and interface text information of the video application. The interface text information of the video application includes the text content in the application interface of the video application. The video application has an association relationship with the video to be identified. For example, the video to be identified can be the owner of the video application. The object recognition device classifies and identifies the video to be identified according to the information to be processed, and obtains a classification recognition result. The classification recognition result includes: the probability that the video to be identified is a target category, the probability that the video to be identified is a reference category, and the reference category is a given category with reference significance, which can be one or more. For example, the reference category is a sports category, a teaching and tutoring category. The object recognition device determines the object recognition result according to the probability that the video to be identified is a target category and the probability that the video to be identified is a reference category. The object recognition result is the probability that the video to be identified is a target category.

[0067] The following is an introduction to the classification model and decision tree model training method involved in this embodiment. Figure 7 , Figure 7 A flow chart of a model training method provided in an embodiment of the present application. The model training method described in the embodiment of the present application includes but is not limited to the following steps:

[0068] S301. Obtain a sample data group, where the sample data group includes: sample program name information and sample interface text information of a sample application, annotated category information of a sample object, and sample resource feature information of the sample object. The sample resource feature information may be a sample resource transfer amount.

[0069] In an embodiment of the present application, the sample application has an associated relationship with the sample object, the sample interface text information includes the text content in the application interface of the sample application, the annotated category information of the sample object includes a label of whether the sample object is a target category and a label of the sample object is a given category, and the sample application that has an associated relationship with the sample object may be one or more. When there are multiple sample applications that have an associated relationship with the sample object, the sample data group includes sample program name information, sample interface text information and sample resource feature information of each sample application, annotated category information of the sample object, and sample resource feature information of the sample object. For example, the sample applications that have an associated relationship with the sample object are the first sample application and the second sample application, then the sample data group includes the sample program name information of the first sample application, the sample program name information of the second sample application, the sample interface text information of the first sample application, the sample interface text information of the second sample application, the sample resource feature information of the first sample application, the sample resource feature information of the second sample application, the annotated category information of the sample object, and the sample resource feature information of the sample object.

[0070] S302. Perform representation extraction processing on the sample program name information of the sample application and the sample interface text information of the sample application using the representation extraction network included in the initial classification model to obtain first sample representation information of the sample program name information of the sample application and second sample representation information of the sample interface text information of the sample application.

[0071] In one embodiment, when there are one or more sample applications associated with the sample object, the object recognition device puts the sample application name information and the sample interface text information of each sample application in the one or more sample applications into the first representation extraction network and the second representation extraction network respectively, and obtains the first sample representation information of each sample application name information and the second sample representation information of each sample interface text information in the one or more sample applications. For example, if the applications associated with the sample object are the first sample application and the second sample application, the object recognition device puts the sample application name information of the first sample application into the first representation extraction network to obtain the first sample representation information of the sample application name information of the first sample application, puts the sample application name information of the second sample application into the first representation extraction network to obtain the first sample representation information of the sample application name information of the second sample application, puts the sample interface text information of the first sample application into the second representation extraction network to obtain the second sample representation information of the sample interface text information of the first sample application, puts the sample interface text information of the second sample application into the second representation extraction network to obtain the second sample representation information of the sample interface text information of the second sample application, and the first representation extraction network and the second representation extraction network can be the same representation network or different representation extraction networks.

[0072] S303: Use the representation concatenation network included in the initial classification model to fuse the first sample representation information and the second sample representation information to obtain sample concatenation representation information.

[0073] In one embodiment, when there are one or more sample applications associated with a sample object, the object recognition device puts the sample resource feature information of each sample application in the one or more sample applications and the first sample representation information of the sample program name information of each sample application in the one or more sample applications into a representation fusion network to obtain sample weighted representation information of the sample program name information of each sample application in the one or more sample applications; the object recognition device puts the sample resource feature information of each sample application in the one or more sample applications and the second sample representation information of the sample interface text information of each sample application in the one or more sample applications into the representation fusion network to obtain sample weighted representation information of the sample interface text information of each sample application in the one or more sample applications.

[0074] In one embodiment, when there is only one sample application associated with the sample object, the object recognition device determines the sample weighted representation information of the sample program name information of the sample application as the first sample fusion representation information.

[0075] In one embodiment, when there is only one sample application associated with the sample object, the object recognition device determines the sample weighted representation information of the sample interface text information of the sample application as the second sample fusion representation information.

[0076] In one embodiment, when there are multiple sample applications associated with a sample object, the object recognition device places weighted representation information of sample program name information of the multiple sample applications into a representation fusion network to determine first sample fusion representation information. For example, when the sample applications associated with the sample object are a first sample application and a second sample application, the object recognition device places weighted representation information of the sample program name information of the first sample application and weighted representation information of the sample program name information of the second sample application into the representation fusion network to determine first sample fusion representation information.

[0077] In one embodiment, when there are multiple sample applications associated with the sample object, the object recognition device puts weighted representation information of sample interface text information of the multiple sample applications into the representation fusion network to determine the second sample fusion representation information. For example, when the sample applications associated with the sample object are the first sample application and the second sample application, the object recognition device puts the weighted representation information of the sample interface text information of the first sample application and the weighted representation information of the sample interface text information of the second sample application into the representation fusion network to determine the second sample fusion representation information.

[0078] In one embodiment, when there are one or more sample applications associated with the sample object, the object recognition device puts the first sample fusion representation information and the second sample fusion representation information into a representation splicing network to obtain sample splicing representation information.

[0079] S304: Process the sample splicing representation information using the classification recognition network included in the initial classification model to obtain a predicted classification result of the sample object, where the predicted classification result includes a probability that the sample object belongs to the target category and a probability that the sample object belongs to the reference category.

[0080] In the embodiment of the present application, the reference category may be one or more. In one implementation, the reference category may be a normal category, and the target category may be an abnormal category. The object recognition device puts the sample splicing representation information into the classification recognition network to obtain the sample classification recognition result.

[0081] S305: Determine difference data according to the labeled category information and the predicted classification results of the sample object, and adjust the model parameters of the initial classification model according to the difference data to obtain a target classification model.

[0082] In the embodiment of the present application, the difference data may be one or more. For example, the labeled category information contained in 10,000 sample data groups and the predicted classification results of the sample objects contained in the corresponding 10,000 sample data groups are obtained to obtain 10,000 difference data. The difference data is determined by the deviation degree obtained by comparing the labeled category information with the predicted classification results obtained by the initial classification model. For example, the labeled category information contained in the sample data group is that the sample object is the target category and the sample object is category A. The predicted classification results obtained by the initial classification model are that the probability that the sample object belongs to the target category is 90% and the probability that the sample object belongs to category A is 80%. At this time, the deviation degree obtained by comparing the labeled category information with the predicted classification results obtained by the initial classification model is 10% and 20%. The object recognition device determines the difference data according to the deviation degree. The object recognition device adjusts the relevant network parameters in the initial classification model according to the difference data to minimize the prediction error of the initial classification model. When the parameters reach the convergence condition, the training is stopped to obtain the target classification model.

[0083] S306: Train the initial decision tree model using the labeled category information, the predicted classification results of the sample objects, and the resource feature information to obtain a target decision tree model.

[0084] In an embodiment of the present application, an initial decision tree model is trained based on a label of whether the sample object is a target category in the labeled category information contained in the sample data group and a probability that the sample object belongs to the target category in a predicted classification result of the sample object obtained by a target classification model. For example, the label of whether the sample object is a target category in the labeled category information contained in the sample data group is that the sample object is a target category, and the probability that the sample object belongs to the target category in the predicted classification result of the sample object obtained by the target classification model is 90%. The initial decision tree model is trained based on the above information to obtain a target decision tree model.

[0085] By using the model training method provided in the embodiment of the present application, the classification model and the decision tree model are jointly trained by combining multiple information to obtain a target classification model and a target decision tree model, which can make the results obtained in object recognition using the target classification model and the target decision tree model more accurate.

[0086] See also Figure 8 , Figure 8 : is a schematic diagram of the structure of an object recognition device provided in an embodiment of the present application. The object recognition device described in the embodiment of the present application corresponds to the object recognition device described above, and the object recognition device includes:

[0087] The acquisition unit 801 is used to: acquire information to be processed of the object to be identified, the information to be processed includes program name information and interface text information of an application, the application has an association relationship with the object to be identified, and the interface text information includes text content in an application interface of the application;

[0088] The processing unit 802 is used to: perform classification and recognition processing on the object to be recognized according to the information to be processed to obtain a classification and recognition result, wherein the classification and recognition result includes a first probability that the object to be recognized belongs to a target category and a second probability that the object to be recognized belongs to a reference category;

[0089] The processing unit 802 is further configured to determine an object recognition result of the object to be recognized according to the first probability and the second probability, wherein the object recognition result is used to indicate whether the object to be recognized belongs to the target category.

[0090] In one embodiment, the processing unit 802 is specifically used to: obtain resource feature information of the object to be identified; and determine an object recognition result of the object to be identified according to the first probability, the second probability and the resource feature information of the object to be identified.

[0091] In one embodiment, there are one or more applications associated with the object to be identified, and the information to be processed includes: program name information and interface text information of each application among the one or more applications associated with the object to be identified, and the processing unit 802 is specifically used to: respectively obtain first representation information of the program name information of each application, and respectively obtain second representation information of the interface text information of each application; determine first fused representation information based on the first representation information, and determine second fused representation information based on the second representation information; splice the first fused representation information and the second fused representation information to obtain spliced ​​representation information; and classify and identify the object to be identified based on the spliced ​​representation information to obtain a classification recognition result.

[0092] In one embodiment, the processing unit 802 is specifically used to: obtain resource feature information of each of the applications; use the resource feature information of the target application to weightedly process the first representation information of the program name information of the target application to obtain weighted representation information of the program name information of the target application, and determine the first fused representation information based on the weighted representation information of the program name information of each of the applications; the target application is any one of the one or more applications; use the resource feature information of the target application to weightedly process the second representation information of the interface text information of the target application to obtain weighted representation information of the interface text information of the target application, and determine the second fused representation information based on the weighted representation information of the interface text information of each of the applications.

[0093] In one embodiment, the processing unit 802 is specifically used to: if there is only one application that has an association relationship with the object to be identified, then determine the weighted representation information of the program name information of the application as the first fused representation information; if there are multiple applications that have an association relationship with the object to be identified, then fuse the weighted representation information of the program name information of the multiple applications to obtain the first fused representation information.

[0094] In one embodiment, the processing unit 802 is specifically used to: if there is only one application that has an associated relationship with the object to be identified, then determine the weighted representation information of the interface text information of the application as the second fused representation information; if there are multiple applications that have an associated relationship with the object to be identified, then fuse the weighted representation information of the interface text information of the multiple applications to obtain the second fused representation information.

[0095] In one embodiment, the processing unit 802 is specifically used to: input the program name information and the interface text information into a target classification model for processing to obtain a classification recognition result;

[0096] The target classification model is obtained by training the initial classification model using the sample data group, and the sample data group includes: sample program name information and sample interface text information of the sample application, and annotated category information of the sample object, and the sample application has an associated relationship with the sample object; the initial classification model includes a representation extraction network, a representation fusion network, a representation splicing network and a classification recognition network; the representation extraction network is used to extract the first sample representation information of the sample program name information and the second sample representation information of the sample interface text information respectively; the representation fusion network is used to determine the first sample fusion representation information according to the first sample representation information, and to determine the second sample fusion representation information according to the second sample representation information; the representation splicing network is used to determine the sample splicing representation information according to the first sample fusion representation information and the second sample fusion representation information; the classification recognition network is used to determine the predicted classification result of the sample object according to the sample splicing representation information, and the predicted classification result includes the probability that the sample object belongs to the target category and the probability that it belongs to the reference category; the difference data referred to by the initial classification model during the training process is determined according to the annotated category information and the predicted classification result.

[0097] In one embodiment, the sample data group further includes resource feature information of the sample object, and the processing unit 802 is specifically used to: input the first probability, the second probability and the resource feature information of the object to be identified into a target decision tree model for processing to obtain an object recognition result of the object to be identified;

[0098] The target decision tree model is obtained by training an initial decision tree model according to the labeled category information, the predicted classification results of the object to be identified, and resource feature information.

[0099] It can be understood that the functions of each functional unit of the object recognition device in the embodiment of the present application can be specifically implemented according to the object recognition method in the above method embodiment, and its specific implementation process can refer to the relevant description in the above object recognition method embodiment, which will not be repeated here.

[0100] See also Fig. 9 , Fig. 9 A schematic diagram of the structure of a computer device provided in an embodiment of the present application. The computer device described in the embodiment of the present application includes: a processor 901, a communication interface 902, and a memory 903. Among them, the processor 901, the communication interface 902, and the memory 903 can be connected via a bus or other means, and the embodiment of the present application takes the connection via a bus as an example.

[0101] Among them, the processor 901 (or CPU (Central Processing Unit)) is the computing core and control core of the computer device, which can parse various instructions in the computer device and process various data of the computer device. For example, the CPU can be used to parse the power on and off instructions sent by the user to the computer device, and control the computer device to perform power on and off operations; for another example, the CPU can transmit various interactive data between the internal structures of the computer device, and so on. The communication interface 902 can optionally include a standard wired interface, a wireless interface (such as Wi-Fi, a mobile communication interface, etc.), which is controlled by the processor 901 to send and receive data. The memory 903 (Memory) is a memory device in the computer device for storing programs and data. It can be understood that the memory 903 here can include both the built-in memory of the computer device and the extended memory supported by the computer device. The memory 903 provides a storage space, which stores the operating system of the computer device, which may include but is not limited to: Android system, iOS system, Windows Phone system, etc., and this application does not limit this.

[0102] In the embodiment of the present application, the processor 901 performs the following operations by running the executable program code in the memory 903:

[0103] Acquire, through the communication interface 902, information to be processed of the object to be identified, the information to be processed including program name information and interface text information of an application program, the application program having an association relationship with the object to be identified, the interface text information including text content in an application program interface of the application program;

[0104] Performing classification and recognition processing on the object to be recognized according to the information to be processed to obtain a classification and recognition result, wherein the classification and recognition result includes a first probability that the object to be recognized belongs to the target category and a second probability that the object to be recognized belongs to the reference category;

[0105] An object recognition result of the object to be recognized is determined according to the first probability and the second probability, and the object recognition result is used to indicate whether the object to be recognized belongs to the target category.

[0106] In one embodiment, the processor 901 is specifically configured to: obtain resource feature information of the object to be identified; and determine an object recognition result of the object to be identified according to the first probability, the second probability and the resource feature information of the object to be identified.

[0107] In one embodiment, there are one or more applications associated with the object to be identified, and the information to be processed includes: program name information and interface text information of each application in the one or more applications associated with the object to be identified, and the processor 901 is specifically used to: respectively obtain first representation information of the program name information of each application, and respectively obtain second representation information of the interface text information of each application; determine first fused representation information based on the first representation information, and determine second fused representation information based on the second representation information; splice the first fused representation information and the second fused representation information to obtain spliced ​​representation information; and classify and identify the object to be identified based on the spliced ​​representation information to obtain a classification recognition result.

[0108] In one embodiment, the processor 901 is specifically used to: obtain resource feature information of each of the applications; use the resource feature information of the target application to weightedly process the first representation information of the program name information of the target application to obtain weighted representation information of the program name information of the target application, and determine the first fused representation information based on the weighted representation information of the program name information of each of the applications; the target application is any one of the one or more applications; use the resource feature information of the target application to weightedly process the second representation information of the interface text information of the target application to obtain weighted representation information of the interface text information of the target application, and determine the second fused representation information based on the weighted representation information of the interface text information of each of the applications.

[0109] In one embodiment, the processor 901 is specifically used to: if there is one application that has an association relationship with the object to be identified, then determine the weighted representation information of the program name information of the application as the first fused representation information; if there are multiple applications that have an association relationship with the object to be identified, then fuse the weighted representation information of the program name information of the multiple applications to obtain the first fused representation information.

[0110] In one embodiment, the processor 901 is specifically used to: if there is only one application that has an association relationship with the object to be identified, then determine the weighted representation information of the interface text information of the application as the second fused representation information; if there are multiple applications that have an association relationship with the object to be identified, then fuse the weighted representation information of the interface text information of the multiple applications to obtain the second fused representation information.

[0111] In one embodiment, the processor 901 is specifically used to: input the program name information and the interface text information into a target classification model for processing to obtain a classification recognition result;

[0112] Among them, the target classification model is obtained by training the initial classification model using the sample data group, and the sample data group includes: sample program name information and sample interface text information of the sample application, and annotated category information of the sample object, and the sample application has an associated relationship with the sample object; the initial classification model includes a representation extraction network, a representation fusion network and a classification recognition network; the representation extraction network is used to extract the first sample representation information of the sample program name information and the second sample representation information of the sample interface text information respectively; the representation fusion network is used to determine the sample splicing representation information according to the first sample representation information and the second sample representation information; the classification recognition network is used to determine the predicted classification result of the sample object according to the sample splicing representation information, and the predicted classification result includes the probability that the sample object belongs to the target category and the probability that it belongs to the reference category; the difference data referenced by the initial classification model for training is determined according to the annotated category information and the predicted classification result.

[0113] In one embodiment, the sample data group further includes resource feature information of the sample object, and the processor 901 is specifically configured to: input the first probability, the second probability and the resource feature information of the object to be identified into a target decision tree model for processing to obtain an object recognition result of the object to be identified;

[0114] The target decision tree model is obtained by training an initial decision tree model according to the labeled category information, the predicted classification results of the object to be identified, and resource feature information.

[0115] In a specific implementation, the processor 901, communication interface 902 and memory 903 described in the embodiments of the present application can execute the implementation of the computer device described in the object recognition method provided in the embodiments of the present application, and can also execute the implementation described in the object recognition device provided in the embodiments of the present application, which will not be repeated here.

[0116] The embodiment of the present application further provides a computer-readable storage medium, in which instructions are stored. When the computer-readable storage medium is executed on a computer, the computer implements the object recognition method provided by the embodiment of the present application.

[0117] The embodiment of the present application further provides a computer program product, which includes a computer program or computer instructions. When the computer program or computer instructions are executed by a processor, the object recognition method provided by the embodiment of the present application is implemented.

[0118] It should be noted that, for the above-mentioned various method embodiments, for the sake of simplicity of description, they are all expressed as a series of action combinations, but those skilled in the art should be aware that this application is not limited by the order of the actions described, because according to this application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by this application.

[0119] A person of ordinary skill in the art may understand that all or part of the steps in the various methods of the embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0120] The above disclosure is only part of the embodiments of the present application, which certainly cannot be used to limit the scope of rights of the present application. Therefore, equivalent changes made according to the claims of the present application are still within the scope covered by the present application.

Claims

1. An object recognition method, characterized in that: The method comprises: Acquire information to be processed of the object to be identified, the information to be processed includes program name information and interface text information of an application, the application has an association relationship with the object to be identified, and the interface text information includes text content in an application interface of the application; Performing classification and recognition processing on the object to be recognized according to the information to be processed to obtain a classification and recognition result, wherein the classification and recognition result includes a first probability that the object to be recognized belongs to a target category and a second probability that the object to be recognized belongs to a reference category; An object recognition result of the object to be recognized is determined according to the first probability and the second probability, and the object recognition result is used to indicate whether the object to be recognized belongs to the target category.

2. The method according to claim 1, characterized in that The determining the object recognition result of the object to be recognized according to the first probability and the second probability includes: Acquire resource feature information of the object to be identified; An object recognition result of the object to be recognized is determined according to the first probability, the second probability and the resource feature information of the object to be recognized.

3. The method according to claim 1 or 2, characterized in that There are one or more applications associated with the object to be identified, and the information to be processed includes: program name information and interface text information of each of the one or more applications associated with the object to be identified; The classifying and identifying the object to be identified according to the information to be processed to obtain a classification and identification result includes: respectively obtaining first representation information of program name information of each application program, and respectively obtaining second representation information of interface text information of each application program; Determine first fused representation information according to the first representation information, and determine second fused representation information according to the second representation information; Splicing the first fused representation information and the second fused representation information to obtain spliced ​​representation information; The object to be identified is classified and identified according to the splicing representation information to obtain a classification and identification result.

4. The method according to claim 3, characterized in that The determining first fused representation information according to the first representation information, and determining second fused representation information according to the second representation information includes: Obtain resource feature information of each of the applications; Using the resource feature information of the target application, weighted processing is performed on the first representation information of the program name information of the target application to obtain the weighted representation information of the program name information of the target application, and the first fusion representation information is determined according to the weighted representation information of the program name information of each application; the target application is any one of the one or more applications; The resource feature information of the target application is used to perform weighted processing on the second representation information of the interface text information of the target application to obtain the weighted representation information of the interface text information of the target application, and the second fused representation information is determined based on the weighted representation information of the interface text information of each application.

5. The method according to claim 4, characterized in that The determining the first fused representation information according to the weighted representation information of the program name information of each application program includes: If there is only one application associated with the object to be identified, determining weighted representation information of program name information of the application as first fused representation information; If there are multiple applications associated with the object to be identified, weighted representation information of program name information of the multiple applications is fused to obtain first fused representation information.

6. The method according to claim 1, characterized in that The information to be processed includes the program name information and the interface text information, and the classification and recognition processing of the object to be recognized according to the information to be processed to obtain the classification and recognition result includes: Inputting the program name information and the interface text information into a target classification model for processing to obtain a classification recognition result; The target classification model is obtained by training the initial classification model using the sample data group, and the sample data group includes: sample program name information and sample interface text information of the sample application, and annotated category information of the sample object, and the sample application has an associated relationship with the sample object; the initial classification model includes a representation extraction network, a representation fusion network, a representation splicing network and a classification recognition network; the representation extraction network is used to extract the first sample representation information of the sample program name information and the second sample representation information of the sample interface text information respectively; the representation fusion network is used to determine the first sample fusion representation information according to the first sample representation information, and to determine the second sample fusion representation information according to the second sample representation information; the representation splicing network is used to determine the sample splicing representation information according to the first sample fusion representation information and the second sample fusion representation information; the classification recognition network is used to determine the predicted classification result of the sample object according to the sample splicing representation information, and the predicted classification result includes the probability that the sample object belongs to the target category and the probability that it belongs to the reference category; the difference data referred to by the initial classification model during the training process is determined according to the annotated category information and the predicted classification result.

7. The method according to claim 6, characterized in that The sample data group also includes resource feature information of the sample object, and determining the object recognition result of the object to be recognized according to the first probability and the second probability includes: Inputting the first probability, the second probability and the resource feature information of the object to be identified into a target decision tree model for processing to obtain an object recognition result of the object to be identified; The target decision tree model is obtained by training an initial decision tree model according to the labeled category information, the predicted classification results of the object to be identified, and resource feature information.

8. An object recognition device, characterized in that: The device comprises: an acquisition unit, configured to acquire information to be processed of the object to be identified, wherein the information to be processed includes program name information and interface text information of an application program, wherein the application program has an association relationship with the object to be identified, and the interface text information includes text content in an application program interface of the application program; a processing unit, configured to perform classification and recognition processing on the object to be recognized according to the information to be processed, and obtain a classification and recognition result, wherein the classification and recognition result includes a first probability that the object to be recognized belongs to a target category and a second probability that the object to be recognized belongs to a reference category; The processing unit is used to determine an object recognition result of the object to be recognized according to the first probability and the second probability, and the object recognition result is used to indicate whether the object to be recognized belongs to the target category.

9. A computer device, characterized in that: include: A processor, a communication interface and a memory, wherein the processor, the communication interface and the memory are connected to each other, wherein the memory stores an executable program code, and the processor is used to call the executable program code to implement the object recognition method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, which, when executed on a computer, enable the computer to implement the object recognition method according to any one of claims 1 to 7.

11. A computer program product, characterized in that The computer program product comprises a computer program or a computer instruction, and when the computer program or the computer instruction is executed by a processor, the object recognition method according to any one of claims 1 to 7 is implemented.

Images