Object matching method and related device

Abstract

The present disclosure provides an object matching method, including: determining feature information of a target object based on attribute information of the target object; in response to determining that the target object is a non-seed object, determining at least one first candidate object matching the target object from a seed object pool based on the feature information of the target object, and establishing an association relationship pair from the first candidate object to the target object; and in response to determining that the target object is a seed object, determining at least one second candidate object matching the target object from an object pool based on the feature information of the target object, and establishing an association relationship pair from the target object to the second candidate object. The present disclosure also provides an object matching device, an electronic device, a storage medium, and a program product.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to an object matching method and related devices. Background Art

[0002] As e-commerce becomes more and more deeply rooted in people's lives, users have more and more needs to find the same products through the Internet. Therefore, how to quickly find the same products that match a certain product among a large number of products has become a problem that many e-commerce applications need to solve. Summary of the invention

[0003] In view of this, an embodiment of the present disclosure provides an object matching method, which can instantly determine at least one object that matches the feature information of the target object at a low cost after the attribute information of the target object is updated, and establish an association relationship pair between the target object and the object it matches, thereby improving the timeliness and accuracy of searching for the same object in a massive object pool.

[0004] The object matching method described in the embodiment of the present disclosure may include: determining feature information of the target object based on the attribute information of the target object; in response to determining that the target object is a non-seed object, determining at least one first candidate object matching the target object from a seed object pool based on the feature information of the target object, and establishing an association relationship pair from the first candidate object to the target object; and in response to determining that the target object is a seed object, determining at least one second candidate object matching the target object from an object pool based on the feature information of the target object, and establishing an association relationship pair from the target object to the second candidate object.

[0005] In the method described in the embodiment of the present disclosure, determining at least one first candidate object matching the target object from a seed object pool based on the feature information of the target object includes: performing object recall in the seed object pool using the feature information of the target object as a query condition to obtain at least one first seed object; determining the similarity between the at least one first seed object and the target object respectively; and determining the at least one first candidate object from the at least one first seed object based on the similarity between the at least one first seed object and the target object.

[0006] In the method described in the embodiment of the present disclosure, respectively determining the similarity between the at least one first seed object and the target object includes: filtering the at least one first seed object based on a preset filtering condition to obtain the at least one first seed object after filtering; and respectively determining the similarity between the at least one first seed object after filtering and the target object.

[0007] In the method described in the embodiment of the present disclosure, respectively determining the similarity between the at least one first seed object and the target object includes: respectively establishing at least one feature information pair from the query condition to the feature information of at least one object in the recall result; performing a reversal operation on the feature information in the at least one established feature information pair to obtain at least one inverted feature information pair; and inputting the at least one inverted feature information pair into a similarity algorithm model to determine the similarity between the at least one first seed object and the target object.

[0008] In the method described in the embodiment of the present disclosure, determining at least one first candidate object matching the target object from a seed object pool based on the feature information of the target object further includes: searching for at least one second seed object having an association relationship with the target object in an established association relationship pair; respectively determining the similarity between the at least one second seed object and the target object; determining at least one third candidate object matching the target object from the at least one second seed object based on the similarity between the at least one second seed object and the target object; and establishing an association relationship pair between the third candidate object and the target object.

[0009] In the method described in the embodiment of the present disclosure, determining at least one second candidate object matching the target object from the object pool based on the feature information of the target object includes: performing object recall in the object pool using the feature information of the target object as a query condition to obtain at least one pool object; respectively determining the similarity between the target object and the at least one pool object; and determining the at least one second candidate object from the at least one pool object based on the similarity between the target object and the at least one pool object.

[0010] In the method described in the embodiment of the present disclosure, respectively determining the similarity between the target object and the at least one pot object includes: respectively establishing at least one feature information pair from the query condition to the feature information of at least one object in the recall result; and inputting the at least one feature information pair into a similarity algorithm model to determine the similarity between the target object and the at least one pot object.

[0011] Corresponding to the above object matching method, the embodiment of the present disclosure further discloses an object matching device. The above object matching device may include:

[0012] A feature determination module, used to determine feature information of a target object based on attribute information of the target object;

[0013] A classification module, used to determine whether the target object is a seed object;

[0014] a first processing module, configured to, in response to determining that the target object is a non-seed object, determine at least one first candidate object matching the target object from a seed object pool based on feature information of the target object, and establish an association relationship pair from the first candidate object to the target object; and

[0015] The second processing module is used for determining at least one second candidate object matching the target object from the object pool based on the feature information of the target object in response to determining that the target object is a seed object, and establishing an association relationship pair from the target object to the second candidate object.

[0016] In addition, an embodiment of the present disclosure further provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned object matching method when executing the program.

[0017] An embodiment of the present disclosure further provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the above-mentioned object matching method.

[0018] An embodiment of the present disclosure further provides a computer program product, including computer program instructions, which, when executed on a computer, enable the computer to execute the object matching method.

[0019] The object matching method and related equipment described in the embodiments of the present disclosure can instantly determine at least one object that matches the feature information of the target object at a low cost after the attribute information of the target object is updated, and establish an association relationship pair between the target object and the object it matches, thereby improving the timeliness and accuracy of searching for the same object in a massive object pool.

[0020] Furthermore, the object matching method described in the embodiment of the present disclosure can support real-time same-item object query capabilities based on reverse recall, and the implementation mode of same-item object search can be converted from the original fixed-period full-volume same-item query calculation to a positive recall and reverse recall and similarity calculation process that uses the target object's feature information as the query condition, which is triggered only when the target object's attribute information changes. Since full calculation is not required, the overall calculation amount is expected to be reduced by 50%, and the system throughput will be increased by 1 time. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the present disclosure or related technologies, the drawings required for use in the embodiments or related technical descriptions are briefly introduced below. Obviously, the drawings described below are only embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0022] Figure 1 The implementation process of the object matching method described in some embodiments of the present disclosure is shown.

[0023] Figure 2 The implementation process of the method for determining at least one first candidate object matching the target object from a seed object pool based on feature information of the target object described in some embodiments of the present disclosure is shown.

[0024] Figure 3 The flowchart of the processing method described in some embodiments of the present disclosure when the target object is a non-seed object is shown.

[0025] Figure 4 The specific method flow of determining at least one second candidate object matching the target object from the object pool based on the feature information of the target object described in some embodiments of the present disclosure is shown.

[0026] Figure 5 A schematic diagram of the internal structure of the object matching device described in some embodiments of the present disclosure is shown.

[0027] Figure 6 A more specific schematic diagram of the hardware structure of an electronic device described in some embodiments of the present disclosure is shown. DETAILED DESCRIPTION

[0028] In order to make the objectives, technical solutions and advantages of the present disclosure more clearly understood, the present disclosure is further described in detail below in combination with specific embodiments and with reference to the accompanying drawings.

[0029] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present disclosure should be understood by people with ordinary skills in the field to which the present disclosure belongs. The "first", "second" and similar words used in the embodiments of the present disclosure do not indicate any order, quantity or importance, but are only used to distinguish different components. "Including" or "comprising" and similar words mean that the elements or objects appearing before the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0030] It is understandable that before using the technical solutions of each embodiment of the present disclosure, the type, scope of use, usage scenarios, etc. of the personal information involved will be informed to the user in an appropriate manner, and the user's authorization will be obtained.

[0031] For example, in response to receiving an active request from a user, a prompt message is sent to the user to clearly remind the user that the operation requested to be performed will require obtaining and using the user's personal information. Thus, the user can independently choose whether to provide personal information to software or hardware such as an electronic device, application, server, or storage medium that performs the operation of the technical solution of the present disclosure according to the prompt message.

[0032] As an optional but non-limiting implementation, in response to receiving the user's active request, the prompt information may be sent to the user in the form of a pop-up window, in which the prompt information may be presented in text form. In addition, the pop-up window may also carry a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0033] It is understandable that the above notification and the process of obtaining user authorization are merely illustrative and do not constitute a limitation on the implementation of the present disclosure. Other methods that meet relevant laws and regulations may also be applied to the implementation of the present disclosure.

[0034] In the embodiments of the present disclosure, the following concepts are defined for the convenience of description:

[0035] 1) Seed object, which can be used to represent the product to be found.

[0036] 2) Seed object pool, which can be used to represent the product pool containing all seed objects, usually with a size of millions.

[0037] 3) Object pool, which can be used to represent a product pool containing all products, usually in the billions. In the embodiment of the present disclosure, searching for the same object can specifically refer to searching for the same object of a seed object in the seed object pool in the object pool.

[0038] 4) Pot object: the object in the object pot is called the pot object.

[0039] The seed object pool and the object pool may have a certain intersection or may not have an intersection. That is, in some embodiments, all or part of the seed objects in the seed object pool are also included in the object pool. In other embodiments, all the seed objects in the seed object pool are not in the object pool.

[0040] 4) An association pair, also referred to as a same-item pair, can be used to represent two objects on the left and right that have a same-item association relationship. The same-item relationship it embodies is specifically that the right-side object is the same-item object as the left-side object. It can be understood that the above-mentioned association pair or same-item object pair usually contains a sequential relationship for the objects involved, that is, it is one-way. For example, for the association pair AB, it only means that the right-side object B is the same-item object as the left-side object A. However, the above-mentioned association pair AB does not limit whether the left-side object A is the same-item object as the right-side object B. For example, a piece of clothing a is marked as made of pure cotton, it may be the same item as a piece of clothing b that does not indicate the material, but clothing b is not the same item as clothing a, that is, an association pair ba can be established between product a and product b, but an association pair ab cannot be established.

[0041] 5) Attribute information, which can be used to represent the description of the corresponding object, and can include one or more items of the image and text information corresponding to the object. The text information can include one or more items of information such as a title and a detailed description.

[0042] 6) The feature information of the object can be used to represent the features associated with the object calculated using the feature algorithm model according to the attribute information of the object. It can be understood that by processing the feature information of different objects, it can be determined whether different objects are the same object.

[0043] As mentioned above, how to quickly find the same object that matches a certain object among a large number of objects has become a problem that needs to be solved in many e-commerce applications.

[0044] The existing method for searching for the same object generally includes two processes: feature calculation and feature matching. Among them, the feature calculation process generally refers to batch calculation of feature information of objects in an offline manner. The above-mentioned batch calculation includes calculating feature information of all objects in the object pool, and calculating feature information of all seed objects in the seed object pool. The above-mentioned feature matching process generally refers to traversing the seed objects in the seed object pool based on a fixed time period, and for each seed object, searching for one or more related objects matching the above seed object in the object pool, thereby establishing at least one association relationship pair from the above seed object to the related object, and saving the established association relationship pair in the system. Among them, the process of searching for one or more related objects matching the above seed object in the object pool generally includes: recall, fusion, and sorting processes including rough sorting and fine sorting. In the case where the association relationship pair between objects is established, if it is necessary to find the same object of a certain seed object, the saved association relationship pair can be directly searched from the system to determine one or more objects matching the above seed object, that is, the same object of the above seed object.

[0045] It can be seen that in the above method, since the characteristics of the object and the association relationship pairs between the objects are pre-established and stored in the system, the above method can be applied to application scenarios with high timeliness requirements when the attribute information of the object is not updated. However, it can be understood that if the attribute information of an object in the object pool or seed object pool changes, the association relationship pairs associated with the object will usually change. In this case, since the association relationship pairs between objects are periodically established and stored in the system, before the moment when the association relationship pairs are updated, when searching for the same object based on the stored association relationship pairs, the results of the search for the same object may be inaccurate, thereby affecting the performance of the e-commerce business.

[0046] In order to solve the above problems, an embodiment of the present disclosure provides an object matching method, which can determine at least one object that matches the feature information of the target object immediately at a low cost after the attribute information of the target object is updated, and complete the update of the association relationship pairs in real time, thereby ensuring the timeliness and accuracy of searching for the same object in a massive object pool while updating the attribute information of certain objects.

[0047] Figure 1 The implementation process of the object matching method described in the embodiment of the present disclosure is shown. Figure 1 As shown, the above object matching method may include:

[0048] In step 110 , feature information of the target object is determined based on the attribute information of the target object.

[0049] In step 120, it is determined whether the target object is a seed object or a non-seed object.

[0050] In step 130 , in response to determining that the target object is a non-seed object, at least one first candidate object matching the target object is determined from a seed object pool based on feature information of the target object.

[0051] In step 140, an association relationship pair is established from the first candidate object to the target object.

[0052] In step 150 , in response to determining that the target object is a seed object, at least one second candidate object matching the target object is determined from an object pool based on feature information of the target object.

[0053] In step 160, an association relationship pair is established from the target object to the second candidate object.

[0054] The specific implementation method of each step in the above object matching method will be further described in detail below with reference to specific examples.

[0055] As mentioned above, the attribute information of the target object may include one or more of the information such as the image and text information corresponding to the target object. Therefore, in the above step 110, feature calculation may be performed based on one or more of the information such as the image and text information corresponding to the target object using a feature algorithm model to determine the feature information of the target object. It should be noted that the embodiments of the present disclosure do not limit the specific algorithm of the feature algorithm model.

[0056] It can be understood that, usually, when matching identical objects, the feature information of the two objects is compared. Therefore, the existing identical object search method usually includes an offline batch feature calculation process of calculating the feature information of all objects in the object pool and calculating the feature information of all seed objects in the seed object pool. In the embodiment of the present disclosure, when the attribute information of the target object is updated, the feature calculation of the target object can be performed using the same feature algorithm model as the offline batch feature calculation.

[0057] With respect to the above step 120, in an embodiment of the present disclosure, it is possible to determine whether the above target object is a seed object or a non-seed object by searching the above target object in a seed object pool. Specifically, if the above target object can be found in the seed object pool, it can be determined that the above target object is a seed object; otherwise, it can be determined that the above target object is a non-seed object. If the above target object is a newly added seed object, it is usually necessary to first add the above target object to the above seed object pool before continuing to perform subsequent steps.

[0058] With respect to the above step 130, in the embodiment of the present disclosure, the specific process of determining at least one first candidate object matching the above target object from the seed object pool based on the feature information of the above target object can be referred to in Figure 2 , which may include:

[0059] In step 210, the characteristic information of the target object is used as a query condition to perform object recall in the seed object pool to obtain at least one first seed object associated with the target object.

[0060] In step 220 , the similarity between the at least one first seed object and the target object is determined respectively.

[0061] In step 230 , the at least one first candidate object is determined from the at least one first seed object based on the similarity between the at least one first seed object and the target object.

[0062] Through the above Figure 2 It can be seen from the method shown that in the object matching method described in the embodiment of the present disclosure, on the premise that the target object is determined to be a non-seed object, it is possible to first try to find at least one seed object by performing a reverse recall in the seed object pool, so that the above-mentioned target object is a candidate object of the same object as the found seed object. In the embodiment of the present disclosure, the reason why the above-mentioned process is called reverse recall is that, usually, for the search of the same object, the feature information of the seed object is used as the query condition to recall in the object pool, and then other objects that are the same object as the seed object are finally determined through subsequent fusion, filtering, sorting and other steps. In the above step 210, the feature information of the target object as a non-seed object is used as the query condition to recall in the seed object pool, which is similar to a reverse operation of the existing operation. Therefore, it is called reverse recall in this embodiment to distinguish it.

[0063] Specifically, in order to perform the above reverse recall, it is necessary to first establish a vector index library corresponding to the feature information of each seed object in the seed object pool. Then, in the above step 210, the feature information of the above target object is used as a query condition to perform object recall in the above vector index library, thereby obtaining a reverse recall result including at least one first seed object. Usually, the reverse recall result includes at least one first seed object and its feature information.

[0064] In addition, usually, in the existing common object search application, in order to carry out the subsequent sorting process, to finally determine the same object, usually based on the query conditions and the feature information contained in the recall result, a feature information pair is established, and then the established feature information pair is input into the similarity algorithm model to determine the similarity of the above feature information pair, and then the respective recall result pairs are sorted based on the above feature information similarity, so as to determine the final association relationship pair (i.e., the same object pair). As mentioned above, the association relationship pair or the same object pair described in the embodiment of the present disclosure is usually a one-way relationship pair containing a sequential relationship for the objects involved. For example, for the association relationship pair AB, it only indicates that the right object B is the same object of the left object A. However, the above association relationship pair AB does not limit whether the left object A is the same object of the right object B. Therefore, if according to the existing method, a feature information pair is established directly based on the query conditions and the feature information contained in the recall result, the obtained feature information pair will be a feature information pair of the feature information of the target object-the feature information of the first seed object. This is opposite to the order of the left and right objects in the association relationship pair of the seed object-target object that actually needs to be established. Therefore, in the above step 220, if the existing method for searching for the same object pair is used, it is necessary to perform a reverse operation on the feature information in the established feature information pair after establishing the feature information pair between the query condition and the feature information contained in the recall result, that is, to reverse the feature information pair of the feature information of the target object-the feature information of the first seed object into the feature information pair of the feature information of the first seed object-the feature information of the target object. Then, next, the reversed feature information pair is input into the similarity algorithm model to determine the similarity of the relationship pairs of each first seed object-target object contained in the reverse recall result, that is, to determine the similarity of the at least one first seed object to the target object. It can be understood that, since there is a sequential relationship between the objects in the association relationship pair, if the above reversal operation is not performed, and the feature information pair established based on the query condition and the feature information contained in the recall result is directly input into the similarity algorithm model according to the conventional method, and the same object pair is determined by the output similarity, it usually leads to inaccurate final determined same object pair.

[0065] As an alternative to the above solution, the existing method can be abandoned and the feature information pairs inputted into the similarity algorithm model can be directly established in the order of feature information of the first seed object-feature information of the target object. In this way, no additional inversion operation is required.

[0066] Next, in the above step 230, the similarity between the above at least one first seed object and the above target object can be sorted first, and the above at least one first candidate object can be determined from the above at least one first seed object based on the sorting result. It should be noted that the embodiments of the present disclosure do not limit the specific implementation of the above sorting process. For example, the above sorting process can also include multiple steps such as rough sorting and fine sorting in a specific application.

[0067] In some other embodiments, before executing the sorting process in step 230, the at least one first seed object may be further filtered based on a preset filtering condition to obtain at least one filtered first seed object. Then, the step 230 is executed again to determine the similarity between the at least one first seed object and the target object.

[0068] With respect to the above step 140, after the above at least one first candidate object is determined, an association relationship pair from the above first candidate object to the above target object is established.

[0069] In an embodiment of the present disclosure, the established association relationship pair from the above-mentioned first candidate object to the above-mentioned target object can be stored in the system. In this way, when it is necessary to determine the same object of a certain seed object in the future, the established association relationship pair can be directly queried, thereby directly determining the same object of the seed object, and ensuring the timeliness of the same object query service. In addition, the above method can also ensure that when the attribute information of an object in any object pool is updated, the association relationship pair associated with the object can be updated in real time, thereby ensuring the accuracy of the same object query service.

[0070] Further, as a further supplement to the above technical solution, in response to determining that the target object is a non-seed object, the above method may further include: Figure 3 Multiple steps shown.

[0071] In step 310, at least one second seed object that has an association relationship pair with the target object is searched in the established association relationship pairs.

[0072] As mentioned above, there is a one-way order relationship between objects in the association relationship pair, so the association relationship pair found is usually in the order of: second seed object-target object. However, since the attribute information of the target object has changed, whether the target object is still the same object as the second seed object in the current situation needs further verification.

[0073] In step 320 , the similarity between the at least one second seed object and the target object is determined respectively.

[0074] In an embodiment of the present disclosure, at least one feature information pair can be established from the feature information of the at least one second seed object to the feature information of the target object, and the established at least one feature information pair can be input into a similarity algorithm model to determine the similarity between the at least one second seed object and the target object. The specific method can refer to the above method, and will not be repeated here.

[0075] In step 330 , at least one third candidate object matching the target object is determined from the at least one second seed object based on the similarity between the at least one second seed object and the target object.

[0076] Similarly, determining at least one third candidate object matching the target object from the at least one second seed object based on the similarity between the at least one second seed object and the target object may also be achieved through the aforementioned sorting process, which is not repeated here.

[0077] In step 340, an association relationship pair is established from the third candidate object to the target object.

[0078] In this way, when the attribute information of the target object is updated, the established association relationship pair related to the target object can be updated in real time based on the updated attribute information, thereby ensuring the accuracy of the established association relationship pair. Through such an operation, not only the real-time performance of the same object search service can be guaranteed, but also the accuracy of the same object search service can be guaranteed.

[0079] With the above Figure 2 The difference between the method shown in the figure is that in the above step 150, the specific method of determining at least one second candidate object matching the above target object from the object pool based on the feature information of the above target object can be referred to Figure 4 , mainly including the following steps.

[0080] In step 410, the feature information of the target object is used as a query condition to perform object recall in the object pool to obtain at least one pool object.

[0081] Corresponding to the reverse recall described in the aforementioned process, in the embodiment of the present disclosure, the object recall process described in the above step 410 may be referred to as a forward recall. Similarly, in order to perform the above forward recall, it is also necessary to first establish a vector index library corresponding to the feature information of each pool object in the object pool. Then, in the above step 410, the feature information of the above target object may be used as a query condition to perform object recall in the above vector index library, thereby obtaining a forward recall result containing at least one second seed object. Typically, the forward recall result will include the second seed object and its feature information.

[0082] In step 420, similarities between the target object and the at least one pot object are determined respectively.

[0083] In the embodiment of the present disclosure, in the above step 420, feature information pairs of the above target object feature information and the above pot object feature information can be established respectively, and then the established feature information pairs are input into the similarity algorithm model to determine the similarity of the above feature information pairs, that is, to determine the similarity between the above target object and the above at least one pot object. It can be seen that in the above process, since a positive recall is performed, there is no need to perform a reverse operation on the established feature information pairs.

[0084] In step 430, the at least one second candidate object is determined from the at least one pot object based on the similarity between the target object and the at least one pot object.

[0085] Specifically, the step 430 of determining the at least one second candidate object from the at least one pot object based on the similarity between the target object and the at least one pot object may include: sorting the at least one pot object based on the similarity between the target object and the at least one pot object; and determining the at least one second candidate object from the at least one pot object based on the sorting result. Further, before sorting, the at least one pot object recalled may be filtered based on a preset filtering condition, and then the filtered at least one pot object may be sorted.

[0086] Through the above method, at least one second candidate object matching the above target object can be determined from the object pool.

[0087] Next, in step 160, an association relationship pair from the target object to the second candidate object may be established.

[0088] In an embodiment of the present disclosure, the established association relationship pair from the target object to the second candidate object can be stored in the system. In this way, when it is necessary to determine the same object as a seed object in the future, the established association relationship pair can be directly queried, thereby directly determining the same object as the seed object, and ensuring the timeliness of the same object query service. In addition, the above method can also ensure that when the attribute information of any seed object is updated, the association relationship pair associated with the seed object can be updated in real time, thereby ensuring the accuracy of the same object query service.

[0089] Furthermore, the object matching method described in the embodiment of the present disclosure can support real-time same-item object query capabilities based on reverse recall, and the implementation mode of same-item object search can be converted from the original fixed-period full-volume same-item query calculation to a positive recall and reverse recall and similarity calculation process that uses the target object's feature information as the query condition, which is triggered only when the target object's attribute information changes. Since full calculation is not required, the overall calculation amount is expected to be reduced by 50%, and the system throughput will be increased by 1 time.

[0090] Based on the above object matching method, an embodiment of the present disclosure further discloses an object matching device. Figure 5 FIG. 1 shows a schematic diagram of the internal structure of the object matching device according to an embodiment of the present disclosure. Figure 5 As shown, the object matching device may include:

[0091] A feature determination module 510 is used to determine feature information of a target object based on attribute information of the target object;

[0092] A classification module 520, for determining whether the target object is a seed object;

[0093] A first processing module 530 is configured to, in response to determining that the target object is a non-seed object, determine at least one first candidate object matching the target object from a seed object pool based on feature information of the target object, and establish an association relationship pair from the first candidate object to the target object; and

[0094] The second processing module 540 is used to determine at least one second candidate object matching the target object from the object pool based on the feature information of the target object in response to determining that the target object is a seed object, and establish an association relationship pair from the target object to the second candidate object.

[0095] Based on the same inventive concept, corresponding to any of the above-mentioned embodiments and methods, the present disclosure also provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the object matching method described in any of the above embodiments is implemented.

[0096] Figure 6 A more specific hardware structure diagram of an electronic device provided in this embodiment is shown, and the device may include: a processor 2010, a memory 2020, an input / output interface 2030, a communication interface 2040, and a bus 2050. The processor 2010, the memory 2020, the input / output interface 2030, and the communication interface 2040 are connected to each other in communication within the device through the bus 2050.

[0097] The processor 2010 can be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0098] The memory 2020 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 2020 can store an operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program codes are stored in the memory 2020 and are called and executed by the processor 2010.

[0099] The input / output interface 2030 is used to connect input / output devices to realize information input and output. The input / output devices can be configured in the device as components, or can be externally connected to the device to provide corresponding functions. The input devices can include microphones, various sensors, etc., and the output devices can include displays, speakers, vibrators, indicator lights, etc.

[0100] The communication interface 2040 is used to connect a communication module (not shown) to realize communication interaction between the device and other devices. The communication module can realize communication through a wired mode (such as USB, network cable, etc.) or a wireless mode (such as mobile network, WIFI, Bluetooth, etc.).

[0101] The bus 2050 includes a path that transmits information between the various components of the device (eg, the processor 2010, the memory 2020, the input / output interface 2030, and the communication interface 2040).

[0102] It should be noted that, although the above device only shows the processor 2010, the memory 2020, the input / output interface 2030, the communication interface 2040, and the bus 2050, in the specific implementation process, the device may also include other components necessary for normal operation. In addition, it can be understood by those skilled in the art that the above device may also only include the components necessary for implementing the embodiments of the present specification, and does not necessarily include all the components shown in the figure.

[0103] The electronic device of the above embodiment is used to implement the corresponding object matching method in any of the above embodiments, and has the beneficial effects of the corresponding method embodiment, which will not be described in detail here.

[0104] Based on the same inventive concept, corresponding to any of the above-mentioned embodiment methods, the present disclosure also provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the object matching method described in any of the above embodiments.

[0105] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device.

[0106] The computer instructions stored in the storage medium of the above embodiments are used to enable the computer to execute the task processing method described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0107] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present disclosure (including the claims) is limited to these examples. Based on the concept of the present disclosure, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the embodiments of the present disclosure as described above, which are not provided in detail for the sake of simplicity.

[0108] In addition, to simplify the description and discussion, and in order not to make the embodiments of the present disclosure difficult to understand, the known power / ground connections to the integrated circuit (IC) chips and other components may or may not be shown in the provided figures. In addition, the device can be shown in the form of a block diagram to avoid making the embodiments of the present disclosure difficult to understand, and this also takes into account the fact that the details of the implementation of these block diagram devices are highly dependent on the platform on which the embodiments of the present disclosure will be implemented (that is, these details should be fully within the scope of understanding of those skilled in the art). Where specific details (e.g., circuits) are set forth to describe exemplary embodiments of the present disclosure, it is apparent to those skilled in the art that the embodiments of the present disclosure can be implemented without these specific details or with changes in these specific details. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0109] Although the present disclosure has been described in conjunction with specific embodiments of the present disclosure, many replacements, modifications and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may use the embodiments discussed.

[0110] The embodiments of the present disclosure are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present disclosure should be included in the scope of protection of the present disclosure.

Claims

1. An object matching method, comprising: Determine feature information of the target object based on attribute information of the target object; In response to determining that the target object is a non-seed object, determining at least one first candidate object matching the target object from a seed object pool based on feature information of the target object, and establishing an association relationship pair from the first candidate object to the target object; as well as In response to determining that the target object is a seed object, at least one second candidate object matching the target object is determined from an object pool based on feature information of the target object, and an association relationship pair from the target object to the second candidate object is established; wherein, The seed object is the product to be searched for the same product; The seed object pool is a product pool containing all seed objects; and The object pool is a product pool including all products.

2. The method according to claim 1, wherein: Determining at least one first candidate object matching the target object from a seed object pool based on the feature information of the target object includes: Using the characteristic information of the target object as a query condition, recalling objects in the seed object pool to obtain at least one first seed object; respectively determining a similarity between the at least one first seed object and the target object; and The at least one first candidate object is determined from the at least one first seed object based on a similarity between the at least one first seed object and the target object.

3. The method according to claim 2, wherein: Determining the similarity between the at least one first seed object and the target object respectively includes: filtering the at least one first seed object based on a preset filtering condition to obtain at least one filtered first seed object; and The similarity between the at least one filtered first seed object and the target object is determined respectively.

4. The method according to claim 2, wherein: Determining the similarity between the at least one first seed object and the target object respectively includes: Establishing at least one feature information pair from the query condition to feature information of at least one object in the recall result respectively; Performing an inversion operation on the feature information in the at least one established feature information pair to obtain at least one inverted feature information pair; and The inverted at least one feature information pair is input into a similarity algorithm model to determine the similarity between the at least one first seed object and the target object.

5. The method according to claim 2, wherein: Determining at least one first candidate object matching the target object from the seed object pool based on the feature information of the target object further includes: Searching for at least one second seed object having an association relationship with the target object in the established association relationship pairs; respectively determining a similarity between the at least one second seed object and the target object; Determining at least one third candidate object matching the target object from the at least one second seed object based on the similarity between the at least one second seed object and the target object; and An association relationship pair is established from the third candidate object to the target object.

6. The method according to claim 1, wherein: Determining at least one second candidate object matching the target object from the object pool based on the feature information of the target object includes: Using the characteristic information of the target object as a query condition, recalling objects in the object pool to obtain at least one pool object; respectively determining a similarity between the target object and the at least one pot object; and The at least one second candidate object is determined from the at least one pot object based on a similarity between the target object and the at least one pot object.

7. The method according to claim 6, wherein: Determining the similarity between the target object and the at least one pot object respectively includes: Establishing at least one feature information pair from the query condition to feature information of at least one object in the recall result respectively; and The at least one feature information pair is input into a similarity algorithm model to determine a similarity between the target object and the at least one pot object.

8. An object matching device, comprising: A feature determination module, used to determine feature information of a target object based on attribute information of the target object; A classification module, used to determine whether the target object is a seed object; A first processing module, configured to, in response to determining that the target object is a non-seed object, determine at least one first candidate object matching the target object from a seed object pool based on feature information of the target object, and establish an association relationship pair from the first candidate object to the target object; as well as A second processing module is configured to, in response to determining that the target object is a seed object, determine at least one second candidate object matching the target object from the object pool based on the feature information of the target object, and establish an association relationship pair from the target object to the second candidate object; wherein, The seed object is the product to be searched for the same product; The seed object pool is a product pool containing all seed objects; and The object pool is a product pool including all products.

9. An electronic device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the object matching method according to any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable a computer to execute the object matching method according to any one of claims 1 to 7.

11. A computer program product, comprising computer program instructions, which, when executed on a computer, enable the computer to execute the object matching method according to any one of claims 1 to 7.

Images