Object processing method, interface display method, computing device, storage medium and program product

Abstract

The embodiment of the invention provides an object processing method, an interface display method, computing equipment, a storage medium and a program product. The method comprises the steps of obtaining target multimedia data provided by a user, and performing object search based on the target multimedia data to determine at least one first object; providing a search result of the at least one first object for the user; providing search prompt information for the user; acquiring search information provided by the user based on the search prompt information; identifying at least one target object attribute hit by the search information; performing object search in combination with the at least one target object attribute and the at least one first object to determine at least one second object; and providing a search result of the at least one second object to the user. According to the technical scheme provided by the embodiment of the invention, the picture searching function is enriched, the user experience is improved, and the object conversion rate can be increased.

Description

Technical Field

[0001] Embodiments of the present application relate to the field of electronic technologies, and in particular, to an object processing method, an interface display method, a computing device, a computer-readable storage medium, and a computer program product. Background Art

[0002] With the rapid development of Internet technologies, in some online processing systems that provide objects for users to interact with, an object search function is usually provided to facilitate users to quickly find objects, so that further operations can be performed on the objects to complete object conversion. Users can provide keywords or pictures containing the target object, etc., so that object recall can be performed based on the keywords or pictures.

[0003] In the prior art, the search results obtained by object recall are displayed on the search result page for users to view through the user terminal and further find the objects of interest. However, currently, the search function is still relatively single, the user experience is not good, and the object conversion rate is not high. Summary of the Invention

[0004] Multiple aspects of the present application provide an object processing method, an interface display method, a computing device, a storage medium, and a program product to solve the technical problem of the single picture search function and poor user experience in the prior art.

[0005] In a first aspect, an object processing method is provided in an embodiment of the present application, including:

[0006] Obtain target multimedia data provided by a user, and perform object search based on the target multimedia data to determine at least one first object;

[0007] Provide the user with search results of the at least one first object;

[0008] Provide the user with search prompt information;

[0009] Obtain search information provided by the user based on the search prompt information;

[0010] Identify at least one target object attribute hit by the search information;

[0011] Combine the at least one target object attribute and the at least one first object to perform object search to determine at least one second object;

[0012] Provide the user with search results of the at least one second object.

[0013] Optionally, it further includes:

[0014] Determine a set identifier corresponding to the at least one object and a first set of attributes of the set identifier index; wherein, the set identifier is used to identify a target object set to which the at least one object belongs in the multiple sets; the multiple object sets are obtained by grouping multiple objects based on object features according to feature similarity.

[0015] Determine a second set of attributes composed of the attributes respectively possessed by the at least one first object.

[0016] Merge the first set of attributes and the second attributes to obtain a target set of attributes.

[0017] The identifying at least one target object attribute hit by the search information includes:

[0018] Identify at least one target object attribute hit by the search information in the target set of attributes.

[0019] Optionally, the combining the at least one target object attribute and the at least one first object for object search to determine at least one second object includes:

[0020] Divide the object attributes in the second set of attributes into a third set of attributes that meet the same model condition and a fourth set of attributes that do not meet the same model condition according to whether the feature similarity between the corresponding first object and the target multimedia data meets the same model condition.

[0021] After screening out the third set of attributes from the first set of attributes, merge it with the fourth set of attributes to obtain a fifth set of attributes.

[0022] If the at least one target object attribute is in the third set of attributes, search for at least one second object having the at least one target object attribute from the multiple objects according to a first recall quantity.

[0023] If any target object attribute is in the fifth set of attributes, search for at least one second object having the at least one target object attribute from the multiple objects according to a second recall quantity; wherein, the second recall quantity is greater than the first recall quantity.

[0024] Optionally, it further includes:

[0025] Determine the target object set to which the at least one first object belongs in multiple object sets; the multiple object sets are obtained by grouping multiple objects based on object features according to feature similarity.

[0026] Determine a target set of attributes composed of the object attributes respectively possessed by different objects in the target object set and the at least one object.

[0027] The identification of at least one target object attribute hit by the search information includes:

[0028] Identifying at least one target object attribute hit by the search information in the target attribute set.

[0029] Optionally, it further includes:

[0030] Determining the number of groupings of multiple object groupings divided in the target object set; the multiple object groupings are obtained by dividing the objects in the target object set based on object characteristics;

[0031] Determining the second recall quantity according to the number of groupings.

[0032] Optionally, the combining the at least one target object attribute and the at least one first object for object search to determine at least one second object includes:

[0033] Judging whether the at least one target object attribute meets the screening requirements;

[0034] If so, searching for at least one second object having the at least one target object attribute from the multiple objects according to the first recall quantity;

[0035] If not, searching for at least one second object having the at least one target object attribute from the multiple objects according to the second recall quantity; wherein, the second recall quantity is greater than the first recall quantity.

[0036] Optionally, the identification of at least one target object attribute hit by the search information includes:

[0037] Based on the search information, identifying the target intent operation;

[0038] In the case where the target intent operation is an object update operation, identifying at least one target object attribute hit by the search information.

[0039] Optionally, after the object search based on the target multimedia data to determine at least one first object, the method further includes:

[0040] Generating a plurality of recommended questions based on the target multimedia data and the object information of the at least one first object;

[0041] The providing of search prompt information to the user includes;

[0042] Providing the user with search prompt information including the plurality of recommended questions and input prompt information;

[0043] Obtaining the search information provided by the user based on the search prompt information includes:

[0044] Obtaining the target recommended question selected by the user from the multiple recommended questions to use the target recommended question as the search information, or obtaining the search information input by the user for the input prompt information.

[0045] Optionally, generating multiple recommended questions based on the target multimedia data and the object information of the at least one first object includes:

[0046] Generating at least one recommended question corresponding to different intent operations based on the target multimedia data and the object information of the at least one first object;

[0047] The method further includes:

[0048] Determining a target category corresponding to the target multimedia data based on the at least one first object;

[0049] Determining the arrangement order of the multiple recommended questions according to the intent priority of different intent operations corresponding to the target category; the multiple recommended questions are displayed in the arrangement order.

[0050] Optionally, identifying at least one target object attribute hit by the search information includes:

[0051] Using a first large model to identify at least one target object attribute hit by the search information;

[0052] The method further includes:

[0053] Using the first large model to generate a first prompt copy according to the search information and the recognition result of the at least one target object attribute;

[0054] Providing the first prompt copy to the user.

[0055] Optionally, it further includes:

[0056] If no target object attribute corresponding to the search information is recognized, performing an object search based on the target multimedia data and the search information to obtain at least one third object;

[0057] Providing the search result of the at least one third object to the user.

[0058] Optionally, it further includes:

[0059] If there exists the at least one third object, generate a second prompt text and provide the second prompt text to the user; otherwise, generate a third prompt text and provide the third prompt text to the user.

[0060] Optionally, it further includes:

[0061] In the case where the target intent is an evaluation viewing operation, generate evaluation information corresponding to the at least one first object from the evaluation data corresponding to the at least one first object respectively;

[0062] Provide the evaluation information to the user.

[0063] Optionally, the evaluation information includes the target evaluation data corresponding to the at least one first object respectively, and the providing the evaluation information corresponding to the at least one first object to the user includes:

[0064] Provide an evaluation result page to the user, and display the target evaluation data corresponding to the at least one first object respectively in the display order of the at least one first object in the evaluation result page.

[0065] Optionally, the generating the evaluation information corresponding to the at least one first object from the evaluation data corresponding to the at least one first object respectively includes:

[0066] For any one of the first objects, screen out the target evaluation data that meets the evaluation requirements from the evaluation data of the first object; and / or,

[0067] For any one of the first objects, screen out the target evaluation data that meets the evaluation requirements from the evaluation data of the first object, and perform an aggregation process on the target evaluation data corresponding to the at least one first object respectively to obtain comprehensive evaluation data.

[0068] Optionally, it further includes:

[0069] In the case where the target intent operation is a consultation operation, use a fourth large model to generate a response content based on the search information and the object information of the at least one first object;

[0070] Provide the response content to the user.

[0071] Optionally, it further includes:

[0072] In the case where the target intent is a matching operation, identify the target style or target accessories corresponding to the at least one first object;

[0073] Perform an object search based on the target multimedia data and the target style or the target accessories to determine at least one fourth object;

[0074] Provide search results of the at least one fourth object to the user.

[0075] Optionally, after providing the search results of the at least one first object to the user, the method further includes:

[0076] When it is detected that all the first objects in the at least one first object that meet the same model condition as the target multimedia data are exposed, identify the target style corresponding to the first object that meets the same model condition as the target multimedia data;

[0077] Perform object search based on the target multimedia data and the target style to determine at least one fifth object;

[0078] Provide object hint information corresponding to the at least one fifth object respectively in the first search result page.

[0079] Optionally, it further includes:

[0080] When the target intent operation is an abnormal input operation, provide abnormal hint information to the user.

[0081] Optionally, it further includes:

[0082] When the target intent operation is a co-search operation, perform object search based on the target multimedia data and the search information to obtain at least one sixth object;

[0083] Provide search results of the at least one sixth object to the user.

[0084] Optionally, the providing search hint information to the user includes:

[0085] When a predetermined number of first objects are exposed, provide search hint information to the user.

[0086] In a second aspect, an object processing method is provided, including:

[0087] Obtain target multimedia data provided by a user, and perform object search based on the target multimedia data to determine at least one first object;

[0088] Provide search results corresponding to the at least one first object to the user;

[0089] Provide search hint information to the user;

[0090] Obtain search information provided by the user based on the search hint information;

[0091] Execute a processing operation by combining the search information and the at least one first object to obtain a processing result;

[0092] Provide the processing result to the user.

[0093] Optionally, the performing corresponding processing operations by combining the search information and the at least one first object to obtain a processing result includes:

[0094] Identify the target intent operation corresponding to the search information;

[0095] Perform corresponding processing operations by combining the search information and the at least one first object according to the processing manner corresponding to the target intent operation to obtain a processing result.

[0096] In a third aspect, an interface display method is provided, including:

[0097] Display search results of at least one first object in a user interface; the at least one first object is determined by performing object search based on target multimedia data provided by a user;

[0098] Display search prompt information in the user interface;

[0099] Obtain search information provided by the user for the search prompt information; the search information is used to perform a processing operation by combining the at least one first object to obtain a processing result;

[0100] Display the processing result in the user interface.

[0101] Optionally, the displaying search prompt information in the user interface includes:

[0102] When a sliding operation triggered in the user interface exposes a predetermined number of first objects in the first search result page, display search prompt information in the user interface.

[0103] Alternatively, display a prompt object in the user interface; the displaying search prompt information in the user interface in response to a target operation triggered in the user interface includes:

[0104] When a trigger operation for the prompt object is responded to, display search prompt information in the user interface.

[0105] Optionally, it further includes:

[0106] Display a plurality of recommended questions in the user interface;

[0107] The obtaining search information provided by the user for the search prompt information includes:

[0108] In response to the user's selection operation for the multiple recommended questions, use the selected target recommended question as the search information;

[0109] Alternatively, in response to the user's input operation for the search prompt information, determine the search information input by the user.

[0110] Optionally, the processing result includes search results of at least one second object; the search information is used to determine at least one target object attribute it hits; the at least one target object attribute is used to perform object search in combination with the target multimedia data to determine at least one second object;

[0111] The displaying the processing result in the user interface includes:

[0112] Display the search results of the at least one second object in the user interface.

[0113] Optionally, the displaying the search results of the at least one first object in the user interface includes:

[0114] Display a first search result page in the user interface, and display the search results of the at least one first object on the first search result page.

[0115] The displaying the search results of the at least one second object in the user interface includes:

[0116] Display a second search result page in the user interface, and cover at least part of the first search result page;

[0117] Display the search results of the at least one second object on the second search result page.

[0118] Optionally, it further includes:

[0119] Display a first prompt text on the second search result page; the first prompt text is generated according to the search information and the recognition result of the at least one target object attribute.

[0120] Optionally, it further includes:

[0121] Obtain the search results of at least one third object sent by the server; the at least one third object is obtained by performing object search based on the target multimedia data and the search information when no target object attribute corresponding to the search information is recognized;

[0122] Display the search results of the at least one third object in the user interface.

[0123] Optionally, displaying the search results of the at least one third object in the user interface includes:

[0124] Displaying a third search result page in the user interface and displaying the search results of the at least one first object on the search result page;

[0125] The method further includes:

[0126] Displaying a second prompt copy or a third prompt copy on the third search result page.

[0127] Optionally, it further includes:

[0128] Obtaining the search results of at least one fifth object sent by the server; the at least one fifth object is determined by identifying the target style corresponding to the first object that meets the same model condition as the target multimedia data when all the first objects that meet the same model condition as the target multimedia data in the at least one first object are exposed, and performing object search determination based on the target multimedia data and the target style;

[0129] Displaying the search results of the at least one fifth object in the user interface.

[0130] Optionally, the processing result includes evaluation information, and the search information is used to identify the corresponding target intent operation; the evaluation information is determined based on the evaluation data corresponding to the at least one first object when the target intent operation is an evaluation viewing operation;

[0131] Displaying the processing result in the user interface includes:

[0132] Displaying the evaluation information in the user interface.

[0133] Optionally, the processing result includes the search results of at least one fourth object, and the search information is used to identify the corresponding target intent operation; the at least one fourth object is obtained by performing object search based on the target multimedia data and the target style or target accessories of the at least one first object when the target intent operation is a matching operation;

[0134] Displaying the processing result in the user interface includes:

[0135] Displaying the search results of the at least one fourth object in the user interface.

[0136] Optionally, the processing result includes a response content; the search information is used to identify the corresponding target intent operation; the response content is generated based on the search information and the object information of the at least one first object when the target intent operation is a consultation operation.

[0137] The displaying of the processing result in the user interface includes:

[0138] Displaying the response content in the user interface.

[0139] Optionally, the processing result includes an exception prompt message; the search information is used to identify the corresponding target intent operation; the exception prompt message is generated when the target intent operation is an abnormal input operation;

[0140] The displaying of the processing result in the user interface includes:

[0141] Displaying the exception prompt message in the user interface.

[0142] Optionally, the processing result includes search results of at least one sixth object; the search information is used to identify the corresponding target intent operation; the at least one sixth object is determined by performing object search based on the target multimedia data and the search information when the target intent operation is a co-search operation;

[0143] The displaying of the processing result in the user interface includes:

[0144] Displaying the search results of the at least one sixth object in the user interface.

[0145] In a fourth aspect, a computing device is provided, including a processing component and a storage component;

[0146] The storage component stores a computer program; the computer program is used to be called and executed by the processing component to implement the object processing method as described in the first aspect above or the object processing method as described in the second aspect above or the interface display method as described in the third aspect above.

[0147] In a fifth aspect, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processing component, it is used to implement the object processing method as described in the first aspect above or the object processing method as described in the second aspect above or the interface display method as described in the third aspect above.

[0148] In a sixth aspect, a computer program product is provided, including a computer program or instruction. When the computer program or instruction is executed by a processing component, it is used to implement the object processing method as described in the first aspect above or the object processing method as described in the second aspect above or the interface display method as described in the third aspect above.

[0149] In the embodiments of the present application, when at least one first object is searched based on the target multimedia data provided by the user and the search results of the at least one first object are provided to the user, search hint information can be provided to the user to support the user in performing a further operation and freely inputting search information. Then, based on the search information and in combination with the at least one first object, a further processing operation can be performed to obtain a processing result, and the processing result can be provided to the user to assist the user in making a decision or further screening, etc., thereby enriching the picture search function, improving the user experience, and helping to increase the object conversion rate. For example, the user can request object update through the search information, and thus the objects recalled by the target multimedia data can be screened or rewritten in combination with at least one target object attribute to obtain at least one second object again and provide the search results of the at least one second object to the user. By supporting object update, the picture search function is enriched, the user experience is improved, and more accurate information can be provided to the user, which can stimulate the user to perform a further object conversion operation, thereby helping to increase the object conversion rate and making full use of the computing resources to ensure the resource usage effect.

[0150] These aspects or other aspects of the present application will be more clearly understood in the following description of the embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0151] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments and descriptions thereof of the present application are used to explain the present application and do not constitute an improper limitation of the present application. In the drawings:

[0152] Figure 1 The flowchart of an embodiment of an object processing method provided by the present application is shown;

[0153] Figure 2 The flowchart of another embodiment of an object processing method provided by the present application is shown;

[0154] Figure 3 The flowchart of another embodiment of an object processing method provided by the present application is shown;

[0155] Figure 4 The flowchart of an embodiment of an interface display method provided by the present application is shown;

[0156] Figures 5a to 5g The interface schematic diagrams provided by the embodiments of the present application are respectively shown;

[0157] Figure 6 The schematic diagram of scenario interaction provided by the embodiments of the present application is shown;

[0158] Figure 7Shows a schematic structural diagram of an embodiment of an object processing device provided by the present application;

[0159] Figure 8 Shows a schematic structural diagram of another embodiment of an object processing device provided by the present application;

[0160] Figure 9 Shows a schematic structural diagram of another embodiment of an interface display device provided by the present application;

[0161] Figure 10 Shows a schematic structural diagram of an embodiment of a computing device provided by the present application. Detailed implementation manners

[0162] To make the objectives, technical solutions, and advantages of the present application clearer, the technical solutions of the present application will be clearly and completely described below in conjunction with specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present application.

[0163] It should be noted that in the case where the embodiments of the present application involve user information, the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the embodiments of the present application are all information and data authorized by the user or fully authorized by all parties. Additionally, the collection, use, and processing of relevant data need to comply with relevant laws, regulations, and standards of relevant countries and regions, and corresponding operation entrances are provided for users to select authorization or rejection. In addition, various models (including but not limited to language models or large models) involved in the present application comply with relevant laws and standards.

[0164] In addition, it should be noted that in the case where the embodiments of the present application involve user interaction operations or trigger operations, the user interaction operations or trigger operations involved in the embodiments of the present application include but are not limited to: interaction operations in various ways such as touch operations, gesture operations, voice operations, head movement operations, and eye movement operations; among them, touch operations include but are not limited to: click operations, double-click operations, long-press operations, swipe operations, pinch operations, or mouse hover operations, etc. Swipe operations include but are not limited to: straight-line swipes, curved swipes, etc.

[0165] Furthermore, the technical solution of the embodiment of the present application is applicable to a network virtual environment. The users described generally refer to "virtual users". Real users can register user accounts on the server through the registration method to obtain user identities in the network environment. In the embodiment of the present application, the same user account can be used to log in to the server through different types of user terminals, so that the server can identify the same user. The interaction operations between the server and the user are implemented based on the user account. The corresponding data received or sent by the server to the user is also implemented based on the user account. Actually, the user terminal corresponding to the user account receives or sends the corresponding data to the server. In addition, communication between users can also be achieved through user accounts. Among them, the user can refer to an individual or an organization, such as an enterprise, etc. The present application does not make specific restrictions on this.

[0166] The technical solution of the embodiment of the present application can be applicable to application scenarios that support search, such as object search scenarios, etc. The object can be, for example, a commodity, an article, or a video, etc. The object can be provided by an online processing system. The online processing system can be a processing system that supports object publishing and object consumption, and can provide a search function, supporting multimedia information such as keywords and pictures for searching, etc. The online processing system can be understood as a data processing center, which can perform corresponding data processing for object providers, users, etc. The object provider can publish objects in the online processing system, and the user, as a consumer, can consume the objects launched in the online processing system. In a practical application, for example, the online processing system can be an e-commerce system, and the object can refer to a commodity. The commodity can correspond to one or more products offline, and these products can be tangible objects or intangible services, etc.

[0167] The inventor found in the process of implementing the present application that currently, the search results are displayed on the search result page for users to view through the user terminal and find the objects they are interested in. However, the current search function is still relatively single, and the user experience is not good, resulting in a reduction in the user's interest in performing further object conversion operations, such as click operations, etc., resulting in a low object conversion rate.

[0168] To solve this technical problem, the inventor proposed the technical solution of this application through a series of studies. The embodiments of this application provide an implementation method that supports a user to input search information for search results, and further processes the search information and at least one first object obtained by the search. On the premise of providing the user with search results of at least one first object, it is also possible to provide the user with further processing results in combination with the search information provided by the user, thereby enriching the search function, rather than simply displaying search results, improving the user experience. When the user experience is improved, it will encourage the user to perform an object conversion operation, which helps to improve the object conversion rate. In one implementation, the user can update the object in the search results through the search information. By supporting object update, not only is the search function enriched and the user experience improved, but also more accurate information can be provided to the user to encourage the user to perform further object conversion operations, which helps to improve the object conversion rate. And the computing resources are also fully utilized. The consumption of computing resources can improve the object conversion rate, thus ensuring the resource usage effect.

[0169] In the e-commerce scenario, it will effectively improve the user's shopping experience, thereby increasing the commodity purchase rate, etc.

[0170] Next, the technical solutions in the embodiments of this application will be clearly and completely described with reference to the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of this application.

[0171] Figure 1 It is a flowchart of an embodiment of an object processing method provided by this application. The technical solution of this embodiment can be executed by a server, and the server can refer to the server of an online processing system.

[0172] In practical applications, the online processing system can include a server and a user terminal. A connection can be established between the user terminal and the server through a network, and the network provides a medium for the communication link between the user terminal and the server. The network can include various connection types, such as wired, wireless communication links, or fiber optic cables, etc.

[0173] The user terminal can be oriented to consumers for users to perform interactive behaviors such as object search, browsing, and purchasing. In the e-commerce scenario, the object is a commodity.

[0174] The user terminal can interact with the server through the network to receive or send messages, etc. The user terminal can sense the interactive behaviors performed by the user and send corresponding interactive requests to the server, and the server can process the interactive requests and feedback the processing results to the user terminal, etc.

[0175] Among them, the user side can be a browser, an APP (Application), or a web application such as an H5 (HyperText Markup Language 5) application, or a light application (also known as a mini-program, a lightweight application) or a cloud application, etc. The first user side or the second user side can be deployed in an electronic device and needs to rely on the device or certain apps in the device to run, etc. The electronic device can, for example, have a display screen and support information browsing, etc., such as a personal mobile terminal such as a mobile phone, a tablet computer, a personal computer, a desktop computer, a smart speaker, a smart watch, and so on.

[0176] The server side can include servers that provide various services, such as a server that supports model platform training, and a server side that processes information sent by the user side, etc.

[0177] It should be noted that the server side can be implemented as a distributed server cluster composed of multiple servers, or can be implemented as a single server. The server can also be a server of a distributed system, or a server combined with a blockchain. The server can also be 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, Content Delivery Network (CDN), and big data and artificial intelligence platforms, or an intelligent cloud computing server or an intelligent cloud host with artificial intelligence technology, etc. This application does not limit this.

[0178] Among them, Figure 1 The object processing method shown can include the following steps:

[0179] 101: Obtain the target multimedia data provided by the user, and perform object search based on the target multimedia data to determine at least one first object.

[0180] The user can provide the target multimedia data through the user side. Optionally, in the user interface of the user side, such as the home page, etc., a media search control can be displayed, and the user side can obtain the target multimedia data in response to a trigger operation on the media search control.

[0181] Among them, the target multimedia data can refer to any multimedia data provided by the user. In practical applications, the target multimedia data can, for example, include target keywords, target pictures, target videos, and / or target audios, etc.

[0182] When the target multimedia data includes the target keyword, the media search control may include a keyword input box. For the trigger operation on the keyword input box, the keyboard can be invoked to facilitate the user to input the target keyword.

[0183] When the target multimedia data includes the target picture or the target video, the media search control may include picture search prompt information. In response to the trigger operation on the picture search prompt information, the user terminal can obtain the target picture or the target video from the local picture library according to the user request, or can also obtain the target picture or the target video by collecting the target object. The target picture or the target video may include the target object. The target object may be an offline item, and the first object may be a virtual item provided for the online processing system, etc.

[0184] When the target multimedia data includes the target audio, the media search control may include an audio collection control. In response to the trigger operation on the audio collection control, the user terminal can collect the target audio and upload it to the server, and the server can convert it into the target keyword by means of speech recognition, so that the search can be carried out based on the target keyword.

[0185] Of course, the user terminal can also collect the target audio, convert it into the target keyword and then send it to the server, so that the server can carry out the search based on the target keyword.

[0186] Among them, in an actual application, the target multimedia data may refer to the target picture. The technical solution of the embodiment of the present application can be specifically applied to the picture search scenario. Picture search, as an intuitive and efficient search method, has become one of the current mainstream search methods. Picture search can recall the same type of objects as the target object in the target picture, thus greatly improving the search experience. For the convenience of description, in one or more embodiments below, the target multimedia data is taken as an example of the target picture for introduction. It can be understood that the present application is not limited thereto.

[0187] After the server obtains the target multimedia data, it can perform object search to recall at least one first object. Optionally, at least one first object can be recalled according to the feature similarity between the multimedia data and different objects. If the feature similarity meets certain conditions such as being greater than a certain threshold, etc., it can be recalled as the first object.

[0188] When the target multimedia data is the target keyword, the feature similarity can be calculated based on the text features corresponding to the target keyword and the text features corresponding to the text description information of different objects. Of course, at least one first object can also be recalled by means of text matching.

[0189] When the target multimedia data is a target picture or a target video, the feature similarity can be calculated based on image features. For example, by calculating the feature similarity between the image features of the target picture and the object pictures of different objects. Among them, the image features of the target picture can be extracted from the target picture, and the present application does not limit the extraction method of image features. Another example is for the target video. By calculating the feature similarity between the image features of each video frame in the target video and the image features of an object picture of a certain object respectively, and then performing weighted processing on the feature similarities corresponding to multiple video frames to obtain the total feature similarity, object recall can be performed based on the total feature similarity. For example, objects that meet certain conditions such as being greater than a certain threshold can be selected as the first objects.

[0190] It can be understood that feature data such as text features and image features are usually in vector form, and the calculation of feature similarity can be implemented using cosine similarity, Euclidean distance, etc. The present application does not limit this.

[0191] In practical applications, there are usually requirements for the recall quantity when performing object search to ensure the user experience, etc. Therefore, the at least one first object can specifically be multiple first objects, and the multiple first objects can be determined by combining the recall quantity and the feature similarity.

[0192] 102: Provide the search results of at least one first object to the user.

[0193] Optionally, it can be to provide the user with a first search result page and provide the search results of at least one first object in the first search result page.

[0194] The search results of the at least one first object can include the object hint information of each first object. The object hint information of each first object can have corresponding slots, that is, display positions, in the first search result page. The object hint information of the at least one first object can be arranged in descending order of feature similarity in the first search result page.

[0195] The object hint information can be obtained by extracting key information from the object detail information. For example, it can include an object picture, and the object picture can be a main picture. In addition, it can also include information such as the object price and object name. The object name can be specifically extracted from the object title, etc. The present application does not limit the specific implementation of the object hint information, and it can be implemented in a traditional way.

[0196] The above steps 101 to 102 describe the specific implementation method of the picture search operation.

[0197] 103: Provide search hint information to the user.

[0198] The search prompt information can be displayed in the user interface of the user terminal. Optionally, the user terminal can display the search prompt information at the boundary position of the user interface, for example, at the bottom boundary or the top boundary. The search prompt information can cover part of the first search result page. Optionally, the display operation of the search prompt information and the paging operation of the first search result page can be independent of each other. Of course, in order to achieve different display effects, the search prompt information can also be displayed in the user interface when a predetermined number of first objects are exposed on the first search result page, or can be terminated when the display duration of the first search result page exceeds a certain duration or the exposure number of the first objects exceeds a certain number, so as to achieve the purpose of effectively prompting the user without disturbing the user.

[0199] In response to a target request triggered by the user, provide search prompt information to the user; the target request can be generated by the user terminal in response to the user's target operation.

[0200] The target operation can be triggered by the user for the user interface of the user terminal, such as a sliding operation, etc. Optionally, it can specifically be triggered for the first search result page displayed in the user interface. Of course, while the first search result page is being displayed in the user interface, a prompt object can also be displayed. The prompt object can be superimposed on the first search result page, can cover part of the first search result page, and can be displayed in the form of a floating layer or a pop-up window, etc. The present application does not limit this.

[0201] Among them, limited by the interface size of the user interface, not all search results on the first search result page can be exposed in the user interface. Through paging operations, such as sliding paging, double-click paging, long-press paging, or button paging, etc., different page contents of the first search result page can be switched and displayed in the user interface. The paging operation of the first search result page and the display operation of the prompt object can be independent of each other, and the target operation can be a triggering operation for the prompt object. The prompt object can have a control function to sense the user's operation. The prompt object can be implemented as a floating button or a control with a virtual image, etc. The present application does not limit this.

[0202] Optionally, providing search prompt information to the user can include: providing search prompt information when a predetermined number of first objects are exposed.

[0203] In the embodiments of the present application, exposure can refer to being displayed in the user interface for the user to view, etc.

[0204] When a predetermined number of first objects are exposed on the first search result page, search prompt information can be provided to the user. When the user performs a sliding operation on the first search result page, different first objects can be exposed in the user interface. If a predetermined number of first objects are exposed, it indicates that there is no object that the user likes in the search results. At this time, search prompt information can be provided to the user again to support further operations of the user, etc.

[0205] Of course, it can also be that when the number of page turns of the first search result page reaches a specified number, search prompt information is provided to the user. The user performs a sliding operation on the first search result page, and different first objects are exposed in the user interface by means of page turning. If the number of page turns reaches the specified number, and at this time a certain number of first objects are exposed, it can also be considered that there is no object that the user likes in the search results. At this time, search prompt information can be provided to the user again to support further operations of the user, etc. Among them, the first search result page can be divided into multiple sub-pages according to the interface size, and the number of page turns can refer to the number of exposed sub-pages.

[0206] 104: Obtain the search information provided by the user based on the search prompt information.

[0207] The search prompt information can be used to prompt the user to provide search information. For example, the search prompt information can include input prompt information, which can be implemented as an input prompt box to sense the user's input operation. When the user terminal displays the search prompt information, the trigger operation for the input prompt information can bring up the virtual keyboard to implement the user's input operation and obtain the search information.

[0208] The search information can be text freely input by the user. Of course, in order to ensure that the search information is related to image search, recommended questions generated based on at least one first object can also be provided to the user. When the user terminal displays the search prompt information, one or more recommended questions can also be displayed, and the search information can also be the target recommended question selected by the user.

[0209] 105: Combine the search information and at least one first object to perform a processing operation to obtain a processing result.

[0210] 106: Provide the processing result to the user.

[0211] The processing result can be sent to the user terminal, and the user terminal displays the processing result in the user interface. Optionally, the user terminal can display a target page in the user interface and can display the processing result in the target page. Among them, the target page can cover at least part of the first search result page. Among them, in response to the closing operation of the target page, the user terminal can continue to display the first search result page.

[0212] In this embodiment, for the search results obtained based on the target multimedia data, it is possible to support the user to freely input search information, so that further processing can be carried out in combination with the search information and at least one first object obtained by the search. On the premise of providing the user with the search results of at least one first object, the relevant processing results can also be provided to the user in combination with the search information provided by the user, thus enriching the search function. It is not just a single display of search results, improving the user experience. When the user experience is improved, it will encourage the user to perform the object conversion operation, thus helping to improve the object conversion rate.

[0213] In one embodiment, object update can be achieved through search information, such as Figure 2 shown, which is a flowchart of another embodiment of an object processing method provided by the present application. The technical solution of this embodiment can be executed by the server, and the method can include the following steps:

[0214] 201: Obtain the target multimedia data provided by the user, and perform object search based on the target multimedia data to determine at least one first object.

[0215] The target multimedia data can refer to target keywords, target pictures, target videos, target audios, etc. In a practical application, it can refer to target pictures.

[0216] 202: Provide the user with the search results of at least one first object.

[0217] For the specific operation implementations of steps 201 to 202, reference can be made to the relevant descriptions in steps 101 to 102, and details will not be repeated here.

[0218] 203: Provide the user with search prompt information.

[0219] 204: Obtain the search information provided by the user based on the search prompt information.

[0220] For the specific operation implementations of steps 201 to 204, reference can be made to the relevant descriptions in steps 101 to 104, and details will not be repeated here.

[0221] 205: Identify at least one target object attribute hit by the search information.

[0222] Object attributes can refer to the inherent characteristics of an object itself. Each object attribute can be composed of an attribute type and its corresponding attribute value. The attribute type can include, for example, physical attribute types such as color (the corresponding attribute value is, for example, red), shape (the object attribute value is, for example, square), material (the corresponding attribute value is, for example, cowhide), weight (the corresponding attribute value is, for example, 10 kilograms), or functional attribute types such as battery life (the corresponding attribute value is, for example, 12 hours), pixel (the corresponding attribute value is, for example, 200 million pixels), or design attribute types such as style (the corresponding attribute value is, for example, simple), etc. This application does not limit this.

[0223] For the convenience of control and management, not all object attributes are allowed to be modified in actual applications. Therefore, at least one target object attribute that matches the search information can be identified from the attribute set, and this attribute set can be preset. Of course, in actual applications, the attributes supported for filtering or rewriting by different categories of objects are also different. For example, the target object in the target picture is a red dress, and the red dress belongs to the dress category. The dress category supports color modification but does not support shape modification, etc. Therefore, the attribute set corresponding to different categories can be preset, so that at least one target object attribute can be obtained by identifying from the attribute set corresponding to the target category to which at least one first object recalled from the target picture belongs.

[0224] Among them, at least one target object attribute that matches the search information can be implemented by text matching. Of course, with the development of AI (Artificial Intelligence) technology, the embodiments of this application can support users to provide search information in the form of natural language description. As other implementation methods, at least one target object attribute that matches the search information can also be identified by means of an AI large model. Therefore, at least one target object attribute that matches the search information can be identified by using the first large model.

[0225] The AI large models involved in this article, including the first large model and the second large model, third large model, fourth large model, etc. that may appear in the corresponding embodiments below, refer to large-parameter models trained using large-scale data and powerful computing capabilities, which are machine learning models with complex structures and can process massive amounts of data and complete various complex tasks, such as natural language processing, computer vision, speech recognition, etc. They can include large language models (LLMs) or multimodal large models (MLMs), etc. The AI large models involved in this article can be pre-trained models and are retrained through model fine-tuning to adapt to different processing tasks. This not only utilizes the powerful capabilities of the pre-trained models but also can adapt to new data distributions. Therefore, it can ensure the generalization ability of the models and reduce the overfitting phenomenon.

[0226] Optionally, the first large model can be obtained by fine-tuning the model based on the sample information and the sample attributes corresponding to the sample information. The sample information is used as the model input, and the sample attributes are used as training labels, and training can be completed with only a small amount of training data.

[0227] 206: Combine at least one target object attribute and at least one first object for object search to determine at least one second object.

[0228] In practical applications, there are usually requirements for the recall quantity when performing object search to ensure the user experience, etc. Therefore, the at least one second object can specifically be multiple second objects.

[0229] 207: Provide the user with search results of at least one second object.

[0230] In the embodiments of the present application, object search can be performed again to recall at least one second object again. By combining at least one target object attribute for object search, the purpose of attribute screening or attribute rewriting can be achieved. For example, if the target object in the target picture is a red chair, the at least one first object includes a red chair. If the target object attribute includes blue, then blue chairs can be recalled again; another example is that if the target object in the target picture is a leather sofa, the at least one first object includes a leather sofa. If the target object attribute includes top-layer cowhide, then sofas made of top-layer cowhide can be re-screened and obtained.

[0231] Among them, providing the user with search results of at least one second object can be to provide the user with a second search result page and provide search results of at least one second object in the second search result page. The above-mentioned target page can also be implemented as the second search result page.

[0232] The search results of the at least one second object can include object hint information of each second object. The object hint information of each second object can have corresponding slots, that is, display positions, in the second search result page. The object hint information of the at least one second object can be arranged in descending order of feature similarity in the second search result page. When the target multimedia data is a target picture, the feature similarity can be determined based on the image features of the object picture of the second object and the image features of the target picture.

[0233] The object hint information can be obtained by extracting key information from the object detail information. For example, it can include an object picture, and the object picture can be a main picture. In addition, it can also include information such as the object price and object name. The object name can be specifically extracted from the object title, etc. The present application does not limit the specific implementation of the object hint information.

[0234] Among them, the search results of the at least one second object can be sent to the user terminal, and the user terminal can display the search results of the at least one second object in the user interface. Optionally, the user terminal can display a second search result page in the user interface and display the search results of the at least one second object in the second search result page. The second search result page can cover at least part of the first search result page, etc. Among them, in response to a close operation on the second search result page, the user terminal can continue to display the first search result page.

[0235] In this embodiment, the user can provide the target object attributes through the search information, so as to support attribute filtering or rewriting to achieve object update, enrich the picture search function, improve the user experience, and can provide more accurate information to the user, which can encourage the user to perform further object conversion operations, thus helping to improve the object conversion rate. In the e-commerce scenario, it will improve the user's shopping experience and thus increase the commodity purchase rate, etc.

[0236] In order to improve the calculation efficiency and reduce the consumption of computing resources, in some embodiments, the method may further include:

[0237] Determine the target object set to which at least one first object belongs in multiple object sets; wherein, the multiple object sets are obtained by grouping multiple objects according to feature similarity based on object features; determine the target attribute set composed of the object attributes respectively possessed by different objects in the target object set and the at least one first object;

[0238] Then, the above-mentioned at least one target object attribute identified by the search information hit may include: identifying at least one target object attribute hit by the search information in the target attribute set.

[0239] Optionally, it may be to use a first large model to identify at least one target object attribute hit by the search information in the target attribute set. Optionally, for the convenience of the large model to identify, the attributes in the target attribute set can be subjected to corresponding format conversions, etc. to be converted into a format that the large model can understand, etc. This application does not limit this.

[0240] Among them, the object features can refer to image features, which can be extracted from the object pictures of each object, etc. Of course, the object features can also be text features extracted from the object details, or obtained by fusing text features and image features, etc. In the picture search scenario, the object features can refer to image features. The multiple objects can refer to all objects that support picture search, and these multiple objects constitute the object pool that supports picture search. The multiple objects can be grouped in advance according to the feature similarity to form multiple object sets. Of course, they can also be grouped according to the feature similarity and combined with the object attributes to form multiple object sets, so that each object included in each object set has similar features and / or the same object attributes, etc. Among them, for example, objects with a feature similarity greater than the first similarity threshold can be divided into one group, or objects with a feature similarity greater than the first similarity threshold and having multiple identical object attributes can be divided into one group.

[0241] In the specific implementation process, a clustering algorithm can be used to group the multiple objects based on the first similarity threshold. Each object set can be a cluster obtained by clustering. The specific implementation of the clustering algorithm in this application is not limited.

[0242] The object attributes of each object included in the target object set are combined with the object attributes of each first object, and then the target attribute set can be formed. Optionally, the object attributes that are included in each object of the target object set and support modification, as well as the object attributes that all the multiple first objects have and support modification, can form the target attribute set. Among them, whether the object attribute supports modification can be determined in advance according to the category to which the object belongs, etc.

[0243] Since the target object set is composed of objects with similar features, the number of objects in the target object set can be controlled by setting a reasonable similarity threshold. In the embodiments of this application, only at least one target object attribute that is hit by the search information is identified from the target attribute set, which can reduce the identification range and improve the identification efficiency on the premise of ensuring accuracy.

[0244] In some embodiments, in order to improve the calculation efficiency and ensure the accuracy of the recalled objects, determining the target object set to which at least one first object belongs in the multiple object sets can include: determining the target object set to which the first object with the greatest feature similarity to the target multimedia data in at least one first object belongs;

[0245] When the target multimedia data is a target picture, it can be to determine the target object set to which the first object with the greatest feature similarity between the object picture and the target picture in at least one first object belongs.

[0246] In addition, to further improve the computing efficiency and the recall efficiency, after dividing multiple objects into multiple object sets in advance, a set identifier of the object set to which each object belongs can be set for each object, and an attribute set composed of the attributes of each object in each object set can be determined in advance, and an index relationship can be established with the set identifier, so that the first attribute set can be quickly determined directly according to the set identifier and the index relationship. Therefore, in some embodiments, the method may further include:

[0247] Determine the set identifier corresponding to at least one object and the first attribute set indexed by the set identifier; wherein, the set identifier is used to identify the target object set to which at least one object belongs in multiple sets; the multiple object sets are obtained by grouping multiple objects according to feature similarity based on object features;

[0248] After that, the attributes respectively possessed by at least one first object can be merged with the first attribute set to obtain the target attribute set.

[0249] Optionally, it may be to first determine the second attribute set composed of the attributes respectively possessed by at least one first object, and then merge the first attribute set and the second attribute to obtain the target attribute set.

[0250] Thereby, at least one target object attribute hit by the search information in the target attribute set can be identified.

[0251] And the above determination of the set identifier corresponding to at least one object may be: determining the set identifier configured by the first object with the highest feature similarity to the target multimedia data among at least one first object.

[0252] Among them, there are multiple implementation manners for the determination manner of the at least one second object:

[0253] In an optional implementation manner, the above object search by combining object attributes and at least one first object to determine at least one second object may include:

[0254] Dividing at least one first object into a first object set whose feature similarity to the target multimedia data meets the same model condition and a second object set whose feature similarity does not meet the same model condition; determining the third attribute set corresponding to the first object set and the fourth attribute set corresponding to the second object set;

[0255] After screening out the third attribute set from the first attribute set, merge it with the fourth attribute set to obtain the fifth attribute set;

[0256] If at least one target object attribute is in the third attribute set, search for at least one second object having at least one target object attribute from multiple objects according to the first recall quantity;

[0257] If any of the target object attributes is in the fifth attribute set, at least one second object with at least one target object attribute is searched and obtained from multiple objects according to the second recall quantity; wherein, the second recall quantity is greater than the first recall quantity.

[0258] When the target multimedia data is a target picture, it may be to divide at least one first object into a first object set in which the feature similarity between the object picture and the target picture meets the same model condition and a second object set that does not meet the same model condition.

[0259] For the sake of easy understanding, assume that the first attribute set is A, the second attribute set is B, the third attribute set is b1, and the fourth attribute set is b2. Then the target attribute set is A + B, the second attribute set B = b1 + b2, and the fifth attribute set is A + b2 - b1.

[0260] In practical applications, in order to ensure the recall effect, the recall quantity is usually set. When searching for objects according to the feature similarity, it may make at least one first object include the same model objects and similar objects of the target objects in the target picture. The same model objects can be determined according to the feature similarity meeting the same model condition, and the similar objects can be determined according to the feature similarity not meeting the same model condition but meeting the similar condition. The same model condition can be, for example, that the feature similarity is greater than the first threshold, and the similar condition can be, for example, that the feature similarity is greater than the second threshold, where the first threshold is greater than the second threshold. The same model objects can refer to the first objects with the feature similarity greater than the first threshold, and the similar objects can refer to the first objects with the feature similarity greater than the second threshold.

[0261] Combined with the foregoing description, it can be understood that the first object set is also the same model objects corresponding to the target picture.

[0262] Thus, according to the division of the first object set and the second object set, the third attribute set and the fourth attribute set can be obtained. The third attribute set is also composed of the object attributes possessed by each object in the first object set, and the fourth attribute set is also composed of the object attributes possessed by each object in the second object set.

[0263] Before re - conducting the object search, it can first be determined whether at least one target object attribute is in the third attribute set. If all are in the third attribute set, it indicates that the user's intention is to perform attribute filtering on the search results of at least one first object. At this time, the search and attribute filtering can be carried out from multiple objects according to the first recall quantity, and at least one second object with the at least one target object attribute can be obtained therefrom. If any one of the target object attributes is in the fifth attribute set, it indicates that the user's intention is to perform attribute rewriting on the search results of at least one first object. At this time, according to the second recall quantity, at least one second object with the at least one target object attribute can be searched and obtained from multiple objects.

[0264] Among them, the first recall quantity, that is, the recall quantity corresponding to the first object, can be to screen the first recall quantity of second objects with the at least one target object attribute from multiple objects in the order of decreasing feature similarity.

[0265] When the user's intention is to perform attribute rewriting on the search results of at least one first object, in order to ensure that at least one second object with at least one target object attribute can be recalled, the recall scope can be expanded, that is, the search scope among multiple objects is expanded, and at least one second object with at least one target object attribute is recalled from multiple objects according to the second recall quantity. Optionally, it can be to screen the second recall quantity of second objects with the at least one target object attribute from multiple objects in the order of decreasing feature similarity.

[0266] In this optional implementation manner, at least one second object can be quickly recalled through the attribute filtering method, which can improve the search efficiency.

[0267] In addition, the second attribute can also be directly divided according to whether the same - model condition is met, without having to perform the above - mentioned re - division of the object set. Therefore, as another optional method, the above - mentioned object search in combination with the at least one target object attribute and the at least one first object to determine at least one second object can include:

[0268] Dividing the object attributes in the second attribute set into a third attribute set that meets the same - model condition and a fourth attribute set that does not meet the same - model condition according to whether the feature similarity between the corresponding first object and the target multimedia data meets the same - model condition;

[0269] After screening out the first attribute set from the first attribute set, it is merged with the fourth attribute set to obtain a fifth attribute set;

[0270] If the at least one target object attribute is in the third attribute set, at least one second object with the at least one target object attribute is searched and obtained from the multiple objects according to the first recall quantity;

[0271] If any of the target object attributes is in the fifth attribute set, at least one second object having at least one of the target object attributes is retrieved from the multiple objects according to the second retrieval quantity; wherein, the second retrieval quantity is greater than the first retrieval quantity.

[0272] When the target multimedia data is a target picture, it may also be that the object attributes in the second attribute set are divided into a third attribute set that meets the same model condition and a fourth attribute set that does not meet the same model condition according to whether the object picture corresponding to the first object and the target picture have the same feature similarity.

[0273] In another alternative implementation manner, it may also be that when at least one target object attribute is in the third attribute set, at least one second object having at least one of the target object attributes is screened from at least one first object set; if any of the target object attributes is in the fifth attribute set, at least one second object having at least one of the target object attributes is screened from the target object set and the at least one first object.

[0274] In this alternative implementation manner, at least one second object can be quickly retrieved through the attribute filtering method, and it is not necessary to re-compare the feature similarity, thereby improving the search efficiency.

[0275] In practical applications, the first retrieval quantity may refer to the retrieval quantity corresponding to the first object, and the second retrieval quantity is a value greater than the first retrieval quantity, which can be preset. Of course, in order to ensure the retrieval accuracy, in some embodiments, the method may further include:

[0276] Determine the number of groups of the multiple object groups divided in the target object set; the multiple object groups are obtained by dividing the objects in the target object set based on object features;

[0277] Determine the second retrieval quantity according to the number of groups.

[0278] Wherein, the number of groups can be used as the second retrieval quantity. In addition, since the number of groups may be too large, resulting in an increase in the search workload, in order to ensure the search efficiency, when the number of groups is less than the predetermined quantity, the number of groups can be used as the second retrieval quantity, and when the number of groups is greater than the predetermined quantity, the predetermined quantity can be used as the second retrieval quantity.

[0279] Of course, the second retrieval quantity corresponding to different numbers of groups can also be preset, or a calculation formula for determining the corresponding second retrieval quantity based on the number of groups can be preset, etc.

[0280] In this embodiment, based on object features, objects in the object set can be further divided according to feature similarity to obtain multiple object groups. Thus, the second recall quantity can be determined according to the number of groups corresponding to the target object set, and then the search quantity of second objects with at least one target object attribute can be filtered according to the second recall quantity.

[0281] By determining the second recall quantity based on the number of groups, a more accurate recall quantity of the second objects can be determined, thus ensuring the recall effect to determine the second objects that can recall at least one target object attribute.

[0282] As can be seen from the foregoing description, the object set can be divided based on the first similarity threshold. Optionally, the object groups can be divided based on the second similarity threshold, so that the feature similarity between different objects in the same object group can be greater than the second similarity threshold, where the second similarity threshold is greater than the first similarity threshold, thus achieving the purpose of further grouping the target objects.

[0283] In another alternative implementation, when the target multimedia data is a target picture, the above object search by combining at least one target object attribute and at least one first object to determine at least one second object may include:

[0284] Combining at least one target object attribute, updating the image features extracted from the target picture, and performing object search based on the updated image features to determine at least one second object.

[0285] At least one target object attribute can be fused into the updated image features, so that at least one of the recalled second objects has the at least one target object attribute.

[0286] In another alternative implementation, when the target multimedia data is a target picture, the above object search by combining at least one target object attribute and at least one first object to determine at least one second object may include:

[0287] Combining at least one target object attribute, updating the target picture, and performing object search based on the image features extracted from the updated target picture to determine at least one second object.

[0288] According to at least one target object attribute, the target picture can also be updated first, that is, the target object in the target picture is updated to have the at least one target object attribute, so that the image features of the updated target picture can be extracted and object search can be performed, so that at least one of the recalled second objects has the at least one target object attribute.

[0289] In some embodiments, the object search in combination with at least one target object attribute and at least one first object to determine at least one second object may include:

[0290] Determine whether the at least one target object attribute meets the screening requirements;

[0291] If so, search for at least one second object having at least one target object attribute from multiple objects according to the first recall quantity; if not, search for at least one second object having at least one target object attribute from multiple objects according to the second recall quantity; wherein, the second recall quantity is greater than the first recall quantity.

[0292] Combined with the above various optional implementation manners, the screening requirement may be, for example, that at least one target object attribute is in the third attribute set, etc. If any target object attribute is not in the third attribute set, it can be considered that the screening requirement is not met, and the intention is to perform attribute rewriting. At this time, if any target object attribute is in the fifth attribute set, then at least one second object having at least one target object attribute can be recalled from multiple objects according to the second recall quantity.

[0293] Of course, it may also be to determine whether the at least one target object attribute meets the screening requirements; if so, screen at least one second object having at least one target object attribute from at least one first object; if not, screen at least one second object having at least one target object attribute from the target object set to which at least one first object belongs and at least one first object.

[0294] In some embodiments, the at least one target object attribute identified as a hit by the search information may include:

[0295] Based on the search information, identify the target intent operation; in the case where the target intent operation is an object update operation, identify the at least one target object attribute hit by the search information.

[0296] The technical solution of the embodiments of the present application can support multiple intent operations on the basis of search results, including not only object update operations, but also other operations, such as evaluation viewing operations, consultation operations, matching operations, or co-search operations, etc. Therefore, for the search information provided by the user, the intent can be identified first to determine the target intent. If the target intent is an object update operation, the identification operation can be performed again to determine the at least one target object attribute hit by the search information.

[0297] Optionally, an intent recognition model can be used to recognize a corresponding target intent operation based on the search information. The intent recognition model can be specifically implemented as a second large model, which can be obtained by fine-tuning the model based on sample search information and the sample intent operations corresponding to the sample search information. Among them, the sample search information serves as the model input data, and the sample intent model serves as the training label.

[0298] In order to facilitate the user to provide search information, in some embodiments, after performing object search on the target multimedia data to determine at least one first object, the method may further include: generating a plurality of recommended questions based on the target multimedia data and the object information of at least one first object;

[0299] The above-mentioned providing search prompt information to the user may include: in response to a target operation triggered by the user, outputting search prompt information including a plurality of recommended questions and input prompt information;

[0300] Then, the above-mentioned obtaining the search information provided by the user based on the search prompt information may include: obtaining a target recommended question selected by the user from the plurality of recommended questions as the search information, or obtaining the search information input by the user for the input prompt information.

[0301] Among them, the plurality of recommended questions can be preset or generated in real time based on the target multimedia data and at least one first object. As can be seen from the foregoing description, the technical solution of the embodiment of the present application can support multiple intent operations, and the recommended questions corresponding to different intent operations are different. In some embodiments, the above-mentioned generating a plurality of recommended questions based on the target multimedia data and the object information of at least one first object may include:

[0302] Generating at least one recommended question corresponding to different intent operations respectively based on the target multimedia data and the object information of at least one first object.

[0303] Since the plurality of recommended questions may correspond to different intent operations, in some embodiments, the method may further include:

[0304] Based on at least one first object, determining a target category corresponding to the target multimedia data; determining the arrangement order of the plurality of recommended questions according to the intent priorities of different intent operations corresponding to the target category; and displaying the plurality of recommended questions in the arrangement order.

[0305] That is, the intention priorities of different intention operations corresponding to different categories can be preset in advance, so that the intention priorities of different intention operations corresponding to the target category can be determined according to the target category to which at least one first object belongs. Furthermore, the arrangement order of multiple recommended questions can be determined according to the intention priorities. Among them, the recommended questions corresponding to the intention operations with higher intention priorities are arranged more forward, and the arrangement order among the recommended questions belonging to the same intention operation can be determined randomly, etc. Thus, the client can display the multiple recommended questions according to the sorting order of the multiple recommended questions, etc.

[0306] In some embodiments, it may be to use the third large model to generate at least one recommended question corresponding to different intention operations based on the target multimedia data and the object information of at least one first object.

[0307] The third large model can be trained based on the sample multimedia data, the object information of multiple sample objects matched with the sample multimedia data, and the sample questions corresponding to different intention operations. Among them, the sample multimedia data and the object information of multiple sample objects matched with the sample multimedia data are used as the model inputs of the third large model, and the sample questions corresponding to different intention operations can be used as training labels to fine-tune the third large model. When the multimedia data is a picture, the above-mentioned target multimedia data is the target picture, and the sample multimedia data is the sample picture.

[0308] As can be seen from the previous description, the first large model can be used to identify at least one target object attribute hit by the search information. To further improve the user experience, in some embodiments, the method may further include:

[0309] Generating a first prompt copy; providing the first prompt copy to the user.

[0310] Optionally, the first prompt copy can be provided on the second search result page. The client can display the second search result page in the user interface, display the first prompt copy on the second search result page, and after obtaining the search results of at least one second object, continue to display the search results of the at least one second object on the second search result page.

[0311] Optionally, the first prompt copy can be generated by using the first large model according to the search information and the recognition result of at least one target object attribute

[0312] Among them, the recognition result of at least one target object attribute may include successful recognition or failed recognition. For example, if the search information is "red" and at least one target object attribute includes: the color is red, in the case of successful recognition, the first prompt copy may be, for example, "Red products have been found for you", etc.

[0313] Of course, the server can also pre-configure the first prompt text corresponding to different recognition results.

[0314] To further improve the user experience, in some embodiments, the method may further include:

[0315] If no target object attribute corresponding to the search information is recognized, perform object search based on the target multimedia data and the search information to obtain at least one third object; provide the at least one third object to the user. When the target multimedia data is a target picture, it may be to perform object search based on the target picture and the search information to obtain at least one third object.

[0316] In practical applications, there are usually requirements for the recall quantity when performing object search to ensure the user experience, etc. Therefore, the at least one third object may specifically be multiple third objects.

[0317] Optionally, it may be to provide the user with a third search result page and provide search results of at least one third object in the third search result. The user terminal can display the third search result page on the user interface and display the search results of at least one third object on the third search result page.

[0318] In practical applications, if the search information does not hit any target object attribute, it indicates that the object attribute in the search information is an object attribute that does not support modification. To ensure the user experience, object search can be performed based on the target multimedia data and the search information.

[0319] Among them, when the target multimedia data is a target picture, the image features extracted from the target picture can be fused with the text features transformed from the search information. The image features of the object picture of each object can be fused with the text features transformed from the object information, so that object search can be performed based on the fused features according to the feature similarity, such as recalling at least one third object whose feature similarity is greater than a certain threshold. Among them, the object information can specifically be the object title, etc.

[0320] Of course, it may also be to calculate the feature similarity between the image features of the target picture and the image features of each object, and between the text features of the search information and the text features of each object respectively, then calculate the sum of the feature similarities, and recall at least one third object according to the sum of the feature similarities, etc.

[0321] In some embodiments, the method may further include:

[0322] If there is at least one third object, generate a second prompt text and provide the second prompt text to the user; otherwise, generate a third prompt text and provide the third prompt text to the user.

[0323] Optionally, the second prompt text or the third prompt text may be provided on the third search result page.

[0324] For example, the second prompt text may be "The following products are recommended for you according to the keywords", and the third prompt text may be "No products found yet. Try changing the word."

[0325] The second prompt text or the third prompt text may be obtained through pre-configuration. Of course, it may also be generated by using the first large model according to the recall results of at least one third object, etc. The recall results include successful recall or failed recall. Successful recall indicates the existence of at least one third object, and failed recall indicates the non-existence of at least one third object, etc.

[0326] As can be seen from the foregoing description, the technical solution of the embodiment of the present application can support multiple intent operations. Several possible intent operations are exemplified below. It should be noted that the present application is not limited thereto, and can be set according to the growing user needs in actual applications, etc.:

[0327] In an optional implementation manner, the method may further include:

[0328] In the case where the target intent is an evaluation view operation, generate evaluation information corresponding to at least one first object from the evaluation data corresponding to each of the at least one first object; provide the evaluation information to the user.

[0329] Optionally, the evaluation information may be displayed on the first search result page, etc.

[0330] Of course, in some embodiments, providing the evaluation information to the user may include: providing the user with an evaluation result page and displaying the evaluation information on the evaluation result page.

[0331] That is, the evaluation information may be displayed on a separate evaluation result page, and the evaluation result page is the target page.

[0332] Among them, the evaluation information may include the target evaluation data corresponding to each first object, or the comprehensive evaluation data aggregated from the evaluation data of at least one first object. Of course, it may also include the target evaluation data and the comprehensive evaluation data.

[0333] In some embodiments, when the evaluation information includes the target evaluation data corresponding to each of the at least one first object, providing the evaluation information corresponding to the at least one first object to the user may include:

[0334] Providing the user with an evaluation result page and displaying the target evaluation data of the evaluation information corresponding to the at least one first object on the evaluation result page in the display order of the at least one first object.

[0335] By displaying the evaluation information corresponding to at least one first object on the evaluation result page, the normal display of the first search result page can be unaffected, etc. When the evaluation result page is closed, the first search result page can continue to be displayed.

[0336] The object hint information of each first object in the first search result page can be linked to the object details page of the first object. Optionally, the evaluation information of each first object in the evaluation result page can also be linked to the object details page of the first object. Thus, after the user browses the evaluation information and determines the interested first object, the user can jump to the object details page in the evaluation result page without having to find the corresponding first object from the first search results.

[0337] In some embodiments, when the target intent is an evaluation viewing operation, generating the evaluation information corresponding to at least one first object from the evaluation data corresponding to each of the at least one first object may include:

[0338] When the target intent is an evaluation viewing operation, re-performing object search based on the target multimedia data to determine the evaluation data corresponding to each of the at least one first object; generating the evaluation information corresponding to at least one first object according to the evaluation data corresponding to each of the at least one first object.

[0339] Compared with the method of finding the corresponding evaluation data according to the object identifier of each first object in the first search result page, by re-performing object search and directly recalling the evaluation data corresponding to each of the at least one first object, the processing efficiency can be improved and the consumption of computing resources can be reduced.

[0340] Among them, the target evaluation data of each first object can be screened from the multiple evaluation data of each first object. Therefore, in some embodiments, generating the evaluation information corresponding to at least one first object from the evaluation data corresponding to each of the at least one first object may include: for any first object, screening the target evaluation data that meets the evaluation requirements from the evaluation data of the first object.

[0341] Optionally, if there is no target evaluation data that meets the evaluation requirements, the content data corresponding to this first object in the evaluation information can be blank. Of course, a predetermined copywriting such as "No real user evaluation has been found for this product" can also be used.

[0342] In addition, a fourth hint copywriting can also be generated according to the search information. The fourth hint copywriting can be, for example, "Real user evaluations of the same model product have been found for you", etc. The third hint copywriting can be generated using a large model, etc. Of course, a copywriting template can also be pre-configured and then combined with the keywords in the search information to generate it. Among them, the evaluation requirements can be, for example, belonging to positive evaluation data, containing at least one specific keyword, and / or containing real pictures, etc.

[0343] Of course, the evaluation information can also be generated from the positive review data of each first object by using an information generation model, etc., and the present application does not limit this.

[0344] The comprehensive evaluation data can specifically be obtained by aggregating the target evaluation data of each first object. Therefore, the method can also include: aggregating the target evaluation data corresponding to at least one first object respectively to obtain the comprehensive evaluation data.

[0345] This aggregation processing operation can be implemented by using an aggregation model, for example. The aggregation model can also be implemented as a large model. Thus, the aggregation model can, according to certain aggregation requirements, aggregate the target evaluation data corresponding to at least one first object respectively to obtain the comprehensive evaluation data. The aggregation requirements can be set in combination with the actual situation. For example, it is required that the comprehensive evaluation data includes evaluation keywords that each first object has, or includes evaluation keywords whose occurrence times are greater than a certain number, and so on.

[0346] In another alternative implementation, the method can also include:

[0347] In the case where the target intent operation is a consultation operation, generating a response content based on the search information and the object information of at least one first object; providing the response content to the user.

[0348] Among them, it can be to provide the user with a target page and provide the response content in the target page. The user terminal can display the target page in the user interface and display the response content in the target page. Of course, the search information can also be displayed in the target page to help the user understand the conversation context, etc.

[0349] Optionally, it can be to use a fourth large model to generate a response content based on the search information and the object information of at least one first object.

[0350] The fourth large model can be trained based on sample information, the object information of multiple sample objects matching the sample information, and sample response content; among them, the sample information and the object information of multiple sample objects matching the sample information can be used as model inputs, while the sample response content is used as a training label, and the training is achieved by fine-tuning the fourth large model.

[0351] In practical applications, for example, in the e-commerce scenario, what often affects the user's purchase decision also includes the purchase guide and the post-purchase guide. The purchase guide can help the user fully understand the product and ensure that the purchased product is suitable for themselves, while the post-purchase guide can help the user understand how to use and maintain the product, etc.

[0352] Therefore, in some embodiments, the consultation operation may include an operation of obtaining a purchase guide or an operation of obtaining a post-purchase guide, etc. Through intention recognition, the target intention operation can be specifically recognized as an operation of obtaining a purchase guide or an operation of obtaining a post-purchase guide.

[0353] For example, if the target picture is a picture of slippers, at least one first object is a slipper product, and the search information can be "What material of slippers is the most anti-slip", through intention recognition, it can be determined that the target intention operation is an operation of obtaining a purchase guide. Another example is that if the target picture is a picture of slippers, at least one first object is a slipper product, and the search information can be "How to clean slippers more cleanly", through intention recognition, it can be determined that the target intention operation is an operation of obtaining a post-purchase guide.

[0354] By using the training data to fine-tune the model, the above-mentioned second large model can be made to implement the operation of obtaining a purchase guide or the operation of obtaining a post-purchase guide, and then combined with the search information and the object information of at least one first object to give the corresponding response content.

[0355] Of course, the search information can be text freely input by the user. The recommended questions corresponding to the operation of obtaining a purchase guide and the operation of obtaining a post-purchase guide can help the user on how to provide appropriate text, etc.

[0356] In addition, the consultation operation may further include a chat operation. When the target intention operation is a chat operation, a guiding prompt text can be provided to the user to prompt the user to input information related to the product, etc. The guiding prompt text can be pre-configured, and of course, it can also be generated by the large model based on the search information and the recognition result of the chat operation.

[0357] In another alternative, the method may further include:

[0358] In the case where the target intention is a matching operation, identify the target style or target accessories corresponding to at least one first object; perform object search based on the target multimedia data and the target style or target accessories to determine at least one fourth object; provide the user with the search results of at least one fourth object. When the target multimedia data is the target picture, it can also be an object search based on the target picture and the target style or based on the target picture and the target accessories.

[0359] In practical applications, there are usually requirements for the recall quantity during object search to ensure the user experience, etc. Therefore, the at least one fourth object may specifically be multiple fourth objects.

[0360] The matching operation may aim to find a fourth object that meets the matching requirements with the target style of the first object, or a fourth object that can be used as the target accessory of the first object.

[0361] The matching requirement can be, for example, the same style and matching categories, etc. Of course, it can also be set according to actual needs, etc.

[0362] For example, if the first object is a sofa in Italian style, the fourth object can be a coffee table in Italian style, etc.; for example, if the first object is a mobile phone, the fourth object can be a mobile phone case as an accessory for the mobile phone, etc.

[0363] Among them, the identification of the target style or target accessory can be achieved by using the fifth large model, for example. By using the fifth large model, the corresponding target style or target accessory can be determined based on the object information of at least one first object. The fifth large model can be obtained by fine-tuning with training data; the training data can include the object information of sample objects, as well as the sample style or sample accessory corresponding to the sample objects, etc. The object information of the sample objects is used as the model input, and the sample style or sample accessory is used as the training label.

[0364] Among them, it can be to provide the user with a fourth search result page and provide search results of at least one fourth object on the fourth search result page.

[0365] The user terminal can display the fourth search result page in the user interface and display the search results of at least one fourth object on the fourth search result page. The fourth search result page is the target page.

[0366] Of course, a corresponding fifth prompt text can also be generated based on the search information to display the fifth prompt text on the fourth search result page, such as "Suitable accessories have been found for you", etc. Of course, the fifth prompt text can be generated using a large model or pre-configured, etc.

[0367] Among them, it can be generated based on the search information and the recognition result of the target style or target accessory. The recognition result can include successful recognition or failed recognition, etc.

[0368] In addition, if the fourth object is not recalled, corresponding prompt text can also be generated to guide the user to provide search information again, etc.

[0369] In actual applications, users may input randomly, resulting in some invalid information or risk control information in the search information. Therefore, in another optional implementation, the method can further include:

[0370] When the target intent operation is an abnormal input operation, providing the user with abnormal prompt information.

[0371] The abnormal prompt information can be pre-configured. Of course, it can also be generated with the help of a large model, etc.

[0372] For example, the abnormal prompt information can be "The product you are looking for may already exist in the results. Try changing the keyword."

[0373] In another alternative implementation, the method may further include:

[0374] When the target intent operation is a co-search operation, perform object search based on the target multimedia data and the search information to obtain at least one sixth object; provide the user with the search results of at least one sixth object. When the target multimedia data is a target picture, that is, perform object search based on the target picture and the search information.

[0375] In practical applications, there are usually requirements for the recall quantity during object search to ensure the user experience, etc. Therefore, the at least one sixth object may specifically be multiple sixth objects.

[0376] It may be to provide the user with a sixth search result page and provide the search results of at least one sixth object in the sixth search result page.

[0377] Among them, the search method of the sixth object is the same as that of the third object described above, and will not be repeated here.

[0378] In the above one or more embodiments, after providing the user with the search results of at least one first object, the method may further include:

[0379] When it is detected that all the first objects in at least one first object that meet the same model condition with the target multimedia data are exposed, identify the target style corresponding to the first object that meets the same model condition with the target multimedia data; perform object search based on the target multimedia data and the target style to determine at least one fifth object; provide the user with the search results corresponding to at least one fifth object. When the target multimedia data is a target picture, that is, when it is detected that all the first objects in at least one first object that meet the same model condition with the target picture are exposed, identify the target style corresponding to the first object that meets the same model condition with the target picture, and perform object search based on the target picture and the target style.

[0380] In practical applications, there are usually requirements for the recall quantity during object search to ensure the user experience, etc. Therefore, the at least one fifth object may specifically be multiple fifth objects.

[0381] It may be to provide the search results corresponding to the at least one fifth object in the first search result page. The search results include the object prompt information of each fifth object.

[0382] As described above, at least one first object may include a first object that meets the same-style condition with the target picture and a first object that does not meet the same-style condition but meets the similar condition. If all the first objects that meet the same-style condition are exposed, the embodiments of the present application can identify the target style of these exposed objects, so as to continue object search. The at least one fifth object may be an object with the target style and the feature similarity between the image features of the object picture and the image features of the target picture is greater than a third threshold. The third threshold may be less than the first threshold and greater than the second threshold, for example.

[0383] Among them, the exposure of the first object may mean that the object prompt information of the first object is displayed in the user interface for the user to view, etc.

[0384] In one or more of the above embodiments, the user terminal may display a media search control in the user interface, which may be displayed on the home page of the user terminal or on a specific page, etc.

[0385] Taking the target multimedia data as the target picture as an example, the target picture provided by the user may be obtained by the user terminal in response to a trigger operation on the media search control, by selecting the target picture from the local picture library or calling the image acquisition component to perform image acquisition on the target object. Therefore, obtaining the target picture provided by the user may be:

[0386] Based on the image acquisition request triggered by the media search control, provide an image acquisition page to the user; obtain the target picture provided by the user for the image acquisition page.

[0387] As described above, the search prompt information may include input prompt information, and may also include acquisition prompt information. The acquisition prompt information is used to call the image acquisition component to re-acquire the target picture; in some embodiments, the method may further include:

[0388] Based on the image acquisition request triggered by the acquisition prompt information, re-obtain the target picture provided by the user and return to the step of performing object search based on the target picture to determine at least one first object and continue to execute.

[0389] Among them, based on the image acquisition request, an image acquisition page may be provided to the user again, for the user terminal to re-display the image acquisition page and re-obtain the target picture, etc.

[0390] As described above, the technical solution of the embodiments of the present application can support multiple intent operations. After the user provides search information, intent recognition can be performed first, and then different processing operations can be executed according to the corresponding target intent operation. See Figure 3As shown in the figure, it is a flowchart of another embodiment of an object processing method provided by an embodiment of the present application. The technical solution of this embodiment can be executed by a server, and the method may include the following steps:

[0391] 301: Obtain target multimedia data provided by a user, and perform object search based on the target multimedia data to determine at least one first object.

[0392] 302: Provide the user with search results corresponding to at least one first object.

[0393] 303: Provide the user with search prompt information.

[0394] 304: Obtain search information provided by the user based on the search prompt information.

[0395] 305: Identify the target intent operation corresponding to the search information.

[0396] 306: According to the processing method corresponding to the target intent operation, combine the search information and at least one first object to perform corresponding processing operations to obtain a processing result.

[0397] 307: Provide the user with the processing result.

[0398] This embodiment is different from Figure 1 the embodiment shown in that for the search information provided by the user, intent recognition can be first performed to determine the corresponding target intent operation, and then according to the processing method corresponding to the target intent operation, a more precise processing operation can be realized. Thus, while ensuring that the picture search function is enriched and the user experience is improved without increasing the object conversion rate, the accuracy of the processing operation can also be ensured, further improving the user experience.

[0399] The same or related steps in this embodiment can be found in detail in Figure 1 what is described in the embodiment shown, and will not be elaborated here in detail.

[0400] Combined with the above description, there may be various implementation manners for the target intent operation.

[0401] In some embodiments, the above-mentioned performing corresponding processing operations according to the processing method corresponding to the target intent operation, combining the search information and at least one first object to obtain a processing result may include:

[0402] In the case where the target intent operation is an object update operation, identify at least one target object attribute hit by the search information;

[0403] Perform object search by combining at least one target object attribute and at least one first object to determine at least one second object;

[0404] Providing the processing result to the user may include: providing the search result of at least one second object to the user.

[0405] Among them, the specific implementation manner of the object update operation, the specific manner of the target object attribute, and the specific determination manner of at least one second object have been described in detail in the corresponding foregoing embodiments, and will not be repeated here.

[0406] In some embodiments, after searching for objects based on the target multimedia data to determine at least one first object, the method may further include: generating a plurality of recommended questions based on the target multimedia data and the object information of at least one first object;

[0407] Providing the search prompt information to the user may include: providing the search prompt information including a plurality of recommended questions and input prompt box information to the user;

[0408] Obtaining the search information provided by the user based on the search prompt information may include: obtaining the target recommended question selected by the user from the plurality of recommended questions as the search information, or obtaining the search information input by the user for the input prompt box information.

[0409] In some embodiments, generating a plurality of recommended questions based on the target multimedia data and the object information of at least one first object may include:

[0410] Generating at least one recommended question corresponding to different intent operations based on the target multimedia data and the object information of at least one first object;

[0411] The method may further include: determining the target category corresponding to the target multimedia data based on at least one first object; determining the arrangement order of the plurality of recommended questions according to the intent priority of different intent operations corresponding to the target category; the plurality of recommended questions may be displayed in the arrangement order.

[0412] In one embodiment, the method may further include: if no target object attribute corresponding to the search information is recognized and obtained, performing object search based on the target multimedia data and the search information to obtain at least one third object; providing the search result of at least one third object to the user.

[0413] In addition, if there is at least one third object, a second prompt copy may be generated and provided to the user; otherwise, a third prompt copy may be generated and provided to the user.

[0414] In some embodiments, the above-mentioned processing method of operating according to the target intention, combining the search information and at least one first object to perform corresponding processing operations to obtain a processing result may include: when the target intention is an evaluation viewing operation, generating evaluation information corresponding to at least one first object from the evaluation data corresponding to each of the at least one first object;

[0415] Then, the above-mentioned providing the processing result to the user may include: providing the evaluation information to the user.

[0416] In some embodiments, the above-mentioned processing method of operating according to the target intention, combining the search information and at least one first object to perform corresponding processing operations to obtain a processing result may include: when the target intention operation is a consultation operation, generating a response content based on the search information and the object information of at least one first object;

[0417] Then, the above-mentioned providing the processing result to the user may be: providing the response content to the user.

[0418] In some embodiments, the above-mentioned processing method of operating according to the target intention, combining the search information and at least one first object to perform corresponding processing operations to obtain a processing result may include:

[0419] When the target intention is a matching operation, identifying the target style or target accessories corresponding to at least one first object; performing an object search based on the target multimedia data and the target style or target accessories to determine at least one fourth object;

[0420] Then, the above-mentioned providing the processing result to the user may include: providing the search result of at least one fourth object to the user.

[0421] Among them, the specific implementation manner of the matching operation can be seen in the corresponding embodiments described above, and will not be elaborated here.

[0422] In some embodiments, the above-mentioned processing method of operating according to the target intention, combining the search information and at least one first object to perform corresponding processing operations to obtain a processing result may include: when the target intention operation is an abnormal input operation, generating an abnormal prompt message.

[0423] The above-mentioned providing the processing result to the user may include: providing the abnormal prompt message to the user.

[0424] That is, when the target intention operation is an abnormal input operation, the processing operation performed on the search information and at least one first object may be empty, and an abnormal prompt message is generated.

[0425] In some embodiments, in accordance with the target intention, the corresponding processing method is operated, and the corresponding processing operation is performed in combination with the search information and at least one first object to obtain a processing result, which may include: when the target intention operation is a co-search operation, object search is performed based on the target multimedia data and the search information to obtain at least one sixth object.

[0426] Then, providing the processing result to the user may include: providing the search result of at least one sixth object to the user.

[0427] Among them, the specific implementation manners of the evaluation viewing operation, the matching operation, the consultation operation, the abnormal input operation, and the co-search operation can be seen in the corresponding embodiments described above, and will not be elaborated here.

[0428] In some embodiments, after providing the search result of at least one first object to the user, the method may further include: when it is detected that all the first objects that meet the same model condition with the target multimedia data in at least one first object are exposed, identifying the target style corresponding to the first object that meets the same model condition with the target multimedia data; performing object search based on the target multimedia data and the target style to determine at least one fifth object; providing the object prompt information corresponding to at least one fifth object respectively in the first search result page.

[0429] In addition, the search prompt information may include input prompt information, and may also include acquisition prompt information and / or audio prompt information, etc.

[0430] Figure 4 This is a flowchart of an embodiment of an interface display method provided by an embodiment of the present application. The technical solution of this embodiment can be executed by a user terminal, and the method may include the following steps:

[0431] 401: Display the search result of at least one first object in the user interface.

[0432] Among them, at least one first object is determined by performing object search based on the target multimedia data provided by the user. The specific determination method can be seen in the corresponding embodiments described above, and will not be repeated here.

[0433] Optionally, it may be to display a first search result page in the user interface, and display the search result of at least one first object in the first search result page. The search result includes the object prompt information of each first object.

[0434] In practical applications, limited by the interface size of the user interface, the search results of at least one first object are usually not fully exposed. The user can perform a paging operation, such as a swiping operation, to expose the object hint information of different first objects in the user interface. Therefore, it can be understood that the display of the search results of at least one first object in the user interface described in this article does not mean that all the search results are presented in the user interface, but rather at least part of the search results of at least one first object are displayed.

[0435] In practical applications, taking the target multimedia data as the target picture as an example, a media search control can be displayed in the user interface, which can be displayed on the home page. In response to the trigger operation on the media search control, an image acquisition page can be displayed in the user interface. The image acquisition page can present the preview image captured in real time by the image acquisition component. In the image acquisition page, the user's upload operation or acquisition operation can be sensed, so that the target picture can be obtained from the local picture library or the acquisition operation can be performed to obtain the target picture. The user terminal can then upload the target picture to the server, and the server can process it and return the search results of at least one first object, etc.

[0436] 402: Display search hint information in the user interface.

[0437] As an alternative, it can be to display search hint information in the user interface in response to a swiping operation triggered in the user interface to expose a predetermined number of first objects in the first search result page.

[0438] As another alternative, a hint object can also be displayed in the user interface; in response to a target operation triggered in the user interface, the display of search hint information in the user interface can include: in response to the trigger operation on the hint object, display search hint information in the user interface.

[0439] 403: Obtain the search information provided by the user for the search hint information.

[0440] Among them, the search information is used to perform a processing operation in combination with at least one first object to obtain a processing result. The specific way to obtain the processing result can be seen in the corresponding embodiments described above, and will not be repeated here.

[0441] 404: Display the processing result in the user interface.

[0442] Similarly, it can be understood that limited by the interface size, at least part of the processing result can be displayed in the user interface.

[0443] Among them, a target page can be displayed in the user interface, and the processing result can be displayed in the target page.

[0444] In this embodiment, by providing a user interface for human-computer interaction operations, when displaying search results of at least one first object, it supports the user to perform further operations and freely input search information. Thus, based on the search information, combined with at least one first object, further processing operations can be carried out to obtain a processing result, and the processing result is displayed on the user interface to help the user make decisions or further screen, etc., thereby enriching the picture search function, improving the user experience, and contributing to enhancing the object conversion rate.

[0445] To facilitate the user to accurately provide search information, in some embodiments, the method may further include: displaying a plurality of recommended questions in the user interface;

[0446] Then, the obtaining of the search information provided by the user in response to the search prompt information may include: in response to the user's selection operation for a plurality of recommended questions, taking the selected target recommended question as the search information; or, in response to the user's input operation for the search prompt information, determining the search information input by the user.

[0447] Among them, the search prompt information may include input prompt information, and the input prompt information may be an input prompt box, etc. It may be in response to the user's input operation triggered for the input prompt information, so as to determine the search information input by the user, etc.

[0448] Among them, the arrangement order and generation method, etc. of the plurality of recommended questions have been described in the corresponding embodiments above, and will not be elaborated here.

[0449] In addition, the search prompt information can also be requested to be closed. Therefore, in some embodiments, in response to the close operation triggered for the search prompt information, the display of the search prompt information can be terminated.

[0450] The search prompt information may include a close button, and the close operation can be triggered based on the close button.

[0451] For ease of understanding, in one or more of the following embodiments, the technical solutions of the embodiments of the present application will be introduced in combination with interface schematic diagrams. In these interface schematic diagrams, the target multimedia data is mainly taken as a target picture as an example for introduction. These interface schematic diagrams are mainly illustrated by taking an e-commerce scenario as an example. It can be understood that these drawings do not represent the user interface in the actual production process, and in the actual production process, it will change in combination with actual data, aesthetic requirements, interface layout, and style design, etc.

[0452] Assume that the target object in the target picture 100 is a seat cushion and the shape is square. The user side can display a media search control in the home page provided by the user interface. For the trigger operation for the media search control, such as Figure 5aAs shown, an image acquisition page can be displayed in the user interface 10. The image acquisition page can display the real-time acquired image screen 501, and in addition, can also display the picture selection prompt information 502. The user can select a target picture from the local picture library through the picture selection prompt information 502 or use the acquisition confirmation prompt message 503 to use the currently acquired image screen as the target picture.

[0453] After recalling at least one first object, a first search result page 20 can be displayed in the user interface, and the search results of at least one first object can be displayed on the first search result page 20. The search results can include, for example, the object prompt information 21 of the first object, etc. In response to the user's target operation, such as Figure 5b As shown, in the user interface 10, for example, at the bottom boundary position, the search prompt information 30 can be displayed. In addition, recommended questions 40, etc. can also be displayed. Among them, the search prompt information 30 and the recommended questions 40 can be displayed in a pop-up window page or a floating layer page, and can cover part of the first search result page.

[0454] In some embodiments, the processing result can include the search results of at least one second object; the search information can be used to determine at least one target object attribute it hits; at least one target object attribute is used to combine with target multimedia data for object search to determine at least one second object.

[0455] Optionally, the search information can first be used to identify its corresponding target intent operation, and in the case where the target intent operation is an object update operation, determine at least one target object attribute it hits, etc.

[0456] The above display of the processing result in the user interface can include: displaying the search results of at least one second object in the user interface.

[0457] It can be to display a second search result page in the user interface and display the search results of at least one second object on the second search result page. Among them, the search results can include the object prompt information of each second object. The target page is the second search result page.

[0458] The second search result page can cover at least part of the first search result page. In addition, in response to the close operation on the second search result page, the first search result page can continue to be displayed.

[0459] Among them, according to the at least one target object attribute, the object update operation can be determined as object rewriting or object screening.

[0460] Such as Figure 5aAs shown, the target object in the target picture provided by the user is a "seat cushion" with a square shape. Among the at least one first object recalled, there is a seat cushion product with a square shape.

[0461] Suppose the search information 11 is "round", and the shape of the target object attribute it hits is round. This target object attribute is a rewritten attribute, and the object update operation is an object rewriting operation. The at least one second object recalled includes seat cushion products with a round shape. As Figure 5c shown, the second search result page 50 can be displayed in the user interface, covering at least part of the first search result page 20. At least one second object can be displayed on the second search result page 50, that is, the search result of "round seat cushion", and the search result includes the object prompt information 51 of the second object.

[0462] Suppose the search information 11 is "between 10 yuan and 20 yuan", and the target object attribute hit is a price between 10 yuan and 20 yuan. This target object attribute is a filtering attribute, and the object update operation is an object filtering operation. The at least one second object recalled is the "square seat cushion with a price range between 10 yuan and 20 yuan" obtained by filtering from at least one first object. As Figure 5d shown, the search result of at least one second object displayed on the second search result page 50 is the "square seat cushion with a price range between 10 yuan and 20 yuan".

[0463] In addition, in some embodiments, a first prompt text can also be displayed on the second search result page; the first prompt text can be generated according to the search information and the recognition result of at least one target object attribute.

[0464] As Figure 5c shows the first prompt text 52, and as Figure 5d the first prompt text 52 in etc. In addition, the search information 11 can also be displayed on the second search result page.

[0465] In some embodiments, the method may further include:

[0466] Obtain the search result of at least one third object sent by the server; the at least one third object is obtained by performing object search based on the target multimedia data and the search information when no target object attribute corresponding to the search information is recognized; display the search result of at least one third object in the user interface.

[0467] It can be to display a third search result page in the user interface and display the search result of at least one third object on the third search result page, and the search result includes the object prompt information of each third object.

[0468] In addition, a second prompt copy, a third prompt copy, etc. can also be displayed on the third search result page.

[0469] In some embodiments, the method may further include:

[0470] Obtaining search results of at least one fifth object sent by the server; the at least one fifth object is determined by identifying a target style corresponding to a first object that meets the same model condition as the target multimedia data when all the first objects that meet the same model condition as the target multimedia data in at least one first object are exposed, and performing object search based on the target multimedia data and the target style;

[0471] Displaying the search results of the at least one fifth object in the user interface.

[0472] Among them, the search results of the at least one fifth object can be displayed on the first search result page, etc.

[0473] In some embodiments, the processing result may include evaluation information, and the search information can be used to identify the corresponding target intent operation; the evaluation information is determined based on the evaluation data corresponding to at least one first object when the target intent operation is an evaluation viewing operation;

[0474] Then, the above-mentioned displaying the processing result in the user interface may include: displaying the evaluation information in the user interface.

[0475] An evaluation result page can be displayed in the user interface, and the evaluation information can be displayed on the evaluation result page. The target page is the evaluation result page.

[0476] Among them, the evaluation information may include, for example, the target evaluation data of each first object. In addition, it may also include the comprehensive evaluation data of at least one first object, etc.

[0477] For example, the search information 11 provided by the user is "view real evaluations", and the corresponding target intent operation can be identified as an evaluation viewing operation according to this search information. As Figure 5e shown, an evaluation result page 60 can be displayed in the user interface 10, covering at least part of the first search result page 20. The target evaluation data 61 corresponding to at least one first object can be displayed on the evaluation result page 60. Among them, the target evaluation data 61 corresponding to the at least one first object can be arranged and displayed on the evaluation result page according to the arrangement order of the at least one first object, etc.

[0478] A fourth prompt copy 62 can also be displayed on the evaluation result page 60. In addition, the search information 11 can also be displayed.

[0479] In some embodiments, the processing result includes the search result of at least one fourth object, and the search information is used to identify the corresponding target intent operation; the at least one fourth object is obtained by performing an object search based on the target multimedia data and the target style or target accessories of at least one first object when the target intent operation is a matching operation;

[0480] Displaying the processing result in the user interface may include: displaying the search result of at least one fourth object in the user interface.

[0481] It may be to display a fourth search result page in the user interface and display the search result of at least one fourth object on the fourth search result page, and the search result may include the object prompt information of each fourth object.

[0482] For example, the search information 11 provided by the user is "looking for accessories". According to this search information, the corresponding target intent operation can be identified as a matching operation, and at least one first object is "seat cushion", and the target accessory can be identified as the "elastic band" for fixing coordinates. As Figure 5f shown, a fourth search result page 70 can be displayed in the user interface 10, covering at least part of the first search result page 20. At least one fourth object, that is, the object prompt information 71 corresponding to the "elastic band" can be displayed on the fourth search result page 70.

[0483] In addition, a fifth prompt text 72 can also be displayed on the fourth search result page 70. In addition, the search information 11 can also be displayed.

[0484] In some embodiments, the processing result may include a response content; the search information is used to identify the corresponding target intent operation; the response content is generated based on the search information and the object information of at least one first object when the target intent operation is a consultation operation;

[0485] Then, the above-mentioned displaying the processing result in the user interface may include: displaying the response content in the user interface.

[0486] Optionally, it may be to display the response content on the target page.

[0487] As can be seen from the foregoing description, the consultation operation may include an operation to obtain a purchase guide or an operation to obtain a buying guide. Of course, it may also be a meaningless chat operation, etc.

[0488] In Figure 5a shown under the search result of at least one first object, assuming the search information is "what kind of material of seat cushion is the most comfortable", then it can be determined that the target intent operation is an operation to obtain a purchase guide. As Figure 5gAs shown, the target page 80 can be displayed in the user interface 10, and the corresponding response content 81 such as "The most comfortable cushion materials include ice silk, linen, velvet, pure cotton with Lycra, and genuine leather" can be displayed in the target page 80. In addition, the search information 11, etc. can also be displayed.

[0489] Suppose the search information is "How to clean a cushion made of cowhide", then it can be determined that the target intent operation is the operation of obtaining post-purchase strategies. Similarly, the target page can be displayed in the user interface, and the corresponding response content such as "Gently wipe the surface dust with a clean and soft dry cloth, and avoid scratching the leather surface with rough materials. For slight stains, gently wipe with a slightly wet soft cloth dipped in a neutral cleaner, and then immediately dry with a dry cloth" can be displayed in the target page. In addition, the search information, etc. can also be displayed in the target page, and no further illustration will be given.

[0490] In some embodiments, search prompt information can also be displayed in the target page, such as Figure 5g the search prompt information 30 shown in, so that it can be convenient for the user to continue to input search information, and based on this re-input search information, the server can continue to perform corresponding operations according to the technical solution of the present application. The operations performed for the search information can be seen in the corresponding embodiments described above.

[0491] In some embodiments, the processing result can include exception prompt information; the search information is used to identify the corresponding target intent operation; the exception prompt information is generated when the target intent operation is an abnormal input operation;

[0492] Then the above display of the processing result in the user interface can include: displaying the exception prompt information in the user interface.

[0493] The exception prompt information can be overlaid and displayed at the display position corresponding to the search prompt information, etc. In addition, the exception prompt information can also be closed to continue to display the search prompt information, etc. Of course, the exception prompt information can also be displayed in the target page. When the target page is closed, the first search result page and the search prompt information can continue to be displayed.

[0494] In some embodiments, the processing result can include the search results of at least one sixth object; the search information is used to identify the corresponding target intent operation; at least one sixth object is determined by performing object search based on the target multimedia data and the search information when the target intent operation is a co-search operation;

[0495] Then the above display of the processing result in the user interface can include: displaying the search results of at least one sixth object in the user interface.

[0496] It may be to display the sixth search result page on the user interface, and display the search results of at least one sixth object in the sixth search result page, where the search results include the object prompt information of each sixth object.

[0497] Among them, as Figure 5a shown in, the search prompt information 30 may further include acquisition prompt information 31, audio prompt information 32, input prompt box 33, etc.

[0498] Thus, in response to an input operation on the input prompt box 33, the search information input by the user can be determined. In response to a trigger operation on the acquisition prompt information 31, the image acquisition page can be redisplayed to re-obtain the target picture provided by the user, etc. In response to a trigger operation on the audio prompt information 32, audio data can be collected, and the search information obtained by performing speech recognition conversion on the audio data can be displayed in the input prompt box, etc.

[0499] Of course, the search prompt information 30 may further include a confirmation control 34. For the search information input by the user, in response to a confirmation operation on the confirmation control, the search information is sent to the server.

[0500] In the case where the target recommended question selected from multiple recommended questions is used as the search information, the search information can be directly sent to the server.

[0501] In the embodiments of the present application, the above-mentioned intention recognition, target object attribute recognition, generation of response content, generation of prompt copywriting, and generation of recommended questions can be implemented by means of an AI large model. Using AI technology can ensure processing accuracy and the friendliness of interaction with users, further ensuring the user experience.

[0502] In addition, Figure 6 In the scene interaction schematic diagram described above, the data communication process between the user terminal 601 and the server 602 is shown. Taking the target multimedia data as the target picture as an example, the user terminal 601 can perceive the target picture provided by the user, and can send a picture search request to the server 602 based on the target picture. The server 602 can perform object search based on the feature similarity determined based on the target picture in the picture search request, and recall at least one first object in combination with the recall quantity. The search results of the at least one first object can be sent to the user terminal 601, and the user terminal 601 can display the search results of the at least one first object in the first search result page in the user interface.

[0503] The client 601 can sense the sliding operation performed by the user on the first search result page. Thus, when a predetermined number of first objects are exposed or the number of page flips of the first search result page reaches a specified number, such as 2 times, it can be considered that the user is not satisfied with the search results. At this time, search prompt information and recommended questions can be displayed in the user interface, and the input operation based on the search prompt information or the selection operation based on the recommended questions by the user can be sensed, so as to determine the search information provided by the user, and the search information is sent to the server 602.

[0504] The server 602 can perform intent recognition and intent diversion based on the search information to determine the processing methods corresponding to different intent operations and execute the processing operations and processing results. Different intent operations can include, for example, object update operation 81, evaluation view operation 82, consultation operation 83, matching operation 84, abnormal input operation 85, and co-search operation 86, etc. Different intent operations can obtain different processing results, such as the search results 91 of at least one second object corresponding to the object update operation; the evaluation information 92 corresponding to the evaluation view operation, the response content 93 corresponding to the resource operation, the search results 94 of at least one fourth object corresponding to the matching operation, the abnormal prompt information 95 corresponding to the abnormal input operation, and the search results 96 of at least one sixth object corresponding to the co-search operation, etc.

[0505] Of course, the specific implementation manners of the client 601 and the server 602 for different intent operations can be seen in the corresponding embodiments described above, and will not be repeated here.

[0506] The server 602 can send the processing result to the client 601, and the client 601 can display the processing result on the target page. Of course, the search information can also be displayed on the target page. In addition, corresponding prompt copy can be generated based on the search information and displayed on the target page, etc.

[0507] Through the technical solution of the embodiment of the present application, the picture search function is enriched, rather than simply displaying search results, improving the user experience. In the case of improved user experience, it will encourage users to perform object conversion operations, thus helping to improve the object conversion rate. And it can also provide more accurate information to users to encourage users to perform further object conversion operations, thus helping to improve the object conversion rate.

[0508] In the e-commerce scenario, the object is a commodity. Adopting the technical solution of the embodiment of the present application will effectively improve the shopping experience of users and thus improve the commodity purchase rate, etc.

[0509] It should be noted that in some of the processes described in the above embodiments and the accompanying drawings, a plurality of operations appear in a specific order. However, it should be clearly understood that these operations may not be executed in the order in which they appear herein or may be executed in parallel. The operation numbers such as 101 and 102 are only used to distinguish different operations, and the numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions such as "first" and "second" herein are used to distinguish different messages, devices, modules, etc., and do not represent a sequence, nor do they limit that "first" and "second" are of different types.

[0510] Figure 7 FIG. 4 is a schematic structural diagram of an embodiment of an object processing device provided by an embodiment of the present application. The device may include:

[0511] A first search module 701, configured to obtain target multimedia data provided by a user, and perform object search based on the target multimedia data to determine at least one first object.

[0512] A first providing module 702, configured to provide a user with search results corresponding to at least one first object;

[0513] A second providing module 703, configured to provide a user with search prompt information;

[0514] A first obtaining module 704, configured to obtain search information provided by a user based on the search prompt information;

[0515] A first processing module 705, configured to perform a processing operation by combining the search information and at least one first object to obtain a processing result;

[0516] A third providing module 706, configured to provide a user with the processing result.

[0517] In some embodiments, the first processing module may specifically be configured to identify a target intent operation corresponding to the search information; and perform a corresponding processing operation by combining the search information and at least one first object according to the processing method corresponding to the target intent operation to obtain a processing result.

[0518] Figure 7 The object processing device described above may execute Figure 1 or Figure 3 the object processing method described in the embodiments shown. The implementation principles and technical effects will not be elaborated herein. For the object processing device in the above embodiments, the specific manners in which each module and unit perform operations have been described in detail in the embodiments related to the method, and will not be elaborated herein.

[0519] Figure 8 Schematic structural diagram of another embodiment of an object processing device provided by an embodiment of the present application. The device may include:

[0520] A first search module 801, configured to obtain target multimedia data provided by a user, and perform object search based on the target multimedia data to determine at least one first object.

[0521] A first providing module 802, configured to provide a user with search results corresponding to at least one first object;

[0522] A second providing module 803, configured to provide a user with search prompt information;

[0523] A first obtaining module 804, configured to obtain search information provided by a user based on the search prompt information;

[0524] An identification module 805, configured to identify at least one target object attribute hit by the search information;

[0525] A second search module 806, configured to perform object search by combining at least one target object attribute and at least one first object to determine at least one second object;

[0526] A fourth providing module 807, configured to provide a user with search results of at least one second object.

[0527] Figure 8 The object processing device described above may execute Figure 2 The object processing method described in the illustrated embodiment. The implementation principle and technical effects will not be elaborated further. For the object processing device in the above embodiment, the specific manners in which each module and unit perform operations have been described in detail in the embodiment related to the method, and will not be elaborated in detail here.

[0528] Figure 9 Schematic structural diagram of an embodiment of an interface display device provided by an embodiment of the present application. The device may include:

[0529] A first display module 901, configured to display search results of at least one first object in a user interface; the at least one first object is determined by performing object search based on target multimedia data provided by a user;

[0530] A second display module 902, configured to display search prompt information in the user interface;

[0531] A second obtaining module 903, configured to obtain search information provided by a user for the search prompt information; the search information is used to perform a processing operation by combining at least one first object to obtain a processing result;

[0532] The third display module 904 is configured to display a processing result in a user interface.

[0533] In some embodiments, the processing result may include search results of at least one second object; the search information is used to determine at least one target object attribute that it hits; the at least one target object attribute is used to perform object search in combination with target multimedia data to determine at least one second object;

[0534] The above-mentioned third display module may specifically display search results of at least one second object in a user interface.

[0535] Figure 9 The described interface display device may execute Figure 4 The interface display method described in the illustrated embodiment, and its implementation principle and technical effects will not be elaborated further. For the object processing device in the above embodiment, the specific manners in which each module and unit perform operations have been described in detail in the embodiment related to the method, and will not be elaborated here.

[0536] Figure 10 This is a schematic structural diagram of an embodiment of a computing device provided by the present application. As Figure 10 shown, in practice, the computing device may include: a storage component 1001 and a processing component 1002.

[0537] The storage component 1001 is configured to store computer programs and may be configured to store various other data to support operations on the computing device. Examples of such data include instructions for any application program or method for operating on the computing device, data structures, contact data, phone book data, messages, pictures, videos, etc.

[0538] The processing component 1002 is coupled to the storage component 1001 and is configured to execute the computer programs in the storage component 1001 to implement the object processing method as Figure 1 shown or the object processing method as Figure 2 shown or the object processing method as Figure 3 shown or the interface display method as Figure 4 shown.

[0539] Furthermore, as Figure 10 shown, the computing device further includes: other components such as a communication component 1003, a display component 1004, a power supply component 1005, an audio component 1006, etc. Figure 10 Only some components are schematically shown, and it does not mean that the computing device only includes Figure 10 the components shown. Additionally, Figure 10The components within the dashed-line box are optional components, rather than mandatory components, and can be determined according to the product form of the working node. The working node in this embodiment can be implemented as a terminal device such as a desktop computer, a laptop computer, a smart phone, or an IOT device, or can also be a server device such as a conventional server, a cloud server, or a server array. If the working node in this embodiment is implemented as a terminal device such as a desktop computer, a laptop computer, or a smart phone, it may include Figure 10 the components within the dashed-line box; if the working node in this embodiment is implemented as a server device such as a conventional server, a cloud server, or a server array, it may not include the components within the dashed-line box in Figure 5.

[0540] The above storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random-Access Memory (SRAM), Electrically Erasable Programmable Read Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic memory, flash memory, a magnetic disk, or an optical disc.

[0541] The above communication component is configured to facilitate communication between the device where the communication component is located and other devices in a wired or wireless manner. The device where the communication component is located can access a wireless network based on a communication standard, such as a mobile communication network such as 2G, 3G, 4G / LTE, 5G, or a combination thereof. In an exemplary embodiment, the communication component receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel.

[0542] The above display component includes a screen, and the screen can include a Liquid Crystal Display (LCD) and a TouchPanel (TP). If the screen includes a touch panel, the screen can be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensor can not only sense the boundaries of touch or swipe actions, but also detect the duration and pressure associated with the touch or swipe operation.

[0543] The above power supply component provides power for various components of the device where the power supply component is located. The power supply component may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the device where the power supply component is located.

[0544] The above audio component can be configured to output and / or input audio signals. For example, the audio component includes a microphone (Microphone, MIC). When the device where the audio component is located is in an operating mode, such as a call mode, a recording mode, and a voice recognition mode, the microphone is configured to receive external audio signals. The received audio signals can be further stored in the memory or sent via the communication component. In some embodiments, the audio component further includes a speaker for outputting audio signals.

[0545] Accordingly, an embodiment of the present application further provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor is enabled to implement the steps in the above method embodiments. Among them, the computer-readable storage medium can be implemented by volatile or non-volatile or a combination thereof, and can be removable or non-removable. Examples of computer-readable storage media include, but are not limited to, phase-change random access 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), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette tapes, magnetic disk storage or other magnetic storage devices or any other non-transmission medium

[0546] Accordingly, an embodiment of the present application further provides a computer program product. The computer program product includes a computer program or instructions. When the computer program or instructions are executed by a processor, the processor is enabled to implement the steps in the above method embodiments. It should be understood that each process or a combination of multiple processes in the above method flow can be implemented by the computer program or instructions. In addition, these computer programs or instructions can be applied to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices, so that the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices can be used as devices to implement the corresponding functions in the above method embodiments.

[0547] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0548] It should also be noted that the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, commodity or device comprising a series of elements not only includes those elements but also includes other elements not expressly listed, or further includes elements inherent in such process, method, commodity or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, commodity or device comprising the element.

[0549] Finally, it should be noted that the above are only embodiments of the present application and are not used to limit the present application. For those skilled in the art, the present application may have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.

Claims

1. An object processing method, characterized in that: include: Acquire target multimedia data provided by a user, and perform an object search based on the target multimedia data to determine at least one first object; providing the user with search results of the at least one first object; Providing search prompt information to the user; Acquiring search information provided by the user based on the search prompt information; Identifying at least one target object attribute hit by the search information; performing an object search in combination with the at least one target object attribute and the at least one first object to determine at least one second object; A search result of the at least one second object is provided to the user.

2. The method according to claim 1, characterized in that Also includes: Determine a set identifier corresponding to the at least one object and a first attribute set indexed by the set identifier; wherein the set identifier is used to identify a target object set to which the at least one object belongs in multiple sets; the multiple object sets are obtained by grouping multiple objects according to feature similarity based on object features; determining a second attribute set consisting of attributes respectively possessed by the at least one first object; Merging the first attribute set and the second attribute to obtain a target attribute set; The identifying at least one target object attribute hit by the search information comprises: At least one target object attribute hit by the search information in the target attribute set is identified.

3. The method according to claim 2, characterized in that The performing an object search in combination with the at least one target object attribute and the at least one first object to determine at least one second object comprises: The object attributes in the second attribute set are divided into a third attribute set that meets the same condition and a fourth attribute set that does not meet the same condition according to whether the feature similarity between the corresponding first object and the target multimedia data meets the same condition; After filtering out the third attribute set from the first attribute set, merging the third attribute set with the fourth attribute set to obtain a fifth attribute set; If the at least one target object attribute is in the third attribute set, searching for at least one second object having the at least one target object attribute from the multiple objects according to the first recall quantity; If any target object attribute is in the fifth attribute set, at least one second object having the at least one target object attribute is searched from the multiple objects according to a second recall quantity; wherein the second recall quantity is greater than the first recall quantity.

4. The method according to claim 1, characterized in that: Also includes: Determine a target object set to which the at least one first object belongs in a plurality of object sets; the plurality of object sets are obtained by grouping a plurality of objects according to feature similarity based on object features; Determine a target attribute set consisting of different objects in the target object set and object attributes respectively possessed by the at least one object; The identifying at least one target object attribute hit by the search information comprises: At least one target object attribute hit by the search information in the target attribute set is identified.

5. The method according to claim 3, characterized in that: Also includes: Determine the number of groups of the plurality of object groups divided in the target object set; The plurality of object groups are obtained by dividing the objects in the target object set based on object features; The second recall quantity is determined according to the grouping quantity.

6. The method according to claim 1, characterized in that The identifying at least one target object attribute hit by the search information comprises: Based on the search information, identifying a target intended operation; In the case where the target intended operation is an object update operation, at least one target object attribute hit by the search information is identified.

7. The method according to claim 1 or 6, characterized in that: After performing object search based on the target multimedia data to determine at least one first object, the method further includes: generating a plurality of recommendation questions based on the target multimedia data and the object information of the at least one first object; The providing search prompt information to the user comprises: Providing the user with search prompt information including the multiple recommended questions and input prompt information; The acquiring the search information provided by the user based on the search prompt information includes: A target recommended question selected by the user from the plurality of recommended questions is obtained to use the target recommended question as search information, or search information input by the user in response to the input prompt information is obtained.

8. The method according to claim 7, characterized in that The generating a plurality of recommendation questions based on the target multimedia data and the object information of the at least one first object comprises: Based on the target multimedia data and the object information of the at least one first object, generating at least one recommendation question corresponding to different intention operations respectively; The method further comprises: Based on the at least one first object, determining a target category corresponding to the target multimedia data; An arrangement order of multiple recommended questions is determined according to the intention priorities of different intention operations corresponding to the target category; and the multiple recommended questions are displayed according to the arrangement order.

9. The method according to claim 1 or 6, characterized in that: The identifying at least one target object attribute hit by the search information comprises: Using the first large model to identify at least one target object attribute hit by the search information; The method further comprises: Generate a first prompt text using the first large model according to the search information and the recognition result of the at least one target object attribute; The first prompt text is provided to the user.

10. The method according to claim 1, characterized in that Also includes: If any target object attribute corresponding to the search information is not identified, performing an object search based on the target multimedia data and the search information to obtain at least one third object; A search result of the at least one third object is provided to the user.

11. The method according to claim 6, characterized in that Also includes: In the case where the target intention is an evaluation viewing operation, generating evaluation information corresponding to the at least one first object from the evaluation data respectively corresponding to the at least one first object; the evaluation information includes the target evaluation data respectively corresponding to the at least one first object; providing an evaluation result page to the user, and displaying the target evaluation data respectively corresponding to the at least one first object in the evaluation result page according to the display order of the at least one first object; Alternatively, in the case where the target intended operation is a consulting operation, a fourth model is used to generate response content based on the search information and the object information of the at least one first object; Providing the response content to the user; Alternatively, when the target intention is a matching operation, identifying a target style or a target accessory corresponding to the at least one first object; Performing an object search based on the target multimedia data and the target style or the target accessories to determine at least one fourth object; providing the search result of the at least one fourth object to the user; Alternatively, when the target intended operation is a co-search operation, performing an object search based on the target multimedia data and the search information to obtain at least one sixth object; A search result of the at least one sixth object is provided to the user.

12. The method according to claim 1, characterized in that After providing the search result of at least one first object to the user, the method further includes: When detecting that all first objects among the at least one first object satisfying the same condition as the target multimedia data are exposed, identifying a target style corresponding to the first object satisfying the same condition as the target multimedia data; performing an object search based on the target multimedia data and the target style to determine at least one fifth object; Object prompt information corresponding to the at least one fifth object is provided in the first search result page.

13. An object processing method, characterized in that: include: Acquire target multimedia data provided by a user, and perform an object search based on the target multimedia data to determine at least one first object; Providing the user with search results corresponding to the at least one first object; Providing search prompt information to the user; Acquiring search information provided by the user based on the search prompt information; performing a processing operation in combination with the search information and the at least one first object to obtain a processing result; The processing result is provided to the user.

14. The method according to claim 13, characterized in that The performing a corresponding processing operation in combination with the search information and the at least one first object to obtain a processing result includes: Identifying a target intention operation corresponding to the search information; According to the processing method corresponding to the target intention operation, the corresponding processing operation is performed in combination with the search information and the at least one first object to obtain a processing result.

15. An interface display method, characterized in that: include: displaying search results for at least one first object in a user interface; The at least one first object is determined by performing an object search based on target multimedia data provided by a user; Displaying search prompt information in the user interface; Acquire search information provided by the user in response to the search prompt information; The search information is used to perform a processing operation in combination with the at least one first object to obtain a processing result; The processing result is displayed in the user interface.

16. The method according to claim 15, characterized in that Displaying search prompt information in the user interface includes: In response to a sliding operation triggered on the user interface to expose a predetermined number of first objects in the first search result page, displaying search prompt information in the user interface; Alternatively, displaying a prompt object in the user interface; and displaying search prompt information in the user interface in response to a target operation triggered in the user interface includes: In response to a trigger operation on a prompt object, search prompt information is displayed in the user interface.

17. The method according to claim 15, characterized in that Also includes: displaying a plurality of recommended questions in the user interface; The obtaining of the search information provided by the user in response to the search prompt information includes: In response to the user's selection operation on the plurality of recommended questions, using the selected target recommended question as search information; Alternatively, in response to the user's input operation on the search prompt information, the search information input by the user is determined.

18. The method according to claim 15, characterized in that The processing result includes evaluation information, and the search information is used to identify the corresponding target intended operation; the evaluation information is determined based on the evaluation data corresponding to the at least one first object when the target intended operation is an evaluation viewing operation; The displaying of the processing result in the user interface includes: displaying the evaluation information in the user interface; Alternatively, the processing result includes a search result of at least one fourth object, and the search information is used to identify a corresponding target intention operation; the at least one fourth object is obtained by performing an object search based on the target multimedia data and a target style or target accessory of the at least one first object when the target intention operation is a matching operation; and displaying the processing result in the user interface includes: displaying the search result of the at least one fourth object in the user interface; Alternatively, the processing result includes a response content; the search information is used to identify a corresponding target intention operation; the response content is generated based on the search information and the object information of the at least one first object when the target intention operation is a consulting operation; the displaying of the processing result in the user interface includes: displaying the response content in the user interface; Alternatively, the processing result includes abnormal prompt information; the search information is used to identify the corresponding target intended operation; the abnormal prompt information is generated when the target intended operation is an abnormal input operation; and displaying the processing result in the user interface includes: displaying the abnormal prompt information in the user interface; Alternatively, the processing result includes search results of at least one sixth object; the search information is used to identify the corresponding target intention operation; the at least one sixth object is determined by performing an object search based on the target multimedia data and the search information when the target intention operation is a co-search operation; and displaying the processing result in the user interface includes: displaying the search results of the at least one sixth object in the user interface.

19. A computing device, characterized in that including a processing component and a storage component; The storage component stores a computer program; the computer program is used to be called and executed by the processing component to implement the object processing method as described in any one of claims 1 to 12, or the object processing method as described in any one of claims 13 to 14, or the interface display method as described in any one of claims 15 to 18.

20. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by the processing component, it implements the object processing method according to any one of claims 1 to 12, the object processing method according to any one of claims 13 to 14, or the interface display method according to any one of claims 15 to 18.

21. A computer program product, characterized in that It includes a computer program / instruction, which, when executed by a processing component, implements the object processing method as described in any one of 1 to 12, the object processing method as described in any one of claims 13 to 14, or the interface display method as described in any one of claims 15 to 18.

Images