Target object labeling method and device

Abstract

The embodiment of the invention provides a target object labeling method which comprises the following steps: determining the type of a target object, and obtaining a target labeling strategy graph from a plurality of pre-configured labeling strategy graphs; starting from the first annotation node of the target annotation strategy graph, carrying out multiple rounds of circulation until the target annotation strategy graph flows to the last annotation node of the target annotation strategy graph, and obtaining a target annotation result; wherein each round of circulation comprises the following steps: determining a corresponding labeling result based on a labeling rule corresponding to a currently flowing labeling node, determining a target edge based on the corresponding labeling result and a circulation condition of one or more associated edges, and circulating to a next labeling node through the target edge; wherein the target labeling result comprises a labeling result corresponding to each flowing labeling node. According to the technical scheme of the embodiment of the invention, automatic labeling process circulation can be carried out, flexible arrangement of labeling processes of various types of target objects is realized, and the operation efficiency is optimized.

Description

Technical Field

[0001] Embodiments of the present application relate to the technical field of data processing, and in particular, to a method, device, computer device, computer-readable storage medium, and computer program product for annotating a target object. Background Art

[0002] In platform operation, annotating target objects (such as live streamers, products, videos, etc.) can provide key support for adjusting operation strategies. However, the existing annotation schemes have fixed processes and are difficult to quickly adjust and optimize according to different business requirements. As a result, when the platform faces new business scenarios or new target objects, it is necessary to redesign the annotation process, consuming a large amount of resources and affecting operation efficiency.

[0003] It should be noted that the above content is not necessarily prior art and is not used to limit the patent protection scope of the present application. Summary of the Invention

[0004] Embodiments of the present application provide a method, device, computer device, computer-readable storage medium, and computer program product for annotating a target object to solve or alleviate one or more of the above technical problems.

[0005] One aspect of the embodiments of the present application provides a method for annotating a target object, the method including: Determine the type of the target object; Based on the type of the target object, obtain a target annotation strategy graph from a plurality of pre-configured annotation strategy graphs; wherein, the annotation strategy graph includes a plurality of annotation nodes, each annotation node is associated with one or more edges, each edge is used to connect two annotation nodes, each annotation node is configured with a corresponding annotation rule, and each edge is configured with a corresponding transition condition; Starting from the first annotation node of the target annotation strategy graph, perform multiple rounds of transitions until reaching the last annotation node of the target annotation strategy graph to obtain a target annotation result; wherein, each round of transition includes: determining a corresponding annotation result based on the annotation rule corresponding to the currently flowing-through annotation node, determining a target edge based on the corresponding annotation result and the transition conditions of one or more associated edges, and transitioning to the next annotation node through the target edge; Wherein, the target annotation result includes the annotation results corresponding to each flowing-through annotation node.

[0006] Optionally, the annotation rules include pre-configured data processing rules, index positioning rules, mapping rules, and operation rules; Correspondingly, determining a corresponding annotation result based on the corresponding annotation rule includes: Obtain the target data of the target object according to the data processing rule; Determine multiple target metrics according to the metric positioning rule; Determine the rating of the target data of the target object under the multiple target metrics to obtain a rating result; Determine the annotation result based on the rating result, the mapping rule, and the operation rule.

[0007] Optionally, the data processing rule includes a data source, a data type, a data format, and a data filtering logic; Correspondingly, obtaining the target data of the target object according to the data processing rule includes: Based on the data processing rule, determine a target data source from multiple data sources; Obtain the initial data of the target object from the target data source; Based on the data processing rule, perform type conversion, format standardization, and / or data filtering on the initial data to obtain the target data of the target object.

[0008] Optionally, determining the rating of the target data of the target object under the multiple target metrics to obtain a rating result includes: Push the target data of the target object and the multiple target metrics to a rating component, where the rating component is used to provide a rating page, and the rating page is used to display the target data of the target object and the multiple target metrics; wherein, each of the target metrics is associated with multiple rating items; In response to selecting one of the rating items associated with each target metric, determine the rating of the target data of the target object under the multiple target metrics to obtain the rating result.

[0009] Optionally, the annotation result includes a first annotation result and a second annotation result; correspondingly, determining the annotation result based on the rating result, the mapping rule, and the operation rule includes: Based on the mapping rule, map the rating result to a first annotation result in a preset data structure; Based on the first annotation result and the operation rule, determine the second annotation result; Wherein, the first annotation result and the second annotation result have different presentation forms.

[0010] Optionally, each edge is associated with a pre-configured weight; correspondingly, determining a target edge based on the corresponding annotation result and the transfer condition corresponding to one or more associated edges includes: Match the corresponding annotation result with the transfer conditions corresponding to one or more associated edges, and determine the initial edges for the successfully matched edges; Determine the target edges based on the weights of the initial edges, where the target edges are used to perform the transfer of the annotation nodes.

[0011] Optionally, the annotation method for the target object further includes: In the case of determining the annotation result corresponding to the currently flowing annotation node, generate an annotation message based on the type of the target object, the currently flowing annotation node, and the corresponding annotation result; Deliver the annotation message to the downstream system, where the downstream system is used to perform interaction operations on the target object according to the annotation message.

[0012] Another aspect of the embodiments of the present application provides an annotation device for a target object, where the device includes: A determination module, configured to determine the type of the target object; An acquisition module, configured to obtain a target annotation policy graph from a plurality of pre-configured annotation policy graphs based on the type of the target object; where the annotation policy graph includes a plurality of annotation nodes, each annotation node is associated with one or more edges, each edge is used to connect two annotation nodes, each annotation node is configured with a corresponding annotation rule, and each edge is configured with a corresponding transfer condition; An annotation module, configured to start from the first annotation node of the target annotation policy graph, perform multiple rounds of transfer until reaching the last annotation node of the target annotation policy graph to obtain a target annotation result; where each round of transfer includes: determining a corresponding annotation result based on the annotation rule corresponding to the currently flowing annotation node, determining the target edge based on the transfer condition of one or more associated edges corresponding to the corresponding annotation result, and transferring to the next annotation node through the target edge; Wherein, the target annotation result includes the annotation results corresponding to each flowing annotation node.

[0013] Another aspect of the embodiments of the present application provides a computer device, including: At least one processor; and A memory communicatively connected to the at least one processor; Wherein: the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method as described above.

[0014] Another aspect of the embodiments of the present application provides a computer-readable storage medium, where computer instructions are stored in the computer-readable storage medium, and when the computer instructions are executed by a processor, the method as described above is implemented.

[0015] Another aspect of the embodiments of the present application provides a computer program product, including a computer program, which implements the method described above when executed by a processor.

[0016] The embodiments of the present application adopting the above technical solutions may include the following advantages: Determine the type of the target object, and obtain the target annotation strategy graph from a plurality of pre-configured annotation strategy graphs based on the type of the target object. Among them, the annotation strategy graph includes a plurality of annotation nodes, each annotation node is associated with one or more edges, and each edge is used to connect two annotation nodes. Each annotation node is also configured with a corresponding annotation rule, and each edge is also configured with a corresponding transfer condition. Enter the first annotation node of the target annotation strategy graph, and determine the corresponding annotation result based on the corresponding annotation rule. Based on the corresponding annotation result and the transfer conditions corresponding to one or more associated edges, determine the target edge. Transfer to the next annotation node through the target edge, and determine the annotation result corresponding to the next annotation node for the next round of transfer until the last annotation node of the target annotation strategy graph is reached, and obtain the target annotation result. Among them, the target annotation result includes the annotation results corresponding to each passed annotation node. It can be seen that in the embodiments of the present application, by defining and configuring various different annotation nodes, edges and related messages (annotation rules, transfer conditions) in the annotation process through the graph structure language, the automated annotation process transfer can be performed, the flexible orchestration of the annotation processes for various types of target objects can be realized, refined annotation results can be provided, and the operation efficiency can be optimized. Description of the Drawings

[0017] The drawings exemplarily show the embodiments and form a part of the description, and are used together with the written description of the description to explain the exemplary embodiments of the embodiments. The shown embodiments are only for illustrative purposes and do not limit the scope of the claims. In all the drawings, the same reference numerals refer to similar but not necessarily identical elements.

[0018] Figure 1 Schematically shows a flowchart of a method for annotating a target object according to Embodiment 1 of the present application; Figure 2 Schematically shows an annotation strategy graph according to Embodiment 1 of the present application; Figure 3 Schematically shows Figure 1 a sub-step flowchart of step S104 in Figure 4 Schematically shows Figure 3 a sub-step flowchart of step S300 in Figure 5 Schematically shows Figure 3 a sub-step flowchart of step S304 in Figure 6 Schematically shows Figure 3 The sub-step flowchart of step S306 in Figure 7 Schematically shows Figure 1 The sub-step flowchart of step S104 in Figure 8 Schematically shows the new flowchart of the annotation method for the target object according to Embodiment 1 of the present application; Figure 9 Is an application example diagram of the annotation method for the target object according to Embodiment 1 of the present application; Figure 10 Schematically shows the block diagram of the annotation device for the target object according to Embodiment 2 of the present application; and Figure 11 Schematically shows the hardware architecture diagram of the computer device according to Embodiment 3 of the present application. Detailed implementation manners

[0019] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative efforts shall fall within the scope of protection of the present application.

[0020] It should be noted that the descriptions involving "first", "second", etc. in the embodiments of the present application are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one of such features. In addition, the technical solutions between various embodiments can be combined with each other, but it must be based on the fact that those of ordinary skill in the art can implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present application.

[0021] In the description of the present application, it should be understood that the numerical labels before the steps do not identify the order of execution of the steps, but are only used to facilitate the description of the present application and distinguish each step, and thus cannot be understood as a limitation to the present application.

[0022] First, provide the term explanations involved in the present application: JSON (JavaScript Object Notation): A lightweight data interchange format that can be used for data storage and data transmission.

[0023] Graph structure: A data structure composed of nodes (or vertices) and edges, where the edges are used to connect different nodes.

[0024] Secondly, to facilitate the understanding of the technical solutions provided by the embodiments of the present application by those skilled in the art, the related technologies are described below: In the current live broadcast industry, the evaluation and annotation of live streamers are important components of platform operation. However, the applicant has learned that the related annotation technologies still have the following defects: (1) Limited to fixed metrics, it is difficult to adapt to the systematic assessment of different types of live streamers.

[0025] (2) The live streamer data of different modules have different data formats and access methods, and there are compatibility and efficiency problems in processing multi-channel data, resulting in low data management and processing efficiency.

[0026] (3) It is insufficient in terms of real-time performance and accuracy, and lacks flexibility and scalability, resulting in inaccurate annotation results, affecting the subsequent refined management of live streamers by the platform and the platform operation efficiency.

[0027] Therefore, the embodiments of the present application provide a technical solution for annotating target objects. In this technical solution: (1) A class of general data structures (such as a graph structure - annotation strategy graph represented by JSON) is designed to achieve configuration-based access to multi-channel live streamer data, achieve configuration of live streamer assessment materials and metrics, achieve orchestration of the annotation processes for various types of live streamers, and provide refined evaluation and annotation results. (2) Define and configure each different annotation node, edge, and related information (annotation rules and transfer conditions) in the annotation process through a graph structure language, design expressions (operation rules) to calculate the output (annotation results) of the annotation nodes, and determine the edges that meet the transfer requirements, so as to achieve automated annotation process transfer and achieve flexible process orchestration. (3) It is easy to expand, and the process change does not require re-development, optimizing the platform operation efficiency and improving the platform's refined management ability for live streamers. See the following for details.

[0028] The technical solutions of the present application are introduced below through multiple embodiments. It should be noted that these embodiments can be implemented in various different forms and should not be construed as being limited only to the embodiments described herein.

[0029] Embodiment 1 Figure 1 Schematically shows a flowchart of a method for annotating a target object according to Embodiment 1 of the present application.

[0030] As Figure 1 shown, the method for annotating a target object may include steps S100 to S104, where: Step S100, determine the type of the target object.

[0031] Step S102: Obtain a target annotation strategy graph from multiple pre-configured annotation strategy graphs based on the type of the target object. The annotation strategy graph includes multiple annotation nodes, each annotation node is associated with one or more edges, each edge is used to connect two annotation nodes, each annotation node is configured with a corresponding annotation rule, and each edge is configured with a corresponding transition condition.

[0032] Step S104: Starting from the first annotation node of the target annotation strategy graph, perform multiple rounds of transitions until reaching the last annotation node of the target annotation strategy graph to obtain a target annotation result. Each round of transition includes: determining a corresponding annotation result based on the annotation rule corresponding to the currently flowing-through annotation node, determining a target edge based on the transition condition of one or more edges associated with the corresponding annotation result, and flowing through the target edge to the next annotation node. The target annotation result includes the annotation results corresponding to each flowing-through annotation node.

[0033] The annotation method for the target object provided in this embodiment determines the type of the target object and obtains a target annotation strategy graph from multiple pre-configured annotation strategy graphs based on the type of the target object. The annotation strategy graph includes multiple annotation nodes, each annotation node is associated with one or more edges, and each edge is used to connect two annotation nodes. Each annotation node is further configured with a corresponding annotation rule, and each edge is further configured with a corresponding transition condition. Enter the first annotation node of the target annotation strategy graph and determine a corresponding annotation result based on the corresponding annotation rule. Based on the corresponding annotation result and the transition conditions corresponding to one or more associated edges, determine a target edge. Flow through the target edge to the next annotation node and determine the annotation result corresponding to the next annotation node for performing the next round of transition until reaching the last annotation node of the target annotation strategy graph to obtain a target annotation result. The target annotation result includes the annotation results corresponding to each flowing-through annotation node. It can be seen that in the embodiment of the present application, by defining and configuring various different annotation nodes, edges, and related messages (annotation rules, transition conditions) in the annotation process using graph structure language, an automated annotation process transition can be performed, the flexible orchestration of the annotation processes for various types of target objects can be realized, refined annotation results can be provided, and the operation efficiency can be optimized.

[0034] The following Figure 1 is used to elaborate in detail each step in steps S100 - S104 and other optional steps.

[0035] Step S100 , determine the type of the target object.

[0036] The target object can be the content to be annotated, such as: the host, products, videos, articles, etc. In the embodiments of the present application, the host will be used as the target object to exemplarily introduce the annotation method of the target object. The types of the target object can be divided according to actual business needs. For example, they can be divided into multiple types according to the field (entertainment, games, virtual, radio hosts, etc.), multiple types according to age, and multiple types according to the live broadcast duration (new hosts, mature hosts, etc.).

[0037] Step S102 , based on the type of the target object, obtain the target annotation strategy graph from multiple pre-configured annotation strategy graphs; wherein, the annotation strategy graph includes multiple annotation nodes, each annotation node is associated with one or more edges, each edge is used to connect two annotation nodes, each annotation node is configured with a corresponding annotation rule, and each edge is configured with a corresponding transfer condition.

[0038] Exemplarily, corresponding annotation strategies can be formulated in advance for each type of target object. The annotation strategy can include two aspects: process-related configuration and data-related configuration. Among them, the process-related configuration can include: the display name and annotation rule of each annotation node, the display name and transfer condition of each edge, etc., which are used to clarify the annotation process. The annotation rule can include data-related configuration, index-related configuration, etc., such as data source, data type, data format, data filtering logic, index, etc., which are used to determine the data and indexes required for annotation and perform annotation to obtain the annotation result. Converting the formulated annotation strategy into a graph structure can obtain the annotation strategy graph. As Figure 2 shown, the annotation strategy graph can include multiple annotation nodes, such as: "awaiting first review", "awaiting second review", "awaiting third review", "awaiting final review". Each annotation node is associated with one or more edges, and each edge can connect two annotation nodes. Exemplarily, except for the last annotation node "awaiting final review", each of the remaining annotation nodes can be associated with one or more directed edges to point to the next annotation node. For example: "awaiting second review" → "awaiting third review", "awaiting second review" → "awaiting final review". Each annotation node is configured with a corresponding annotation rule, and the annotation rule is used to obtain the data (materials) and indexes required for annotating the target object, and perform rating and annotation based on the data and indexes to obtain the annotation result corresponding to each annotation node. The annotation rules of each annotation node can be the same or different from those of other annotation nodes: in the case where the annotation rules are the same, it is equivalent to ensuring the accuracy and consistency of annotation through a review mechanism. In the case where the annotation rules are different, it is equivalent to performing multi-dimensional comprehensive review, which can improve the comprehensiveness and fineness of annotation. Each edge can be configured with a transfer condition, such as: "the ratings of the first review and the second review are inconsistent". In the case where the transfer condition is met, it can be transferred from the previous annotation node to the next annotation node through the edge, realizing the automated transfer of the annotation process.

[0039] Step S104 Starting from the first annotation node of the target annotation strategy graph, multiple rounds of circulation are carried out until the last annotation node of the target annotation strategy graph is reached, and a target annotation result is obtained; wherein, each round of circulation includes: determining the corresponding annotation result based on the annotation rule corresponding to the currently flowing-through annotation node, determining the target edge based on the circulation condition of one or more associated edges corresponding to the corresponding annotation result, and flowing to the next annotation node through the target edge; wherein, the target annotation result includes the annotation result corresponding to each flowing-through annotation node.

[0040] Exemplarily, in the case of obtaining a target annotation strategy graph that matches the anchor type, enter the first annotation node (such as Figure 2 "Pending First Review"). Based on the annotation rule corresponding to the first annotation node, the annotation result of the first annotation node can be determined. The annotation result is matched with the circulation condition of the edge associated with the first annotation node. In the case where the edge is not configured with special requirements (circulation condition), it can automatically flow to the next annotation node through the edge (that is, the second annotation node, such as Figure 2 "Pending First Review" → "Pending Second Review"). Similarly, based on the annotation rule corresponding to the second annotation node, the annotation result of the second annotation node can be determined. The annotation result is matched with the circulation condition of the edge associated with the second annotation node. As Figure 2 shown, the second annotation node is associated with two edges. The circulation condition on the left can be "The ratings of the first and second reviews are inconsistent" (that is, the annotation result of the first annotation node is inconsistent with the annotation result of the second annotation node), and the circulation condition on the right can be "The ratings of the first and second reviews are consistent" (not shown in the figure). Compare the annotation results of the two annotation nodes. In the case where the two are inconsistent, flow to the next annotation node through the left, that is, "Pending Third Review" is the third annotation node. In the case where the two are consistent, flow to the next node through the right, that is, "Pending Final Review" is the third annotation node. In the target annotation strategy graph, every time a annotation node is reached, the annotation result is determined according to the annotation rule corresponding to the currently flowing-through annotation node, combined with the annotation result and the circulation condition corresponding to the associated edge for judgment, and the next round of circulation is executed until the last annotation node of the target annotation strategy graph is reached (such as Figure 2 "Pending Final Review"), and finally the target annotation result can be obtained. The target annotation result can include the annotation result corresponding to each flowing-through annotation node.

[0041] In this embodiment, a general data structure (such as a graph structure - annotation strategy graph represented by JSON) is designed for various types of anchors to achieve configuration-based access to anchor data from multiple channels, achieve configuration of anchor assessment materials and indicators, achieve orchestration of the annotation processes for various types of anchors, and provide refined evaluation and annotation results.

[0042] In an alternative embodiment, the annotation rules may include pre-configured data processing rules, metric location rules, mapping rules, and operation rules. As Figure 3 shown, step S104 may include: Step S300, obtaining the target data of the target object according to the data processing rules.

[0043] Step S302, determining a plurality of target metrics according to the metric location rules.

[0044] Step S304, determining the rating of the target data of the target object under the plurality of target metrics to obtain a rating result.

[0045] Step S306, determining the annotation result based on the rating result, the mapping rules, and the operation rules.

[0046] Exemplarily, the annotation rules may include pre-configured data processing rules, metric location rules, mapping rules, and operation rules. Among them, the data processing rules can be used to specify from which data sources to obtain the data required for annotating the target object, the type, format, and filtering logic of the data, the production time of the data, etc., to meet the correctness and timeliness of the data statistics scope. The metric location rules can be used to initially annotate the target object (the host) through which metrics, that is, to determine the target metrics, such as: cover title, image, scene, presentation, interaction, manuscript, revenue, live broadcast duration, scarcity, etc. The mapping rules and operation rules can be used to process the preliminary annotation results (compatibility access and calculation) to obtain further annotation results. Specifically, according to the data processing rules, the target data of the target object, that is, the data to be annotated, can be obtained. According to the metric location rules, a plurality of target metrics can be determined. Determine the rating of the target data of the target object under the plurality of target metrics to obtain a rating result. For example: the "cover title" rating is "unqualified", the "live broadcast duration" rating is "excellent", and the scarcity rating is "qualified", etc. Based on the mapping rules, the rating result can be converted into a form applicable to the operation rules, so as to perform compatibility access, and finally the annotation result can be calculated according to the operation rules. It should be noted that the annotation rules of each annotation node can be different or the same, and the specific data processing rules, metric location rules, mapping rules, and operation rules can all be configured according to the actual annotation requirements.

[0047] In this embodiment, by configuring the annotation rules, the host data from multiple channels can be accessed compatibly, and the configuration of the host assessment materials and metrics at different annotation nodes can be realized, providing refined evaluation and annotation results.

[0048] In an alternative embodiment, the data processing rules may include data sources, data types, data formats, and data filtering logic. AsFigure 4 As shown, step S300 may include: Step S400, determining a target data source from multiple data sources based on the data processing rule.

[0049] Step S402, obtaining the initial data of the target object from the target data source.

[0050] Step S404, performing type conversion, format standardization, and / or data filtering on the initial data based on the data processing rule to obtain the target data of the target object.

[0051] Exemplarily, the data source may be various anchor data output systems, such as: room, barrage, interaction, gift system, etc. Analyze the data processing rule to determine which data sources to obtain anchor data from, that is, determine the target data source. Obtain the initial data of the target object from the target data source. Based on the data processing rule, perform type conversion, format standardization, and / or data filtering on the initial data to obtain the target data of the target object for subsequent annotation. In some embodiments, the data processing rule may also verify the timeliness of the initial data to filter out expired data.

[0052] In this embodiment, through the configured data processing rule, the anchor data can be quickly queried, obtained, and efficiently processed, and the multi-dimensional anchor data can be aggregated for subsequent annotation.

[0053] In an alternative embodiment, as Figure 5 shown, step S304 may include: Step S500, pushing the target data of the target object and the multiple target metrics to a rating component, where the rating component is used to provide a rating page, and the rating page is used to display the target data of the target object and the multiple target metrics; wherein, each of the target metrics is associated with multiple rating items.

[0054] Step S502, in response to selecting one of the rating items associated with each of the target metrics, determining the rating of the target data of the target object under the multiple target metrics to obtain the rating result.

[0055] Exemplarily, the target data and multiple target metrics can be pushed to the rating component. The rating component can be an internal module or an external system (such as a third-party scoring system). The rating component can provide a rating page, and the rating page can display the target data and multiple target metrics, and configure rating items for the multiple target metrics, such as "unqualified", "qualified", "excellent", etc. The annotator can select one of the rating items for each target metric on the rating component to determine the rating of the target data under each target metric and obtain the rating result. After the rating is completed, the rating component can return the rating result.

[0056] In this embodiment, the accurate rating of the target data is realized through the rating component, effectively improving the annotation efficiency and accuracy.

[0057] In some embodiments, methods such as visual analysis and data analysis can be combined to determine the rating of the data of the target object under each target index without manual intervention, further improving the annotation efficiency and quality.

[0058] In an alternative embodiment, the annotation result may include a first annotation result and a second annotation result. As Figure 6 shown, step S306 may include: Step S600, mapping the rating result to a first annotation result of a preset data structure based on the mapping rule.

[0059] Step S602, determining the second annotation result based on the first annotation result and the operation rule; wherein, the forms of expression of the first annotation result and the second annotation result are different.

[0060] Exemplarily, ratings such as "unqualified", "qualified", "excellent", etc. are converted into a general data structure through the mapping rule, such as multiple scores (e.g., 0 points, 5 points, 10 points, etc.). The operation rule may include steps such as summation and comparison. For example: summing the scores to obtain the total score. Among them, multiple scores or the total score can be used as the first annotation result to provide comparable annotation results. The total score can also be compared with a preset threshold to obtain the second annotation result. For example: greater than 22 points, the annotation result can be "SS"; greater than 17 points and less than 22 points, the annotation result can be "S"; greater than 15 points and less than 17 points, the annotation result can be "A", and in other cases the annotation result is "unqualified". The second annotation result can be classification labels such as "SS", "S", "A", "unqualified", which can provide intuitive and concise annotation results and reduce errors.

[0061] In this embodiment, by combining the mapping rule and the operation rule, it is possible to provide both fine (the first annotation result) and intuitive (the second annotation result), improving the accuracy and comprehensibility of the annotation and optimizing the platform operation efficiency.

[0062] In an alternative embodiment, each edge may be associated with a pre-configured weight. As Figure 7 shown, step S104 may include: Step S700, matching the corresponding annotation result with the transfer conditions corresponding to one or more associated edges, and determining the initial edge as the successfully matched edge.

[0063] Step S702: Determine a target edge based on the weight of the initial edge, where the target edge is used to perform the transfer of the annotation node.

[0064] Exemplarily, when flowing through an annotation node and calculating the corresponding annotation result, the annotation result can be matched with the transfer conditions corresponding to one or more edges associated with the annotation node, and the edges that meet the transfer conditions are initially determined as the initial edges. Based on the weight corresponding to each initial edge, the initial edge with the highest weight is selected and determined as the target edge for performing the transfer of the annotation node.

[0065] In this embodiment, by finding an edge that meets the transfer condition and has the highest weight from all the edges associated with the annotation node according to the annotation result of the annotation node as the target edge to perform the transfer, the automated transfer of the annotation process can be completed, the quality of the annotation can be improved, and the overall annotation efficiency and accuracy can be enhanced.

[0066] In an alternative embodiment, as Figure 8 shown, the annotation method for the target object may further include: Step S800: In the case of determining the annotation result corresponding to the currently flowing-through annotation node, generate an annotation message based on the type of the target object, the currently flowing-through annotation node, and the corresponding annotation result.

[0067] Step S802: Deliver the annotation message to a downstream system, where the downstream system is used to perform an interaction operation on the target object according to the annotation message.

[0068] Exemplarily, each time an annotation node is flowed through, after calculating the annotation result corresponding to the annotation node, an annotation message can be generated based on the annotation result, the annotation node, and the type of the target object. The annotation message is delivered to the downstream system, and the downstream system can view the current annotation node and annotation result of the target object (for example: first instance, second instance, or final instance), and perform an interaction operation on the target object. For example: the target object is currently in the third instance, and the annotation result is "SS", and the downstream system can provide certain creation incentives, live broadcast support, etc. for the target object, such as: pushing the live broadcast room of the target object to more audiences, providing training, live broadcast subsidies, etc. for the target object.

[0069] In this embodiment, by generating and delivering the annotation message in real time, the linkage between the annotation result and the downstream system can be realized, effectively improving the platform operation efficiency.

[0070] In some embodiments, information such as the annotation result calculated by each annotation node, the display name of the annotation node, and the type of the target anchor can also be saved to a preset database to achieve persistent storage, which is convenient for subsequent operation and call.

[0071] To make this application easier to understand, the following is combined with Figure 9 to provide an exemplary application.

[0072] S1: The policy group formulates annotation policies in advance, including two aspects: data and process. Data can be: data sources, data types, data formats, data filtering logics, etc. The process can be annotation nodes, such as display names, result calculation logics, associated information (annotation rules), etc., and edges, such as edge display names, edge requirements, result calculation (transfer conditions), etc.

[0073] S2: Convert the formulated annotation policy into a graph structure (annotation policy graph), including: points, edges, graphs, mapping rules, corresponding anchor types, etc.

[0074] S3: Operations use the annotation engine to submit various types of anchors to be annotated. According to the anchor type, find the corresponding annotation policy graph and enter the first annotation node. The annotation engine queries, aggregates multi-dimensional data based on the annotation rules corresponding to the annotation node and pushes it to the rating component.

[0075] S4: The annotator performs rating annotations on the rating component. After completion, the rating component pushes the content annotated by the annotator to the annotation engine.

[0076] S5: The annotation engine parses and processes the annotation content of the annotator into a general data structure according to the configured mapping rules, and calculates the annotation result of the anchor under this policy and this annotation node based on the operation rules configured in the first annotation node. Then find an edge that meets the requirements and has the highest weight among all the edges associated with the node and execute it to complete the automated annotation process flow until it reaches the last node of the annotation policy graph.

[0077] Among them, for each transferred annotation node, the annotation result of this annotation node can be calculated. Store the calculated annotation result, annotation details (such as target metrics, etc.), and information related to the annotation node in the persistent database, and send messages to other downstream systems for use.

[0078] In this exemplary application: (1) Designed a type of general data structure (such as a graph structure represented by JSON - annotation policy graph) to achieve configuration-based access to multi-channel anchor data, achieve configuration of anchor assessment materials and metrics, achieve orchestration of annotation processes for various types of anchors, and provide refined evaluation and annotation results. (2) Define and configure various different annotation nodes, edges, and related information (annotation rules and transfer conditions) in the annotation process through graph structure language, design expressions (operation rules) to calculate the output (annotation result) of the annotation node, and determine the edges that meet the transfer requirements to achieve automated annotation process flow and achieve flexible process orchestration. (3) Easy to expand, no need to re-develop for process changes, optimize the platform operation efficiency, and improve the platform's refined management ability for anchors.

[0079] Example 2 Figure 10 Schematically shows a block diagram of a labeling device for a target object according to Example 2 of the present application. The device can be divided into one or more program modules. One or more program modules are stored in a storage medium and executed by one or more processors to complete the embodiments of the present application. The program modules referred to in the embodiments of the present application refer to a series of computer program instruction segments that can complete specific functions. The following description will specifically introduce the functions of each program module in this embodiment. As Figure 10 shown, the device 1000 may include: a determination module 1100, an acquisition module 1200, and a labeling module 1300, where: The determination module 1100 is configured to determine the type of the target object; The acquisition module 1200 is configured to obtain a target labeling policy graph from a plurality of pre-configured labeling policy graphs based on the type of the target object; wherein, the labeling policy graph includes a plurality of labeling nodes, each labeling node is associated with one or more edges, each edge is used to connect two labeling nodes, each labeling node is configured with a corresponding labeling rule, and each edge is configured with a corresponding transition condition; The labeling module 1300 is configured to start from the first labeling node of the target labeling policy graph and perform multiple rounds of transitions until reaching the last labeling node of the target labeling policy graph to obtain a target labeling result; wherein, each round of transition includes: determining a corresponding labeling result based on the labeling rule corresponding to the currently flowing-through labeling node, determining a target edge based on the corresponding labeling result and the transition conditions of one or more associated edges, and transitioning to the next labeling node through the target edge; wherein, the target labeling result includes the labeling results corresponding to each flowing-through labeling node.

[0080] As an optional embodiment, the labeling rules include pre-configured data processing rules, index positioning rules, mapping rules, and operation rules; Correspondingly, determining a corresponding labeling result based on the corresponding labeling rule includes: Obtaining target data of the target object according to the data processing rule; Determining a plurality of target indicators according to the index positioning rule; Determining the rating of the target data of the target object under the plurality of target indicators to obtain a rating result; Determining the labeling result based on the rating result, the mapping rule, and the operation rule.

[0081] As an optional embodiment, the data processing rule includes a data source, a data type, a data format, and a data filtering logic; Correspondingly, according to the data processing rule, obtaining the target data of the target object includes: Determining a target data source from multiple data sources based on the data processing rule; Obtaining the initial data of the target object from the target data source; Based on the data processing rule, performing type conversion, format standardization, and / or data filtering on the initial data to obtain the target data of the target object.

[0082] As an optional embodiment, determining the rating of the target data of the target object under the multiple target metrics to obtain a rating result includes: Pushing the target data of the target object and the multiple target metrics to a rating component, where the rating component is used to provide a rating page for displaying the target data of the target object and the multiple target metrics; wherein, each of the target metrics is associated with multiple rating items; In response to selecting one of the rating items associated with each of the target metrics, determining the rating of the target data of the target object under the multiple target metrics to obtain the rating result.

[0083] As an optional embodiment, the annotation result includes a first annotation result and a second annotation result; correspondingly, based on the rating result, the mapping rule, and the operation rule, determining the annotation result includes: Mapping the rating result to a first annotation result in a preset data structure based on the mapping rule; Determining the second annotation result based on the first annotation result and the operation rule; Wherein, the presentation forms of the first annotation result and the second annotation result are different.

[0084] As an optional embodiment, each edge is associated with a pre-configured weight; correspondingly, based on the corresponding annotation result and the transfer condition corresponding to one or more associated edges, determining a target edge includes: Matching the corresponding annotation result with the transfer condition corresponding to one or more associated edges, and determining the successfully matched edge as the initial edge; Determining the target edge based on the weight of the initial edge, where the target edge is used to perform the transfer of the annotation node.

[0085] As an optional embodiment, the apparatus 1000 is further configured to: When determining the annotation result corresponding to the currently flowing-through annotation node, generating an annotation message based on the type of the target object, the currently flowing-through annotation node, and the corresponding annotation result; Deliver the annotation message to a downstream system, which is used to perform an interaction operation on the target object according to the annotation message.

[0086] Embodiment III Figure 11 Schematically shows a hardware architecture diagram of a computer device 10000 suitable for implementing the annotation method of a target object according to Embodiment III of the present application. In some embodiments, the computer device 10000 may be a terminal device such as a smart phone, a wearable device, a tablet computer, a personal computer, a vehicle-mounted terminal, a game console, a virtual device, a workbench, a digital assistant, a set-top box, a robot, etc. In other embodiments, the computer device 10000 may be a rack server, a blade server, a tower server or a cabinet server (including an independent server or a server cluster composed of multiple servers), etc. As Figure 11 shown, the computer device 10000 includes, but is not limited to: a memory 10010, a processor 10020, and a network interface 10030 that can be communicatively linked to each other through a system bus. Among them: The memory 10010 includes at least one type of computer-readable storage medium. The readable storage medium includes flash memory, a hard disk, a multimedia card, a card-type memory (such as an SD or DX memory), a random access memory (RAM), a static random access memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, an optical disk, etc. In some embodiments, the memory 10010 may be an internal storage module of the computer device 10000, such as the hard disk or memory of the computer device 10000. In other embodiments, the memory 10010 may also be an external storage device of the computer device 10000, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc. equipped on the computer device 10000. Of course, the memory 10010 may also include both the internal storage module and the external storage device of the computer device 10000. In this embodiment, the memory 10010 is generally used to store the operating system and various application software installed on the computer device 10000, such as the program code of the annotation method of the target object. In addition, the memory 10010 may also be used to temporarily store various types of data that have been output or will be output.

[0087] In some embodiments, the processor 10020 may be a Central Processing Unit (CPU), a controller, a microcontroller, a microprocessor, or other chips. The processor 10020 is generally used to control the overall operation of the computer device 10000, such as performing control and processing related to data interaction or communication with the computer device 10000. In this embodiment, the processor 10020 is used to run the program code stored in the memory 10010 or process data.

[0088] The network interface 10030 may include a wireless network interface or a wired network interface. The network interface 10030 is generally used to establish a communication link between the computer device 10000 and other computer devices. For example, the network interface 10030 is used to connect the computer device 10000 to an external terminal through a network, and establish a data transmission channel and a communication link between the computer device 10000 and the external terminal. The network can be an enterprise intranet (Intranet), the Internet, the Global System of Mobile communication (GSM for short), Wideband Code Division Multiple Access (WCDMA for short), 4G network, 5G network, Bluetooth, Wi-Fi and other wireless or wired networks.

[0089] It should be noted that Figure 11 Only the computer device with components 10010 - 10030 is shown, but it should be understood that it is not required to implement all the shown components, and more or fewer components can be alternatively implemented.

[0090] In this embodiment, the annotation method of the target object stored in the memory 10010 can also be divided into one or more program modules and executed by one or more processors (such as the processor 10020) to complete the embodiments of the present application.

[0091] Embodiment 4 The embodiments of the present application also provide a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the annotation method of the target object in the embodiments are implemented.

[0092] In this embodiment, the computer-readable storage medium includes flash memory, hard disks, multimedia cards, card-type memories (e.g., SD or DX memories, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memories, magnetic disks, optical disks, etc. In some embodiments, the computer-readable storage medium may be an internal storage unit of a computer device, such as the hard disk or memory of the computer device. In other embodiments, the computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc., equipped on the computer device. Of course, the computer-readable storage medium may also include both the internal storage unit and the external storage device of the computer device. In this embodiment, the computer-readable storage medium is generally used to store the operating system installed on the computer device and various application software, such as the program code of the annotation method for the target object in the embodiment. In addition, the computer-readable storage medium may also be used to temporarily store various data that have been output or will be output.

[0093] Embodiment 5 The embodiment of the present application further provides a computer program product, including a computer program, which when executed by a processor implements the method in the above embodiment.

[0094] Obviously, those skilled in the art should understand that the above-mentioned modules or steps of the embodiments of the present application can be implemented by a general-purpose computer device. They can be concentrated on a single computer device or distributed on a network composed of multiple computer devices. Optionally, they can be implemented by program codes executable by the computer device. Thus, they can be stored in a storage device and executed by the computer device. And in some cases, the steps shown or described can be executed in a different order from here, or they can be separately made into individual integrated circuit modules, or multiple modules or steps among them can be made into a single integrated circuit module to implement. In this way, the embodiments of the present application are not limited to any specific combination of hardware and software.

[0095] It should be noted that the above are only the preferred embodiments of the present application, and do not limit the patent protection scope of the present application accordingly. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of the present application, or directly or indirectly applied in other related technical fields, shall be equally included in the patent protection scope of the present application.

Claims

1. A method for labeling a target object, characterized in that: The method comprises: Determine the type of target object; Based on the type of the target object, a target annotation strategy graph is obtained from a plurality of pre-configured annotation strategy graphs; wherein the annotation strategy graph includes a plurality of annotation nodes, each annotation node is associated with one or more edges, each edge is used to connect two annotation nodes, each annotation node is configured with a corresponding annotation rule, and each edge is configured with a corresponding flow condition; Starting from the first annotation node of the target annotation strategy graph, multiple rounds of transfer are performed until the last annotation node of the target annotation strategy graph is transferred to obtain the target annotation result; wherein each round of transfer includes: determining the corresponding annotation result based on the annotation rule corresponding to the currently flowing annotation node, determining the target edge based on the corresponding annotation result and the transfer condition of one or more edges associated therewith, and transferring to the next annotation node through the target edge; The target labeling result includes the labeling result corresponding to each labeling node that flows through.

2. The method according to claim 1, characterized in that: The annotation rules include pre-configured data processing rules, indicator positioning rules, mapping rules, and operation rules; Correspondingly, determining a corresponding annotation result based on the corresponding annotation rule includes: According to the data processing rule, obtaining target data of the target object; According to the indicator positioning rule, a plurality of target indicators are determined; Determine the rating of the target data of the target object under the multiple target indicators to obtain a rating result; The labeling result is determined based on the rating result, the mapping rule and the operation rule.

3. The method according to claim 2, characterized in that The data processing rules include data source, data type, data format and data filtering logic; Correspondingly, according to the data processing rule, obtaining the target data of the target object includes: Based on the data processing rule, determining a target data source from a plurality of data sources; Acquire initial data of the target object from the target data source; Based on the data processing rules, type conversion, format standardization and / or data filtering are performed on the initial data to obtain target data of the target object.

4. The method according to claim 2, characterized in that: Determining the rating of the target data of the target object under the multiple target indicators to obtain a rating result includes: Pushing the target data of the target object and the multiple target indicators to a rating component, the rating component is used to provide a rating page, and the rating page is used to display the target data of the target object and the multiple target indicators; wherein each of the target indicators is associated with multiple rating items; In response to selecting one of the rating items associated with each of the target indicators, the rating of the target data of the target object under the multiple target indicators is determined to obtain the rating result.

5. The method according to claim 2, characterized in that: The annotation result includes a first annotation result and a second annotation result; correspondingly, based on the rating result, the mapping rule and the operation rule, determining the annotation result includes: Based on the mapping rule, mapping the rating result to a first annotation result of a preset data structure; Determining the second annotation result based on the first annotation result and the operation rule; The first labeling result and the second labeling result have different presentation forms.

6. The method according to claim 1, characterized in that Each edge is associated with a pre-configured weight; correspondingly, determining a target edge based on the corresponding annotation result and the flow conditions corresponding to the associated one or more edges includes: Matching the corresponding annotation result with the flow conditions corresponding to one or more associated edges, and determining the edge with successful matching as the initial edge; A target edge is determined based on the weight of the initial edge, and the target edge is used to execute the flow of the labeled node.

7. The method according to any one of claims 1 to 6, characterized in that: The method further comprises: In the case of determining the annotation result corresponding to the currently flowing annotation node, generating an annotation message based on the type of the target object, the currently flowing annotation node, and the corresponding annotation result; The annotation message is delivered to a downstream system, and the downstream system is used to perform an interactive operation on the target object according to the annotation message.

8. A target object labeling device, characterized in that: The device comprises: A determination module, used to determine the type of the target object; An acquisition module is used to acquire a target annotation strategy graph from a plurality of pre-configured annotation strategy graphs based on the type of the target object; wherein the annotation strategy graph includes a plurality of annotation nodes, each annotation node is associated with one or more edges, each edge is used to connect two annotation nodes, each annotation node is configured with a corresponding annotation rule, and each edge is configured with a corresponding flow condition; The labeling module is used to start from the first labeling node of the target labeling strategy graph, perform multiple rounds of transfer until it flows to the last labeling node of the target labeling strategy graph, and obtain the target labeling result; wherein each round of transfer includes: determining the corresponding labeling result based on the labeling rule corresponding to the labeling node currently flowing through, determining the target edge based on the corresponding labeling result and the transfer conditions of one or more edges associated with it, and transferring to the next labeling node through the target edge; The target labeling result includes the labeling result corresponding to each labeling node that flows through.

9. A computer device, characterized in that: include: at least one processor; and a memory communicatively connected to the at least one processor; wherein: The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the method according to any one of claims 1 to 7 is implemented.

11. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to claims 1 to 7 are implemented.

Images