Tag relationship establishment method, apparatus, system, medium, and computer program product

Abstract

The application discloses a label relationship establishing method, device, system, medium and computer program product, the method comprises: determining a sampling unit, and distributing the sampling unit on a target scene area; determining a model set corresponding to each sampling unit; for each model set, determining the position relationship between models and the label of each model, and generating the contact data between the labels according to the position relationship, the position relationship comprises adjacent position relationship and intersection position relationship, the contact data comprises adjacent contact data obtained from the adjacent position relationship and intersection contact data obtained from the intersection position relationship; establishing a label relationship according to the contact data corresponding to all model sets; the application can automatically mine the relationship between labels from the target scene area, avoid the influence of artificial subjective factors, and improve the accuracy and comprehensiveness of the label relationship.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to methods, apparatus, systems, media, and computer program products for establishing tag relationships. Background Technology

[0002] The placement of different items in a scene is often related; for example, chairs are usually placed next to tables, cups are placed on tables, and pillows are placed on beds. Establishing relationships between different categories of objects—that is, creating label relationships—can effectively improve development efficiency when building game scenes.

[0003] Currently, the construction of these label relationships relies on manual work, meaning that the relationships between each object model and other models in the scene are manually set during scene construction. This manual method of establishing label relationships is limited by personal subjective factors and knowledge levels, which can easily lead to situations where models that should appear together are not present, or where the relationship of a certain model is omitted, resulting in the inability to guarantee the accuracy and comprehensiveness of the label relationships.

[0004] Therefore, improving the accuracy and comprehensiveness of tag relationships is an urgent problem to be solved. Summary of the Invention

[0005] The main objective of this invention is to provide a method, apparatus, system, medium, and computer program product for establishing tag relationships, aiming to solve the problem of how to construct tag relationships accurately and comprehensively.

[0006] To achieve the above objectives, the present invention provides a method for establishing tag relationships, the method comprising the following steps:

[0007] Determine the sampling units and distribute them across the target scene area;

[0008] Determine the model set corresponding to each sampling unit;

[0009] For each model set, the positional relationships between models and the label of each model are determined, and the connection data between the labels is generated based on the positional relationships. The positional relationships include adjacent positional relationships and intersecting positional relationships, and the connection data includes adjacent connection data obtained from the adjacent positional relationships and intersecting connection data obtained from the intersecting positional relationships.

[0010] Establish label relationships based on the connection data corresponding to all model sets.

[0011] Optionally, the step of determining the model set corresponding to each sampling unit includes:

[0012] Collect the model at the corresponding position of each sampling unit, and determine the model preset corresponding to each sampling unit;

[0013] The models in the model pre-set are filtered to determine the model set corresponding to each sampling unit.

[0014] Optionally, the steps for determining the positional relationships between models include:

[0015] Based on the sampling units corresponding to the model set, the positional relationship between the models in the model set is determined as the adjacent positional relationship;

[0016] Obtain the bounding box corresponding to each model in the model set, and determine the positional relationship between the models in the model set as the intersection positional relationship based on the bounding box.

[0017] Optionally, the step of determining the positional relationship between models in the model set as the intersecting positional relationship based on the bounding box includes:

[0018] The bounding box is magnified according to a preset magnification factor, and it is determined whether the magnified bounding boxes intersect.

[0019] If the enlarged bounding boxes intersect, then the positional relationship between the models corresponding to the enlarged bounding boxes is determined as the intersecting positional relationship.

[0020] Optionally, the steps for establishing label relationships based on the connection data corresponding to all model sets include:

[0021] Determine frequent itemsets based on the connection data corresponding to all model sets;

[0022] Establish label relationships based on the frequent itemsets.

[0023] Optionally, the frequent itemsets include adjacent frequent itemsets determined by the adjacent relationship data, and the step of establishing label relationships based on the frequent itemsets includes:

[0024] Establish adjacency label relationships between labels in the adjacent frequent itemset.

[0025] Optionally, the frequent itemsets include the intersecting frequent itemsets determined by the intersecting relationship data, and the step of establishing label relationships based on the frequent itemsets includes:

[0026] In the set of frequently intersecting items, identify tags with the top-level placement attribute;

[0027] For each tag with the top-level placement attribute, establish intersection tag relationships between it and other tags.

[0028] Optionally, the step of establishing label relationships based on the frequent itemsets includes:

[0029] For tags that do not have the attribute of being placed at the top level, establish adjacency relationships between tags.

[0030] Furthermore, to achieve the above objectives, the present invention also provides a tag relationship establishment device, the tag relationship establishment device comprising:

[0031] The first determining module is used to determine the sampling units and distribute the sampling units on the target scene area;

[0032] The second determining module is used to determine the model set corresponding to each sampling unit;

[0033] The third determining module is used to determine the positional relationship between models and the label of each model for each model set, and to generate connection data between the labels according to the positional relationship. The positional relationship includes adjacent positional relationship and intersecting positional relationship, and the connection data includes adjacent connection data obtained from the adjacent positional relationship and intersecting connection data obtained from the intersecting positional relationship.

[0034] The fourth module is used to establish label relationships based on the connection data corresponding to all model sets.

[0035] Furthermore, the second determining module further includes a filtering module, the filtering module being used for:

[0036] Collect the model at the corresponding position of each sampling unit, and determine the model preset corresponding to each sampling unit;

[0037] The models in the model pre-set are filtered to determine the model set corresponding to each sampling unit.

[0038] Furthermore, the second determining module is also used for:

[0039] Based on the sampling units corresponding to the model set, the positional relationship between the models in the model set is determined as the adjacent positional relationship;

[0040] Obtain the bounding box corresponding to each model in the model set, and determine the positional relationship between the models in the model set as the intersection positional relationship based on the bounding box.

[0041] Furthermore, the third determining module is also used for:

[0042] The bounding box is magnified according to a preset magnification factor, and it is determined whether the magnified bounding boxes intersect.

[0043] If the enlarged bounding boxes intersect, then the positional relationship between the models corresponding to the enlarged bounding boxes is determined as the intersecting positional relationship.

[0044] Furthermore, the fourth determining module is also used for:

[0045] Determine frequent itemsets based on the connection data corresponding to all model sets;

[0046] Establish label relationships based on the frequent itemsets.

[0047] Furthermore, the fourth determining module is also used for:

[0048] Establish adjacency label relationships between labels in the adjacent frequent itemset.

[0049] Furthermore, the fourth determining module is also used for:

[0050] In the set of frequently intersecting items, identify tags with the top-level placement attribute;

[0051] For each tag with the top-level placement attribute, establish intersection tag relationships between it and other tags.

[0052] Furthermore, the fourth determining module is also used for:

[0053] For tags that do not have the attribute of being placed at the top level, establish adjacency relationships between tags.

[0054] In addition, to achieve the above objectives, the present invention also provides a tag relationship establishment system, the tag relationship establishment system comprising: a memory, a processor, and a tag relationship establishment program stored in the memory and executable on the processor, wherein the tag relationship establishment program, when executed by the processor, implements the steps of the tag relationship establishment method as described above.

[0055] In addition, to achieve the above objectives, the present invention also provides a medium, which is a computer-readable storage medium, on which a tag relationship establishment program is stored, and when the tag relationship establishment program is executed by a processor, it implements the steps of the tag relationship establishment method as described above.

[0056] In addition, to achieve the above objectives, the present invention also provides a computer program product, which includes a tag relationship establishment program, wherein when the tag relationship establishment program is executed by a processor, it implements the steps of the tag relationship establishment method as described above.

[0057] The label relationship establishment method proposed in this invention involves: determining sampling units and distributing these units across a target scene region; determining a model set corresponding to each sampling unit; for each model set, determining the positional relationships between models and the label for each model; generating connection data between the labels based on the positional relationships, where the positional relationships include adjacent positional relationships and intersecting positional relationships; and the connection data includes adjacent connection data obtained from the adjacent positional relationships and intersecting connection data obtained from the intersecting positional relationships. Label relationships are then established based on the connection data corresponding to all model sets. This invention can automatically mine the relationships between labels from a target scene region, avoiding the influence of subjective human factors and improving the accuracy and comprehensiveness of label relationships. Attached Figure Description

[0058] Figure 1 This is a schematic diagram of the device structure of the hardware operating environment involved in the embodiments of the present invention;

[0059] Figure 2 This is a flowchart illustrating the first embodiment of the tag relationship establishment method of the present invention;

[0060] Figure 3 This is a flowchart illustrating the second embodiment of the tag relationship establishment method of the present invention;

[0061] Figure 4 This is a flowchart illustrating the third embodiment of the tag relationship establishment method of the present invention;

[0062] Figure 5 This is a flowchart illustrating the fourth embodiment of the tag relationship establishment method of the present invention.

[0063] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0064] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0065] like Figure 1 As shown, Figure 1 This is a schematic diagram of the device structure of the hardware operating environment involved in the embodiments of the present invention.

[0066] The device in this embodiment of the invention can be a PC or a server.

[0067] like Figure 1As shown, the device may include: a processor 1001, such as a CPU; a network interface 1004; a user interface 1003; a memory 1005; and a communication bus 1002. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen or an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface). The memory 1005 may be high-speed RAM or non-volatile memory, such as a disk drive. Optionally, the memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0068] Those skilled in the art will understand that Figure 1 The device structure shown does not constitute a limitation on the device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0069] like Figure 1 As shown, the memory 1005, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and a tag relationship establishment program.

[0070] The operating system is a program that manages and controls portable storage devices and software resources, and supports the operation of the network communication module, user interface module, tag relationship establishment program, and other programs or software; the network communication module is used to manage and control the network interface 1002; and the user interface module is used to manage and control the user interface 1003.

[0071] exist Figure 1 In the storage device shown, the storage device calls the tag relationship establishment program stored in the memory 1005 through the processor 1001 and executes the operations in the various embodiments of the tag relationship establishment method described below.

[0072] Based on the above hardware structure, an embodiment of the label relationship establishment method of the present invention is proposed.

[0073] Reference Figure 2 , Figure 2 This is a flowchart illustrating the first embodiment of the label relationship establishment method of the present invention. This label relationship establishment method is applied to a label relationship establishment system, which can be applied to PC devices, mobile terminals, and other smart terminals. This label relationship establishment system can be used to determine the relationships between labels of models in a game scene. For ease of description, the label relationship establishment system is used as an example. The method may include:

[0074] Step S10: Determine the sampling units and distribute the sampling units on the target scene area;

[0075] It should be noted that a sampling unit is a spatial unit used to collect information within a target scene area.

[0076] The parameters of the sampling unit include its size and shape. These parameters can be manually set according to the target scene area. For example, the size of the sampling unit can be manually set according to the model scale of the target scene area; similarly, the shape of the sampling unit can be manually set according to the target scene area. The sampling unit can be set as a cuboid, cube, etc.

[0077] The target scene area is a 3D game scene. The target scene area is typically composed of different models and their relative positions. For example, a 3D study game scene could include game models such as bookshelves, tables, chairs, and books, as well as the relative positions of these models.

[0078] In this embodiment, the label relationship establishment system can first generate corresponding sampling units based on relevant parameters set manually according to the target scene area, and distribute the sampling units on the target scene area until the sampling units completely cover the target scene area, so as to facilitate the sampling of models in the target scene area and avoid the omission of models during sampling.

[0079] Furthermore, after generating sampling units, the label relationship establishment system can distribute the sampling units on the target scene area according to preset distribution rules, so that the sampling units can completely cover the target scene area, avoid model omissions during sampling, and ensure the comprehensiveness and accuracy of the subsequent determination of the relationship between the labels of the model.

[0080] For example, the preset distribution rule is to distribute the sampling units in the form of a spatial tree.

[0081] Step S20: Determine the model set corresponding to each sampling unit;

[0082] In this embodiment, after the label relationship establishment system distributes the sampling units on the target scene area, each sampling unit covers different areas of the target scene area. Various models exist in different areas, thereby determining the set of models collected by each sampling unit.

[0083] Step S30: For each model set, determine the positional relationship between models and the label of each model, and generate connection data between the labels according to the positional relationship. The positional relationship includes adjacent positional relationship and intersecting positional relationship. The connection data includes adjacent connection data obtained from the adjacent positional relationship and intersecting connection data obtained from the intersecting positional relationship.

[0084] It should be noted that a tag can be understood as a category. For example, 10 different table models are all categorized under the same category, "table", meaning that the tag for all of them is "table".

[0085] The positional relationships between models include adjacent positional relationships and intersecting positional relationships. An adjacent positional relationship occurs when two models are next to each other, meaning one model is located around the other. For example, a chair and a table are adjacent. An intersecting positional relationship occurs when two models are in contact with each other. For example, a water glass is located on a tabletop; they are in an intersecting positional relationship.

[0086] The relationship data uses labels to record the positional relationships between models of different categories. Relationship data includes adjacent relationship data and intersecting relationship data. That is, if the positional relationship between models is adjacent, the corresponding labels are recorded as adjacent relationship data; if the positional relationship between models is intersecting, the corresponding labels are recorded as intersecting relationship data.

[0087] In this embodiment, the label relationship establishment system can identify each model in the model set collected by each sampling unit, determine the label of each model, and determine the positional relationship between the models in the model set collected by each sampling unit. The label relationship establishment system can generate corresponding connection data based on the positional relationship between the models and the labels corresponding to the models.

[0088] The positional relationships between models include adjacent and intersecting relationships, and the connection data includes adjacent connection data and intersecting connection data. For models with adjacent positional relationships, adjacent connection data is generated based on the model's corresponding labels. For models with intersecting positional relationships, intersecting connection data is generated based on the model's corresponding labels.

[0089] For example, a sampling unit collects three different styles of table models, two different styles of chair models, one telephone model, and two book models. The label relationship establishment system can analyze the models collected by the sampling unit through the analysis engine to determine the object type corresponding to each model. The object type of the three different styles of table models is "table", the object type of the two different styles of chair models is "chair", the object type of the telephone model is "telephone", and the object type of the two book models is "book". Thus, the label of each model is determined. That is, the labels of the models collected by the sampling unit include "table", "chair", "telephone", and "book".

[0090] It is understandable that there are many ways to determine a tag. It can be obtained by directly reading the built-in attributes, by feature recognition, or by other means, which will not be limited here.

[0091] The label relationship establishment system identifies the placement of models labeled "table", "chair", "telephone" and "book" in the sampling unit, and determines the positional relationship between the models labeled "table", "chair", "telephone" and "book" as either adjacent or intersecting.

[0092] The label relationship establishment system generates adjacent relationship data based on the labels corresponding to the adjacent position relationship model, and generates intersecting relationship data based on the labels corresponding to the intersecting position relationship model.

[0093] Step S40: Establish label relationships based on the connection data corresponding to all model sets.

[0094] In this embodiment, after the tag relationship establishment system determines the connection data corresponding to each model set, it can establish tag relationships based on the connection data.

[0095] Tag relationships can be used to assist in building game scenes. For example, during the process of building a game scene, after placing a table model, the tag relationships established above can be used to determine other tags associated with the table model's corresponding tag. Then, models under other tag names can be recommended to the user or automatically placed in the scene, improving the efficiency of building game scenes.

[0096] In this embodiment, the label relationship establishment system determines sampling units and distributes them across the target scene area; it determines the model set corresponding to each sampling unit; for each model set, it determines the positional relationships between models and the label of each model, and generates connection data between labels based on the positional relationships. The positional relationships include adjacent positional relationships and intersecting positional relationships, and the connection data includes adjacent connection data obtained from adjacent positional relationships and intersecting connection data obtained from intersecting positional relationships; based on the connection data corresponding to all model sets, it establishes the relationships between labels. This invention can automatically mine the relationships between labels from the target scene area, avoiding the influence of subjective human factors and improving the accuracy and comprehensiveness of label relationships.

[0097] Further, refer to Figure 3 Based on the first embodiment of the label relationship establishment method of the present invention, a second embodiment of the label relationship establishment method of the present invention is proposed.

[0098] The difference between the second embodiment of the label relationship establishment method and the first embodiment of the label relationship establishment method is that the step of determining the model set corresponding to each sampling unit may include: sub-step S201 and sub-step S202;

[0099] Sub-step S201: Collect the model at the corresponding position of each sampling unit and determine the model pre-set corresponding to each sampling unit;

[0100] It should be noted that the model pre-set refers to all models at the corresponding positions on the target scene area collected by each sampling unit.

[0101] In this embodiment, after the label relationship establishment system distributes sampling units across the target scene area, it can collect models at corresponding locations within the target scene area using each sampling unit, thus determining the model pre-set corresponding to each sampling unit. During the sampling process, if a portion of the model appears at the location of a sampling unit, then the model pre-set corresponding to that sampling unit may include that model.

[0102] It is understandable that there may be large-volume models in the target scene area, such as buildings. The positional relationship between these large-volume models and other models is usually not meaningful for data mining, but the model pre-assembly may contain such volumetric models.

[0103] Sub-step S202: Filter the models in the model pre-set to determine the model set corresponding to each sampling unit.

[0104] In this embodiment, the label relationship establishment system can filter the models in the model pre-set collected by each sampling unit to determine the model set corresponding to each sampling unit.

[0105] Specifically, the label relationship establishment system can first determine the number of times each model in the target scene area is sampled by different sampling units, and compare this number with a preset number. If the number is less than the preset number, the corresponding model is retained. If the number is greater than or equal to the preset number, the model can be removed from the corresponding model pre-set and will not participate in subsequent processing.

[0106] Furthermore, if the quantity is greater than or equal to the preset quantity, it can also identify whether the corresponding model has the top-level placement attribute. If it is determined that the model has the top-level placement attribute, the model is saved. If it is determined that the model does not have the top-level placement attribute, the model is filtered out and deleted, thereby determining the model set corresponding to each sampling unit.

[0107] Understandably, the preset number can be determined based on the number and size of sampling units distributed in the target scene area and pre-set in the label relationship establishment system. If the sampling units are small, a table model may appear in multiple sampling units simultaneously. In other words, setting the preset number without considering the actual number and size of sampling units will not achieve a filtering effect. Models deemed unnecessary by the user, such as houses, trees, beds, and cars, can be removed by manually setting the preset number. By setting the preset number, the number of labels participating in the calculation can be controlled, excluding labels that are not intended to participate in relationship establishment.

[0108] For example, if the preset number is set to 5, the label relationship establishment system determines that there is a model that has been collected by more than 5 sampling units. If the model is determined to be meaningless for mining or is considered unnecessary by the user, it will be removed from the corresponding model pre-set and the model set corresponding to each sampling unit will be determined.

[0109] The label relationship establishment system in this embodiment first determines the model pre-set corresponding to each sampling unit, then filters the models in the model pre-set to determine the model set corresponding to each sampling unit, and filters models that do not constitute a general positional relationship with other models, thereby avoiding the introduction of interference terms, improving computational efficiency, and improving the accuracy and comprehensiveness of label relationships.

[0110] Further, refer to Figure 4 Based on the first and second embodiments of the label relationship establishment method of the present invention, a third embodiment of the label relationship establishment method of the present invention is proposed.

[0111] The difference between the third embodiment of the label relationship establishment method and the first and second embodiments of the label relationship establishment method is that the step of determining the positional relationship between models in the model set may include: sub-step S301 and sub-step S302;

[0112] Sub-step S301: Based on the sampling units corresponding to the model set, determine the positional relationship between the models in the model set as the adjacent positional relationship;

[0113] Sub-step S302: Obtain the bounding box corresponding to each model in the model set, and determine the positional relationship between the models in the model set as the intersection positional relationship based on the bounding box.

[0114] It should be noted that the bounding box corresponding to the model is usually created during the game generation process for collision detection.

[0115] In sub-steps S301 and S302, the positional relationship between each model in each model set can be determined as an adjacent positional relationship.

[0116] Furthermore, for each model in the model set, its corresponding bounding box can be obtained, and it can be determined whether the bounding boxes of the models in the set intersect. If there is an intersection between the bounding boxes, it can be determined that the corresponding models are in contact with each other, and the positional relationship between the contacting models can be determined as the intersection positional relationship.

[0117] For example, the model set collected by a certain sampling unit includes six models: abcdef. First, the label relationship establishment system determines the positional relationship between the abcdef models as an adjacent positional relationship. Then, the label relationship establishment system determines the bounding boxes corresponding to the abcdef models respectively and determines whether the bounding boxes between the models intersect. If there is an intersection between the bounding boxes of the bcef models, where bc may intersect directly, ce may intersect directly, ef may intersect directly, and bf may intersect indirectly, but they can all be regarded as intersecting. The label relationship establishment system determines the positional relationship between the bcef models as an intersecting positional relationship.

[0118] It is understandable that a sampling unit is a spatial unit used to collect information. It not only samples the models but also samples the positional relationships between the models. The positional relationships between the models collected by the sampling unit can be determined as adjacent positional relationships. Direct intersection means that the bounding boxes of two models intersect, while indirect intersection means that three models intersect sequentially. Two models at the edge intersect indirectly. For example, if bc intersects directly and ce intersects directly, then be intersects indirectly.

[0119] Furthermore, sub-step S302 may include sub-step S3021 and sub-step S3022.

[0120] Sub-step S3021: Magnify the bounding box according to a preset magnification factor, and determine whether the magnified bounding boxes intersect.

[0121] In this embodiment, the label relationship establishment system can, after determining the bounding box corresponding to each model, magnify the bounding box of each model according to a preset magnification factor, and determine whether there is an intersection between the magnified bounding boxes;

[0122] It should be noted that the preset magnification is a hyperparameter that can be set by the user, and it can usually be set to 1.05.

[0123] Sub-step S3022: If the enlarged bounding boxes intersect, determine the positional relationship between the models corresponding to the enlarged bounding boxes as the intersecting positional relationship.

[0124] In this embodiment, if the label relationship establishment system determines that the enlarged bounding boxes intersect, it can determine that the models corresponding to the enlarged bounding boxes are in contact with each other, and the positional relationship between the in-contact models can be determined as an intersecting positional relationship.

[0125] For example, a sampling unit collects a set of models including six types: abcdef. First, the label relationship establishment system can determine the positional relationship between the abcdef models as an adjacent positional relationship. Then, the label relationship establishment system determines the bounding boxes corresponding to the abcdef models respectively, and enlarges the bounding boxes by a preset magnification factor of 1.05. It then determines whether the enlarged bounding boxes intersect. If there is an intersection between the enlarged bounding boxes corresponding to the bcef models, where bc intersects directly, ce intersects directly, ef intersects directly, and bf intersects indirectly, but can be uniformly regarded as intersecting, the label relationship establishment system determines the positional relationship between the bcef models as an intersecting positional relationship.

[0126] The label relationship establishment system in this embodiment magnifies the bounding box of each model according to a preset magnification factor, determines whether the magnified bounding boxes intersect, and then determines the positional relationship between the models, thereby improving the accuracy of determining the positional relationship between the models and thus improving the accuracy and comprehensiveness of the label relationship.

[0127] Further, refer to Figure 5 Based on the first to third embodiments of the label relationship establishment method of the present invention, a fourth embodiment of the label relationship establishment method of the present invention is proposed.

[0128] The fourth embodiment of the tag relationship establishment method differs from the first to third embodiments in that the step of determining the tag relationship based on the contact data may include sub-step S401 and sub-step S402.

[0129] Sub-step S401: Determine the frequent itemset based on the connection data;

[0130] The relationships presented by intersecting and adjacent data may be accidental and not true pairings. If label relationships are generated directly from the data, the relationships between the labels may not be strong or universal and may not be accurate. To address this, we can identify frequently occurring label pairings.

[0131] In this embodiment, after determining the connection data between labels, the label relationship establishment system uses adjacent connection data and intersecting connection data. The adjacent connection data stores the labels corresponding to models with adjacent positional relationships, and the intersecting connection data stores the labels corresponding to models with intersecting positional relationships. The label relationship establishment system can obtain adjacent frequent itemsets based on the adjacent connection data and intersecting frequent itemsets based on the intersecting connection data.

[0132] In one embodiment, the tag relationship establishment system can obtain the number of tags and the frequency of each tag in the connection data, and based on the number of tags and the frequency of each tag, combine pre-established frequent rules and frequency confidence to filter out frequent itemsets in the connection data, thereby obtaining adjacent frequent itemsets based on adjacent connection data and intersecting frequent itemsets based on intersecting connection data.

[0133] In another embodiment, the label relationship establishment system can also establish corresponding frequent pattern trees, item header tables, and node lists based on adjacent relationship data, and based on intersecting relationship data. The label relationship establishment system mines sequentially upwards from the bottom item of the item header table to determine the conditional pattern base for each item in the frequent pattern tree corresponding to the item header table. The conditional pattern base is the frequent pattern tree corresponding to the currently mined label as the leaf node. The count of each node in the frequent pattern tree is set to the count of the leaf node, and nodes with counts lower than the support are deleted. From this conditional pattern base, the corresponding frequent itemsets can be recursively mined. The label relationship establishment system can then obtain adjacent frequent itemsets based on adjacent relationship data and intersecting frequent itemsets based on intersecting relationship data.

[0134] For example, a tag relationship establishment system can obtain the number of tags and the frequency of each tag in the contact data, and based on the number of tags and the frequency of each tag, combine pre-established frequency rules and frequency confidence levels to filter out frequent itemsets from the contact data. Frequent rules refer to the combinations of all tags in the target scene area, i.e., common combinations. The number of models is denoted by n, where n is the number of classification labels corresponding to all models within the target scene area, and n>1. Frequency confidence refers to the probability of a pairing between two models. It can be understood as a pattern. For example, the pairing of tables and chairs is very common in life, and they are likely to appear together in the same scene area. Therefore, the frequency confidence of manually setting up tables and chairs is relatively high. Similarly, the pairing of tables and cars may occur in some rare cases. Therefore, the frequency confidence of manually setting up tables and cars is high, but not zero.

[0135] The label relationship establishment system can determine frequent itemsets in the connection data based on frequency rules. For example, if there are six models in the target scene area, three of which are different styles of tables (corresponding to a "table" label), and the other three are different styles of chairs (corresponding to a "chair" label), then the connection data for the target scene area includes both "table" and "chair" labels. The label relationship establishment system can determine the common tags for these two types of labels based on frequency rules. Placement scenarios: The scene contains only tables, the scene contains only chairs, and the scene contains both tables and chairs.

[0136] The tag relationship establishment system determines the pairing of tags in the target scene area based on the connection data. It determines the frequency of each pairing based on the frequency of each tag, thereby obtaining the frequency confidence of the pairing. Then, it compares the frequency confidence with a preset frequency confidence. Pairings between tags with a frequency confidence greater than or equal to the preset frequency confidence are added to the frequent itemset, thus obtaining the adjacent frequent itemset corresponding to adjacent connection data and the intersecting frequent itemset corresponding to intersecting connection data.

[0137] For example, with a preset frequency confidence level of 25%, 10 sampling units are used to completely cover the target scene area. If both label a and label b are collected by 10 sampling units, then the frequency confidence level of label a and label b is 100%, which is greater than the preset frequency confidence level of 25%. Therefore, ab is added as a subset to the frequent itemset. If label a and label b are collected by only 1 sampling unit, then the frequency confidence level of label a and label b is 10%, which is less than the preset frequency confidence level of 25%. Therefore, the ab pairing is not added to the frequent itemset. Pairings between labels with a frequency confidence level greater than or equal to the preset frequency confidence level in adjacent relationship data are added to the adjacent frequent itemset. Pairings between labels with a frequency confidence level greater than or equal to the preset frequency confidence level in intersecting relationship data are added to the intersecting frequent itemset. Sub-step S302: Establish label relationships based on the frequent itemsets.

[0138] In this embodiment, after determining frequent itemsets and adjacent frequent itemsets, the label relationship establishment system determines and establishes label relationships based on these frequent itemsets. It is understood that the combinations of labels within adjacent frequent itemsets and overlapping frequent itemsets occur frequently.

[0139] Specifically, sub-step S402 may include the following sub-steps:

[0140] Sub-step S4021: Establish adjacency label relationships between labels in the adjacent frequent itemset;

[0141] Label relationships can include adjacency relationships. An adjacency relationship means that a model corresponding to one label is placed around a model corresponding to another label. For example, a chair is placed next to a table. Establishing adjacency relationships allows us to leverage the proximity of models to help construct game scenes.

[0142] In practical implementation, models in the same sampling unit set can be considered to have spatial connections, can be uniformly regarded as adjacent positional relationships, and adjacency relationships can be established for labels in adjacent frequent itemsets.

[0143] Sub-step S402 may also include the following sub-steps:

[0144] Sub-step S4022: In the set of frequently intersecting items, determine the tags with the top-level placement attribute;

[0145] Sub-step S4023: For each tag with the top-level placement attribute, establish intersecting tag relationships between it and other tags.

[0146] It should be noted that the top-level placement attribute means that the top surface of an object can be used to place another object. For example, a table can hold objects such as cups and telephones. The table has the top-level placement attribute, and the corresponding table tag supports adding the top-level placement attribute.

[0147] Tag relationships can include intersecting tag relationships. An intersecting tag relationship involves placing the model corresponding to another tag (aside from itself, tags with top-level placement attributes, and tags without top-level placement attributes) onto the model corresponding to the tag with top-level placement attributes. Establishing intersecting relationships leverages the placement characteristics between models to aid in constructing game scenes.

[0148] In intersecting relational data, the relationship between the models corresponding to the labels can be either horizontal or vertical. Horizontal intersection means that the models corresponding to two labels are placed side-by-side on the same plane. Vertical intersection means that the model corresponding to one label is placed on the top surface of the model corresponding to another label.

[0149] In a specific implementation, tags with top-level placement attributes can be identified in the frequently intersecting itemset. Then, for each tag with top-level placement attributes, tags in the frequently intersecting itemset that do not have top-level placement attributes or other tags that have placement attributes become top-level placement subordinate tags associated with the tag with top-level placement attributes. That is, an intersecting tag relationship is established between other tags and tags with top-level placement attributes.

[0150] Sub-step S402 may also include the following sub-steps:

[0151] Sub-step S4024: For tags that do not have the top-level placement attribute, establish adjacency tag relationships.

[0152] In a frequently intersecting itemset, the models corresponding to labels that do not have the top-level placement attribute may be adjacent. For example, books and cups placed on a table.

[0153] In some possible implementations, adjacency relationships can be established between tags that do not have a top-level placement attribute in a set of frequently intersecting items.

[0154] For example, the set of frequently intersecting items includes the tags abcdef, where tag a is a table, tag c is a box, and therefore tags a and c are tags with top-level placement attributes. Tag bdefg is a small item such as a telephone or a toy, and therefore tag bdefg is a tag without top-level placement attributes. The tag relationship establishment system identifies tag bcdefg as a top-level placement subordinate tag associated with tag a, and then establishes an intersecting tag relationship between tag bcdefg and tag a. Similarly, it establishes an intersecting tag relationship between tag abdefg and tag c. For tags bdefg that do not have top-level placement attributes, the tag relationship establishment system establishes adjacency tag relationships between tags bdefg.

[0155] The label relationship establishment system in this embodiment determines frequent itemsets based on connection data; it determines the relationship between labels based on frequent itemsets; it avoids the problem of reducing the comprehensiveness of the relationship between labels by determining the relationship between labels based on the combination of models that appear less frequently in the target scene area, thereby improving the accuracy and comprehensiveness of the label relationship.

[0156] The present invention also provides a tag relationship establishment device. The tag relationship establishment device of the present invention includes:

[0157] The first determining module is used to determine the sampling units and distribute the sampling units on the target scene area;

[0158] The second determining module is used to determine the model set corresponding to each sampling unit;

[0159] The third determining module is used to determine the positional relationship between models and the label of each model for each model set, and to generate connection data between the labels according to the positional relationship. The positional relationship includes adjacent positional relationship and intersecting positional relationship, and the connection data includes adjacent connection data obtained from the adjacent positional relationship and intersecting connection data obtained from the intersecting positional relationship.

[0160] The fourth module is used to establish label relationships based on the connection data corresponding to all model sets.

[0161] Furthermore, the second determining module further includes a filtering module, the filtering module being used for:

[0162] Collect the model at the corresponding position of each sampling unit, and determine the model preset corresponding to each sampling unit;

[0163] The models in the model pre-set are filtered to determine the model set corresponding to each sampling unit.

[0164] Furthermore, the second determining module is also used for:

[0165] Based on the sampling units corresponding to the model set, the positional relationship between the models in the model set is determined as the adjacent positional relationship;

[0166] Obtain the bounding box corresponding to each model in the model set, and determine the positional relationship between the models in the model set as the intersection positional relationship based on the bounding box.

[0167] Furthermore, the third determining module is also used for:

[0168] The bounding box is magnified according to a preset magnification factor, and it is determined whether the magnified bounding boxes intersect.

[0169] If the enlarged bounding boxes intersect, then the positional relationship between the models corresponding to the enlarged bounding boxes is determined as the intersecting positional relationship.

[0170] Furthermore, the fourth determining module is also used for:

[0171] Determine frequent itemsets based on the connection data corresponding to all model sets;

[0172] Establish label relationships based on the frequent itemsets.

[0173] Furthermore, the fourth determining module is also used for:

[0174] Establish adjacency label relationships between labels in the adjacent frequent itemset.

[0175] Furthermore, the fourth determining module is also used for:

[0176] In the set of frequently intersecting items, identify tags with the top-level placement attribute;

[0177] For each tag with the top-level placement attribute, establish intersection tag relationships between it and other tags.

[0178] Furthermore, the fourth determining module is also used for:

[0179] For tags that do not have the attribute of being placed at the top level, establish adjacency relationships between tags.

[0180] The present invention also provides a tag relationship establishment system.

[0181] The tag relationship establishment system includes: a memory, a processor, and a tag relationship establishment program stored in the memory and executable on the processor. When the tag relationship establishment program is executed by the processor, it implements the steps of the tag relationship establishment method as described above.

[0182] The method implemented when the tag relationship establishment program running on the processor is executed can be referred to in various embodiments of the tag relationship establishment method of the present invention, and will not be repeated here.

[0183] The present invention also provides a medium.

[0184] The medium is a computer-readable storage medium that stores a tag relationship establishment program, which, when executed by a processor, implements the steps of the tag relationship establishment method as described above.

[0185] The method implemented when the tag relationship establishment program running on the processor is executed can be referred to in various embodiments of the tag relationship establishment method of the present invention, and will not be repeated here.

[0186] The present invention also provides a computer program product.

[0187] The computer program product includes a tag relationship establishment program, which, when executed by a processor, implements the steps of the tag relationship establishment method as described above.

[0188] The method implemented when the tag relationship establishment program running on the processor is executed can be referred to in various embodiments of the tag relationship establishment method of the present invention, and will not be repeated here.

[0189] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0190] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0191] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0192] The above are merely preferred embodiments of the present invention and do not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for establishing label relationships, characterized in that, The method for establishing tag relationships includes the following steps: Sampling units are determined and distributed on the target scene area. The relevant parameters of the sampling units include size and shape. The relevant parameters of the sampling units are set according to the target scene area, which is a three-dimensional game scene. Determine the model set corresponding to each sampling unit; For each model set, the positional relationships between models and the label of each model are determined, and the connection data between the labels is generated based on the positional relationships. The positional relationships include adjacent positional relationships and intersecting positional relationships, and the connection data includes adjacent connection data obtained from the adjacent positional relationships and intersecting connection data obtained from the intersecting positional relationships. Establish label relationships based on the connection data corresponding to all model sets; The step of determining the positional relationship between the models includes: Based on the sampling units corresponding to the model set, the positional relationship between the models in the model set is determined as the adjacent positional relationship; Obtain the bounding box corresponding to each model in the model set, and determine the positional relationship between the models in the model set as the intersection positional relationship based on the bounding boxes; The step of establishing label relationships based on the connection data corresponding to all model sets includes: Based on the connection data corresponding to all model sets, frequent itemsets are determined according to preset frequent rules and frequency confidence. The frequent itemsets include adjacent frequent itemsets and intersecting frequent itemsets. Establish label relationships based on the frequent itemsets; The step of establishing tag relationships based on the contact data further includes: Based on the adjacent connection data and the intersecting connection data, establish a corresponding frequent pattern tree, item header table and node linked list; The label relationships are obtained by recursively mining based on the frequent pattern tree, the item header table, and the node linked list.

2. The tag relationship establishment method as described in claim 1, characterized in that, The step of determining the model set corresponding to each sampling unit includes: Collect the model at the corresponding position of each sampling unit, and determine the model preset corresponding to each sampling unit; The models in the model pre-set are filtered to determine the model set corresponding to each sampling unit.

3. The tag relationship establishment method as described in claim 1, characterized in that, The step of determining the positional relationship between models in the model set as the intersecting positional relationship based on the bounding box includes: The bounding box is magnified according to a preset magnification factor, and it is determined whether the magnified bounding boxes intersect. If the enlarged bounding boxes intersect, then the positional relationship between the models corresponding to the enlarged bounding boxes is determined as the intersecting positional relationship.

4. The tag relationship establishment method as described in claim 1, characterized in that, The frequent itemsets include adjacent frequent itemsets determined by the adjacent relationship data, and the step of establishing label relationships based on the frequent itemsets includes: Establish adjacency label relationships between labels in the adjacent frequent itemset.

5. The tag relationship establishment method as described in claim 1 or 4, characterized in that, The frequent itemsets include the intersecting frequent itemsets determined by the intersecting relationship data, and the step of establishing label relationships based on the frequent itemsets includes: In the set of frequently intersecting items, identify tags with the top-level placement attribute; For each tag with the top-level placement attribute, establish intersecting tag relationships between it and other tags.

6. The tag relationship establishment method as described in claim 5, characterized in that, The step of establishing label relationships based on the frequent itemsets includes: For tags that do not have the attribute of being placed at the top level, establish adjacency relationships between tags.

7. A tag relationship establishment device, characterized in that, The tag relationship establishment device includes: The first determining module is used to determine sampling units and distribute the sampling units on the target scene area. The relevant parameters of the sampling units include size and shape. The relevant parameters of the sampling units are set according to the target scene area, which is a three-dimensional game scene. The second determining module is used to determine the model set corresponding to each sampling unit; The third determining module is used to determine the positional relationship between models and the label of each model for each model set, and to generate connection data between the labels according to the positional relationship. The positional relationship includes adjacent positional relationship and intersecting positional relationship, and the connection data includes adjacent connection data obtained from the adjacent positional relationship and intersecting connection data obtained from the intersecting positional relationship. The fourth module is used to establish label relationships based on the connection data corresponding to all model sets; The determining module is further configured to determine the positional relationship between models in the model set as the adjacent positional relationship based on the sampling unit corresponding to the model set; Obtain the bounding box corresponding to each model in the model set, and determine the positional relationship between the models in the model set as the intersection positional relationship based on the bounding boxes; The fourth determining module is further configured to determine frequent itemsets based on preset frequent rules and frequency confidence based on the connection data corresponding to all model sets, wherein the frequent itemsets include adjacent frequent itemsets and intersecting frequent itemsets; Establish label relationships based on the frequent itemsets; The fourth determining module is further configured to establish a corresponding frequent pattern tree, item header table, and node linked list based on the adjacent connection data and the intersecting connection data. The label relationships are obtained by recursively mining based on the frequent pattern tree, the item header table, and the node linked list.

8. A tag relationship establishment system, characterized in that, The tag relationship establishment system includes: a memory, a processor, and a tag relationship establishment program stored in the memory and executable on the processor. When the tag relationship establishment program is executed by the processor, it implements the steps of the tag relationship establishment method as described in any one of claims 1 to 6.

9. A medium, characterized in that, The medium is a computer-readable storage medium, and the computer program product stores a tag relationship establishment program, which, when executed by a processor, implements the steps of the tag relationship establishment method as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, The computer program product includes a tag relationship establishment program, which, when executed by a processor, implements the steps of the tag relationship establishment method as described in any one of claims 1 to 6.

Images