Similar media material processing method and device, equipment and storage medium
By using multiple similarity algorithms and weighted calculation methods, the media materials are grouped, which solves the problem of low grouping accuracy of a single algorithm and achieves more accurate delivery management of similar media materials.
Patent Information
- Application Number
- CN202411767795.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-04
- Publication Date
- 2025-03-14
AI Technical Summary
When the prior art processes different pictures or splicing pictures, the grouping effect of a single similarity algorithm is poor, resulting in inaccurate delivery and management of similar media materials.
Multiple similarity algorithms (such as K nearest neighbor algorithm and approximate nearest neighbor algorithm) are used to calculate the similarity of media materials, and use similarity weighting to form more accurate groups.
It improves the accuracy of grouping of similar media materials, ensures accurate delivery of media materials, and avoids excessive traffic consumption.
Smart Images
Figure CN119625353A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of image processing technology, and in particular to a method, device, equipment and storage medium for processing similar media materials. Background Art
[0002] In the process of advertising, users usually upload a large amount of media materials on the platform in order to gain more product exposure. However, the indiscriminate delivery of a large amount of media materials will lead to excessive consumption of platform traffic. In order to avoid excessive consumption of traffic, it is usually necessary to accurately find similar media materials and limit the delivery of similar media materials.
[0003] In the related art, a single similarity calculation method is usually used to determine the similarity of media materials, and media materials with higher similarity are grouped together.
[0004] However, for dissimilar images or spliced images, the grouping accuracy of a single similarity algorithm is poor. Summary of the invention
[0005] In order to solve the above technical problems, the present application provides a method, apparatus, device and storage medium for processing similar media materials to improve the accuracy of grouping similar media materials.
[0006] In a first aspect, the present application provides a method for processing similar media materials, the method comprising: obtaining a first media material set; wherein the first media material set includes multiple media materials; using a first similarity algorithm to perform similarity calculation on the media materials in the first media material set to obtain at least one first grouping, and a first similarity value of the media materials in the first grouping is greater than a first threshold; using a second similarity algorithm to perform similarity calculation on the media materials in the first media material set to obtain at least one second grouping, and a second similarity value of the media materials in the second grouping is greater than a second threshold; performing weighted similarity calculation based on the media materials in the first grouping and the media materials in the second grouping to obtain at least one third grouping, and a third similarity value of the media materials in the third grouping reaches a third threshold, and the media materials in the third grouping belong to the first grouping and the second grouping.
[0007] In a second aspect, the present application provides a similar media material processing device, the device comprising: a first material set acquisition module, used to acquire a first media material set; wherein the first media material set includes multiple media materials; a first group determination module, used to use a first similarity algorithm to perform similarity calculation on the media materials in the first media material set to obtain at least one first group, and the first similarity value of the media materials in the first group is greater than a first threshold; a second group determination module, used to use a second similarity algorithm to perform similarity calculation on the media materials in the first media material set to obtain at least one second group, and the second similarity value of the media materials in the second group is greater than a second threshold; a third group determination module, used to perform similarity weighted calculation based on the media materials in the first group and the media materials in the second group to obtain at least one third group, the third similarity value of the media materials in the third group reaches a third threshold, and the media materials in the third group belong to the first group and the second group.
[0008] In a third aspect, the present application provides an electronic device, which may include a similar media material processing device, the device including: one or more processors; a storage device for storing one or more programs; when the one or more programs are executed by one or more processors, the one or more processors implement the similar media material processing method as described in the first aspect above.
[0009] In a fourth aspect, the present application provides a storage medium, which may be a computer-readable storage medium, storing a computer program thereon, which, when executed by a processor, implements a similar media material processing method as in the first aspect above.
[0010] In a fifth aspect, an embodiment of the present application provides a computer program product, which includes a computer program or instructions, and when the computer program or instructions are executed by a processor, a similar media material processing method as described in any one of the first aspects above is implemented.
[0011] Compared with the prior art, the technical solution provided by the embodiments of the present application has the following advantages:
[0012] The embodiment of the present application provides a method, device, equipment and storage medium for processing similar media materials, the method comprising: obtaining a first media material set; wherein the first media material set includes multiple media materials; using a first similarity algorithm to perform similarity calculation on the media materials in the first media material set to obtain at least one first group, the first similarity value of the media materials in the first group is greater than a first threshold; using a second similarity algorithm to perform similarity calculation on the media materials in the first media material set to obtain at least one second group, the second similarity value of the media materials in the second group is greater than a second threshold; performing weighted similarity calculation based on the media materials in the first group and the media materials in the second group to obtain at least one third group, the third similarity value of the media materials in the third group reaches a third threshold, and the media materials in the third group belong to the first group and the second group. Since the media materials are divided into multiple first groups using the first similarity algorithm, and the media materials are divided into multiple second groups using the second similarity algorithm, and then the third group is determined from the first group and the second group according to the weighted similarity calculation. Since after the similarity values are calculated using multiple similarity algorithms, the similarity values are weightedly calculated to determine the similarity between the media materials, the accuracy of the grouping of similar media materials is improved. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.
[0014] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.
[0015] Figure 1 A schematic diagram of a process flow of a similar media material processing method provided in an embodiment of the present application;
[0016] Figure 2 A schematic diagram of a process flow of another similar media material processing method provided in an embodiment of the present application;
[0017] Figure 3 A schematic diagram of the structure of a similar media material processing device provided in an embodiment of the present application;
[0018] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0019] In order to more clearly understand the above-mentioned purposes, features and advantages of the present application, the scheme of the present application will be further described below. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.
[0020] In the following description, many specific details are set forth to facilitate a full understanding of the present application, but the present application may also be implemented in other ways different from those described herein; obviously, the embodiments in the specification are only part of the embodiments of the present application, rather than all of the embodiments.
[0021] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.
[0022] It should be noted that the concepts such as "first" and "second" mentioned in this application are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.
[0023] It should be noted that the modifications of "one" and "plurality" mentioned in the present application are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".
[0024] The following is a detailed description of the similar media material processing method provided by the present application in conjunction with the accompanying drawings and specific implementation methods.
[0025] Figure 1 This is a flowchart of a similar media material processing method in an embodiment of the present application. This embodiment can be applicable to grouping similar media materials. The method can be executed by a similar media material processing device. The similar media material processing device can be implemented in software and / or hardware. The similar media material processing device can be configured in an electronic device.
[0026] like Figure 1 As shown, the similar media material processing method provided in the embodiment of the present application mainly includes steps S101-S104.
[0027] S101: Acquire a first media material set; wherein the first media material set includes a plurality of media materials.
[0028] Among them, media materials generally refer to basic materials and elements used to produce various media contents (such as videos, audios, websites, applications, publications, etc.), and media materials may include picture materials, video materials, text materials, graphic materials, audio materials, etc. In this embodiment, the media material is taken as an example for description.
[0029] The first media material set includes multiple media materials, and the multiple media materials in the first media material set have the same user identifier.
[0030] Since multiple users can upload media materials to the platform at the same time, after receiving the media materials uploaded by multiple users, the received media materials can be grouped according to user identifiers, and the media materials with the same user identifiers can be grouped into a media material set as the second media material set. Exemplarily, the user identifier can be an advertiser identifier.
[0031] S102: Calculate similarity of media materials in a first media material set by using a first similarity algorithm to obtain at least one first group, wherein a first similarity value of the media materials in the first group is greater than a first threshold.
[0032] The first similarity value can be understood as the similarity value between each media material calculated by using the first similarity algorithm. The first threshold refers to a value that characterizes the similarity of the media materials. If the first similarity value is greater than the first threshold, it indicates that the similarity between the multiple media materials is very high. If the first similarity value is less than or equal to the first threshold, it indicates that the similarity between the multiple media materials is very low and they are not similar media materials. The first threshold can be set according to actual conditions.
[0033] The first similarity algorithm may be a K-nearest neighbor algorithm. The K-nearest neighbor algorithm (KNN) is a basic classification and regression algorithm, and its working principle is to perform classification or regression prediction by measuring the distance between different feature values.
[0034] In one possible implementation, similarity is calculated for media materials in a first media material set using a first similarity algorithm to obtain at least one first group, including: traversing the media materials in the first media material set as first media materials; calculating first similarity values between the first media materials and each second media material, wherein the second media materials are media materials in the first media material set other than the first media materials; and grouping the second media materials having first similarity values greater than a first threshold and the first media materials into the same first group.
[0035] In this embodiment, for the media materials in the first media material set, feature vectors corresponding to the respective media materials are extracted, and all feature vectors are stored in a predetermined data structure.
[0036] Traverse the feature vectors corresponding to the media materials from the data structure, and calculate the first similarity value between the feature vectors corresponding to the traversed media materials and each feature vector in the data structure; after the first similarity values between the feature vectors corresponding to the traversed media materials and all other feature vectors in the data structure are calculated, divide the media materials corresponding to all feature vectors whose first similarity values are greater than a first threshold and the traversed media materials into the same first group.
[0037] The first similarity value may be calculated using distance formulas such as Euclidean distance and Manhattan distance, and is not specifically limited in this embodiment.
[0038] The above feature vector includes various features in the material, such as edges, textures, shapes, and colors, and summarizes the content of the picture.
[0039] The KNN algorithm does not need to learn a discriminant function based on samples, but directly stores the feature vectors corresponding to the media materials and directly applies them when calculating the similarity, so as to find all similar media materials in the first media material set.
[0040] S103: Calculate similarity of media materials in the first media material set by using a second similarity algorithm to obtain at least one second group, wherein a second similarity value of the media materials in the second group is greater than a second threshold.
[0041] The second similarity value can be understood as the similarity value between each media material calculated by using the second similarity algorithm. The second threshold value refers to a value representing the similarity degree of the media materials. If the second similarity value is greater than the second threshold value, it indicates that the similarity degree between the multiple media materials is very high. If the second similarity value is less than or equal to the second threshold value, it indicates that the similarity degree between the multiple media materials is very low and they are not similar media materials. The second threshold value can be set according to actual conditions.
[0042] The first threshold and the second threshold may be the same or different, and are not specifically limited in this embodiment.
[0043] Among them, the second similarity algorithm can be an approximate nearest neighbor algorithm. The approximate nearest neighbor (ANN) search algorithm is a method for finding the approximate nearest neighbor point of a given point in a high-dimensional space. The goal of the approximate nearest neighbor algorithm is to find a result that is close enough to the exact nearest neighbor point in a reasonable time. This means sacrificing a certain accuracy in exchange for a faster search speed. ANN algorithms usually construct data structures or use specific search strategies to speed up the query process.
[0044] In one possible implementation, similarity is calculated for media materials in a first media material set using a second similarity algorithm to obtain at least one second group, including: traversing the media materials in the first media material set as third media materials; searching for media materials that are nearest neighbors to the third media materials in the first media material set using a preset index structure; calculating a second similarity value between the third media material and each nearest neighbor media material; and grouping the nearest neighbor media material whose second similarity value is greater than the second threshold and the third media material into the same second group.
[0045] In this embodiment, for the media materials in the first media material set, feature vectors corresponding to each media material are extracted, and all feature vectors are stored in a predetermined data structure. The feature vectors include various features in the material, such as edges, textures, shapes, and colors, etc., which summarize the content of the picture. The extraction of feature vectors can be achieved through an AI model or a deep learning network, which is not specifically limited in this embodiment.
[0046] In order to efficiently perform approximate nearest neighbor search, an index structure is pre-built, which can be a tree-based structure (such as KD tree), a hash-based structure (Locality Sensitive Hashing, LSH), a graph-based structure (Hierarchical Navigable Small World Graphs, HNSW) or a quantization-based method (ProductQuantization, PQ). The feature vectors of all media materials extracted in the above steps are stored in the index structure.
[0047] Traverse the feature vectors of all media materials in the index structure. Use the constructed index structure to perform an approximate nearest neighbor search on the feature vector of the currently traversed media material. Obtain a list of feature vectors in the index structure that are closest to the currently traversed media material.
[0048] The second similarity value between the feature vector of the currently traversed media material and the feature vector of the media material is calculated, which is usually accomplished by measuring the distance between the feature vectors, such as Euclidean distance, cosine similarity or Manhattan distance.
[0049] The media materials corresponding to all feature vectors whose second similarity values are greater than the second threshold and the traversed media materials are grouped into the same second group.
[0050] S104: Perform weighted similarity calculation based on the media materials in the first group and the media materials in the second group to obtain at least one third group, wherein a third similarity value of the media materials in the third group reaches a third threshold, and the media materials in the third group belong to the first group and the second group.
[0051] The third similarity value can be understood as a similarity value obtained by weighted calculation using the first similarity value and the second similarity value. The third threshold value refers to a value representing the similarity of the media materials. If the third similarity value is greater than the third threshold value, it indicates that the similarity between the multiple media materials is very high. If the third similarity value is less than or equal to the third threshold value, it indicates that the similarity between the multiple media materials is very low and they are not similar media materials. The third threshold value can be set according to actual conditions.
[0052] In a possible implementation manner, the third similarity value of the media material in the third group is obtained by weighted calculation of the first similarity value of the media material and the second similarity value of the media material.
[0053] In the embodiment of the present application, media materials that exist in both the first group and the second group are first determined as preliminary media materials. If a media material belongs to the first group and the second group, it indicates that the first similarity value calculated by the first similarity algorithm is greater than the first threshold, and the second similarity value calculated by the second similarity algorithm is greater than the second threshold. In other words, no matter whether the first similarity algorithm or the second similarity algorithm is used, it is similar to the media materials in the first group and the second group.
[0054] For multiple preliminary selected media materials, use the preliminary selected media materials in pairs to construct preliminary selected media material pairs, obtain the first similarity value and the second similarity value of the preliminary selected media material pair, perform weighted calculation, and obtain a third similarity value. If the third similarity value is greater than the third threshold value, determine whether the preliminary selected media materials in the preliminary selected media material pair are in a third group. If one of the preliminary selected media materials has a third group to which it belongs, then add the other preliminary selected media material in the preliminary selected media material pair to the third group. If neither of the two preliminary selected media materials in the preliminary selected media material pair belongs to any third group, then create a new third group, and add the preliminary selected media material pair to the newly created third group.
[0055] The third similarity value of each media material in the first group and the second group is calculated according to the above method, and grouping is performed according to the third similarity value to obtain a plurality of third groups.
[0056] If a media material only belongs to the first group but not the second group, it means that the second similarity value calculated by the second similarity algorithm is less than or equal to the second threshold, that is, the media material is not similar to the media material in the second group. Alternatively, if a media material only belongs to the second group but not the first group, it means that the first similarity value calculated by the first similarity algorithm is less than or equal to the first threshold, that is, the media material is not similar to the media material in the first group. For such media materials, they are not classified into the third group to further improve the accuracy of grouping similar materials.
[0057] The embodiment of the present application provides a similar media material processing method, the method comprising: obtaining a first media material set; wherein the first media material set includes multiple media materials; using a first similarity algorithm to perform similarity calculation on the media materials in the first media material set to obtain at least one first group, the first similarity value of the media materials in the first group is greater than a first threshold; using a second similarity algorithm to perform similarity calculation on the media materials in the first media material set to obtain at least one second group, the second similarity value of the media materials in the second group is greater than a second threshold; performing weighted similarity calculation based on the media materials in the first group and the media materials in the second group to obtain at least one third group, the third similarity value of the media materials in the third group reaches a third threshold, and the media materials in the third group belong to the first group and the second group. Since the media materials are divided into multiple first groups using the first similarity algorithm, and the media materials are divided into multiple second groups using the second similarity algorithm, and then the third group is determined from the first group and the second group according to the weighted similarity calculation. Since after the similarity values are calculated using multiple similarity algorithms, the similarity values are weightedly calculated to determine the similarity between the media materials, the accuracy of the grouping of similar media materials is improved.
[0058] Based on the above embodiments, the present application embodiment optimizes the method for processing similar media materials, such as Figure 2 As shown, the optimized similar media material processing method includes the steps of:
[0059] S201: Acquire a second media material set.
[0060] The second media material set may be understood as all media materials uploaded to the platform by the user.
[0061] Acquiring the second media material set includes: when a media material upload message is detected, acquiring the media material set uploaded by the user as the second media material set.
[0062] In a possible implementation manner, the second media material set includes multiple media materials with the same user identifier.
[0063] Since multiple users can upload media materials to the platform at the same time, after receiving the media materials uploaded by multiple users, the received media materials can be grouped according to user identifiers, and the media materials with the same user identifiers can be grouped into a media material set as the second media material set. Exemplarily, the user identifier can be an advertiser identifier.
[0064] Since the media materials uploaded by different users are necessarily different, there is no need to judge whether the media materials between different users are similar. Therefore, only judging whether the media materials uploaded by the same user are similar can reduce the amount of calculation and improve the calculation speed.
[0065] S202: Use the large model to identify material types of media materials in the second media material set.
[0066] The above-mentioned material types can be understood as media materials that distinguish whether the media materials can be processed by computers. For example, the material types of media materials that cannot be processed by computers are target types, and the media material types that can be processed by computers are non-target types. Among them, the media materials that can be processed by computers can be understood as the features in the media materials that can be extracted by computers, and the similarity values between the media materials can be calculated.
[0067] The target type includes a splicing type and / or a single item type, wherein the splicing type means that the media material is composed of multiple materials combined into one structure, and the single item type means that only one object is displayed in the media material.
[0068] For example, when the media material is a picture, the splicing type may include a picture formed by combining two or more small images. Furthermore, the splicing type of picture may also include: a picture generated by using an AI model.
[0069] Single item type refers to a media material that only displays one object. For example, a picture material only displays a white computer. There is no text description in the picture material, and there is no background image or the background image is a solid color.
[0070] When the media material type is a splicing type or a single item type, the computer cannot calculate the similarity value between the media materials. Therefore, such media materials need to be processed separately.
[0071] The large model may include an open source large language model (LLM). The media material in the second media material set is input into the large model, and a first prompt word is input into the large model, the first prompt word is used to instruct the large model to identify the media material and output the material type of the media material, and the material type of the media material output by the large model is received.
[0072] Using the big model to identify the material type of the media material in the second media material set includes: obtaining a first prompt word, the first prompt word is used to instruct the big model to identify the type of the media material and output the material type of the media material; for the media material in the second media material set, combining the media material with the first prompt word to form first input information, inputting the first input information into the big model, and receiving the material type of the media material output by the big model.
[0073] Exemplarily, the first prompt word may be “Please identify whether the following picture is a splicing type picture and / or a single product type picture”; further, the first prompt word may also explain the definition of the splicing type and / or the single product type.
[0074] S203: Select a first media material set from the second media material set based on the material type of the media material, wherein the material type of the media material in the first media material set is a non-target type.
[0075] In this embodiment, the media materials of non-target type in the second media material set are divided into the first media material set. The other pictures in the second media material set except the splicing type pictures and the single product type pictures are taken as the first media material set.
[0076] The media materials in the first media material set may be grouped according to similar programs between the media materials.
[0077] S204: Calculate similarity of media materials in the first media material set by using a first similarity algorithm to obtain at least one first group, wherein a first similarity value of the media materials in the first group is greater than a first threshold.
[0078] S205: Calculate similarity of the media materials in the first media material set by using a second similarity algorithm to obtain at least one second group, wherein a second similarity value of the media materials in the second group is greater than a second threshold.
[0079] S206: Perform weighted similarity calculation based on the media materials in the first group and the media materials in the second group to obtain at least one third group, wherein a third similarity value of the media materials in the third group reaches a third threshold, and the media materials in the third group belong to the first group and the second group.
[0080] The execution process of steps S204-S206 provided in the embodiment of the present application is the same as that of steps S102-S104 provided in the above embodiment. For details, please refer to the description in the above embodiment, which will not be repeated in the embodiment of the present application.
[0081] S207: Select a third media material set from the second media material set based on the material type of the media material, wherein the material type of the media material in the third media material set is the target type.
[0082] In this embodiment, the media materials of the target type in the second media material set are divided into the third media material set. The splicing type pictures and / or single product type pictures in the second media material set are taken as the first media material set.
[0083] The media materials in the third media material set may be grouped according to tag information marked on the media materials.
[0084] S208: Use the large model to label the media materials in the third media material set to obtain label information of the media materials in each third media material set.
[0085] The tag information may be the unique identification information of the product. For example, the tag information may be the product model, product code, etc. The tag information of the media material may be the identifier corresponding to the content displayed in the media material. For example, when the media material is a picture, and a model A mobile phone is displayed in the picture, then the tag information of the media material is model A mobile phone.
[0086] In a possible implementation, obtaining label information corresponding to the media material in the third media material set includes: obtaining label information input by a user for the media material in the third media material set. In other words, the media material in the third media material set is manually labeled.
[0087] In a possible implementation, the media material in the third media material set is input into the large language model so that the large language model recognizes the media material, identifies the content displayed in the media material, and uses the content to annotate the media material to obtain label information of the media material.
[0088] Furthermore, the large model is used to mark the media materials in the third media material set to obtain label information of the media materials in each third media material set, including: obtaining a second prompt word, the second prompt word is used to instruct the large model to identify the content of the media material and output the label information of the media material, for the media material in the third media material set, combining the media material with the second prompt word to form second input information, inputting the second input information into the large model, and receiving the label information of the media material output by the large model.
[0089] Exemplarily, the second prompt word may be “Please identify the product model of the product shown in the following picture”.
[0090] S209: Group the media materials of the third media material set according to the tag information to obtain a plurality of fourth groups; the fourth groups include a plurality of media materials with the same tag information.
[0091] In this embodiment, media materials with the same tag information are grouped into the same group as the fourth group. For example, if the tag information of media material 1, media material 3, and media material 5 is a mobile phone of model A, then media material 1, media material 3, and media material 5 are grouped into the fourth group. If the tag information of media material 2, media material 4, and media material 6 is a mobile phone of model B, then media material 2, media material 4, and media material 6 are grouped into the fourth group.
[0092] In an embodiment of the present application, a second media material set is obtained, a material type of the media material in the second media material set is identified using a large model, a third media material set is selected from the second media material set based on the material type of the media material, wherein the material type of the media material in the third media material set is a target type, label information corresponding to the media material in the third media material set is obtained, and the media materials in the third media material set are grouped according to the label information to obtain a plurality of fourth groups. In this embodiment, for media materials that cannot be grouped using similarity, the marked label information is used for grouping, and the grouping of similar media materials is achieved by combining manual and AI methods.
[0093] Figure 3 is a structural diagram of a similar media material processing device in an embodiment of the present application, such as Figure 3 As shown, the similar media material processing device 30 provided in the embodiment of the present application mainly includes: a first material set acquisition module 31, a first group determination module 32, a second group determination module 33 and a third group determination module 34.
[0094] Among them, the first material set acquisition module 31 is used to acquire a first media material set; wherein the first media material set includes multiple media materials; the first group determination module 32 is used to use a first similarity algorithm to perform similarity calculation on the media materials in the first media material set to obtain at least one first group, and the first similarity value of the media materials in the first group is greater than a first threshold; the second group determination module 33 is used to use a second similarity algorithm to perform similarity calculation on the media materials in the first media material set to obtain at least one second group, and the second similarity value of the media materials in the second group is greater than a second threshold; the third group determination module 34 is used to perform similarity weighted calculation based on the media materials in the first group and the media materials in the second group to obtain at least one third group, the third similarity value of the media materials in the third group reaches a third threshold, and the media materials in the third group belong to the first group and the second group.
[0095] The embodiment of the present application provides a similar media material processing device, which is used to execute the following process: obtaining a first media material set; wherein the first media material set includes multiple media materials; using a first similarity algorithm to perform similarity calculation on the media materials in the first media material set to obtain at least one first group, and the first similarity value of the media materials in the first group is greater than a first threshold; using a second similarity algorithm to perform similarity calculation on the media materials in the first media material set to obtain at least one second group, and the second similarity value of the media materials in the second group is greater than a second threshold; performing weighted similarity calculation based on the media materials in the first group and the media materials in the second group to obtain at least one third group, the third similarity value of the media materials in the third group reaches the third threshold, and the media materials in the third group belong to the first group and the second group. Since the media materials are divided into multiple first groups using the first similarity algorithm, and the media materials are divided into multiple second groups using the second similarity algorithm, and then the third group is determined from the first group and the second group according to the weighted similarity calculation. Since after the similarity values are calculated using multiple similarity algorithms, the similarity values are weightedly calculated to determine the similarity between the media materials, the accuracy of the grouping of similar media materials is improved.
[0096] In a possible implementation, the first group determination module 32 is specifically configured to traverse media materials in the first media material set as first media materials; calculate first similarity values between the first media materials and each second media material, wherein the second media materials are media materials other than the first media materials in the first media material set; and group the second media materials whose first similarity values are greater than a first threshold into the same first group as the first media materials.
[0097] In a possible implementation, the second group determination module 33 is specifically used to traverse the media materials in the first media material set as the third media material; in the first media material set, use a preset index structure to search for the media material that is the nearest neighbor to the third media material; calculate the second similarity value between the third media material and each nearest neighbor media material; and group the nearest neighbor media material whose second similarity value is greater than the second threshold and the third media material into the same second group.
[0098] In a possible implementation manner, the third similarity value of the media material in the third group is obtained by weighted calculation of the first similarity value of the media material and the second similarity value of the media material.
[0099] In one possible implementation, the first material set acquisition module 31 is specifically used to acquire a second media material set; use a large model to identify the material type of the media material in the second media material set; select the first media material set from the second media material set based on the material type of the media material, wherein the material type of the media material in the first media material set is a non-target type, wherein the target type includes a splicing type and / or a single product type, wherein the splicing type means that the multiple media materials in the second media material set are composed of multiple materials combined together, and the single product type means that only one object is displayed in the media material.
[0100] In one possible implementation, the fourth grouping determination module is used to select a third media material set from the second media material set based on the material type of the media material, wherein the material type of the media material in the third media material set is a target type; use a large model to mark the media materials in the third media material set to obtain label information of each media material in the third media material set; group the media materials of the third media material set according to the label information to obtain multiple fourth groups; the fourth group includes multiple media materials with the same label information.
[0101] In a possible implementation, the media material set includes multiple media materials with the same user identifier.
[0102] The similar media material processing device provided in the embodiment of the present application can execute the similar media material processing method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.
[0103] Figure 4 is a schematic diagram of the structure of an electronic device provided in this embodiment. The electronic device may include similar media material processing devices, such as Figure 4As shown, the electronic device 400 includes a processor 410, a memory 420, an input device 430, and an output device 440; the number of the processor 410 in the electronic device can be one or more. Figure 4 A processor 410 is taken as an example; the processor 410, the memory 420, the input device 430 and the output device 440 in the electronic device can be connected via a bus or other means. Figure 4 The example of connecting through bus is taken in the following.
[0104] The memory 420 is a computer-readable storage medium that can be used to store software programs, computer executable programs and modules, such as program instructions / modules corresponding to the similar media material processing method in the embodiment of the present invention. The processor 410 executes various functional applications and data processing of the electronic device by running the software programs, instructions and modules stored in the memory 420, that is, implements the similar media material processing method provided in the embodiment of the present invention.
[0105] The memory 420 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created according to the use of the terminal, etc. In addition, the memory 420 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 420 may further include a memory remotely arranged relative to the processor 410, and these remote memories may be connected to the electronic device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.
[0106] The input device 430 may be used to receive input digital or character information and generate key signal input related to the advertiser settings and function control of the electronic device, and may include a keyboard, a mouse, etc. The output device 440 may include a display device such as a display screen.
[0107] This embodiment also provides a storage medium containing computer executable instructions. When the computer executable instructions are executed by a computer processor, they are used to implement the similar media material processing method provided by the embodiment of the present invention.
[0108] Of course, the computer executable instructions of a storage medium including computer executable instructions provided by an embodiment of the present invention are not limited to the above method operations, and can also execute related operations in similar media material processing methods provided by any embodiment of the present invention.
[0109] Through the above description of the implementation method, the technicians in the relevant field can clearly understand that the present invention can be implemented by means of software and necessary general hardware, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform the methods of various embodiments of the present invention.
[0110] It is worth noting that in the above-mentioned embodiments of similar media material processing devices, the various units and modules included are only divided according to functional logic, but are not limited to the above-mentioned division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.
[0111] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.
[0112] The above is only a specific implementation of the present application, so that those skilled in the art can understand or implement the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to these embodiments herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for processing similar media materials, characterized in that: The method comprises: Acquire a first media material set; wherein the first media material set includes a plurality of media materials; Using a first similarity algorithm to calculate similarity of media materials in the first media material set, to obtain at least one first group, wherein first similarity values of the media materials in the first group are greater than a first threshold; Using a second similarity algorithm to calculate similarity of media materials in the first media material set to obtain at least one second group, wherein a second similarity value of the media materials in the second group is greater than a second threshold; A weighted similarity calculation is performed based on the media materials in the first group and the media materials in the second group to obtain at least one third group, wherein a third similarity value of the media materials in the third group reaches a third threshold, and the media materials in the third group belong to the first group and the second group.
2. The method according to claim 1, characterized in that The using a first similarity algorithm to calculate the similarity of the media materials in the first media material set to obtain at least one first group includes: Traversing the media materials in the first media material set as first media materials; Calculating a first similarity value between the first media material and each second media material, wherein the second media material is a media material in the first media material set other than the first media material; The second media material whose first similarity value is greater than the first threshold is divided into the same first group with the first media material.
3. The method according to claim 1, characterized in that The using a second similarity algorithm to perform similarity calculation on the media materials in the first media material set to obtain at least one second group includes: Traversing the media materials in the first media material set as third media materials; In the first media material set, searching for the media material that is closest to the third media material by using a preset index structure; Calculating a second similarity value between the third media material and each of the nearest neighboring media materials; The nearest neighbor media material whose second similarity value is greater than the second threshold value and the third media material are grouped into the same second group. 4 . The method according to claim 1 , wherein the third similarity value of the media material in the third group is obtained by weighted calculation of the first similarity value of the media material and the second similarity value of the media material.
5. The method according to claim 1, characterized in that Get the first media material set, including: Acquire a second media material set; identifying material types of media materials in the second media material set by using the large model; A first media material set is selected from the second media material set based on the material type of the media material, wherein the material type of the media material in the first media material set is a non-target type, wherein the target type includes a splicing type and / or a single product type, wherein the splicing type means that the media material is composed of multiple media materials in the second media material set combined together, and the single product type means that only one object is displayed in the media material.
6. The method according to claim 5, characterized in that Also includes: Selecting a third media material set from the second media material set based on the material type of the media material, wherein the material type of the media material in the third media material set is a target type; Using the large model to label the media materials in the third media material set, to obtain label information of each media material in the third media material set; The media materials of the third media material set are grouped according to the tag information to obtain a plurality of fourth groups; the fourth groups include a plurality of media materials with the same tag information.
7. The method according to any one of claims 1 to 6, characterized in that The media material set includes multiple media materials with the same user identifier.
8. A similar media material processing device, characterized in that: The device comprises: A first material set acquisition module, configured to acquire a first media material set; wherein the first media material set includes a plurality of media materials; A first group determination module, configured to calculate similarity of the media materials in the first media material set by using a first similarity algorithm to obtain at least one first group, wherein a first similarity value of the media materials in the first group is greater than a first threshold; A second group determination module, configured to calculate similarity of the media materials in the first media material set by using a second similarity algorithm to obtain at least one second group, wherein a second similarity value of the media materials in the second group is greater than a second threshold; The third group determination module is used to perform weighted similarity calculation based on the media materials in the first group and the media materials in the second group to obtain at least one third group, wherein the third similarity value of the media materials in the third group reaches a third threshold, and the media materials in the third group belong to the first group and the second group.
9. An electronic device, characterized in that: The device comprises: one or more processors; A storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the similar media material processing method according to any one of claims 1 to 7.
10. A storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the similar media material processing method as described in any one of claims 1 to 7 is implemented.