A label text processing method, device and equipment

Abstract

The application discloses a label text processing method, device and equipment. In the method, after obtaining a new label text, a first multimedia feature of multimedia information corresponding to the new label text is obtained. A first similarity between the first multimedia feature and a second multimedia feature corresponding to an existing label text in a label library is calculated. If the first similarity between the first multimedia feature and each second multimedia feature is less than a first similarity threshold, it represents that the new label text and the existing label text in the label library are quite different, and the new label text is not a similar label of the existing label text. At this time, the new label text and the corresponding first multimedia feature can be added to the label library, so that the new label text and the corresponding multimedia feature are quickly stored in the label library.

Description

Technical Field

[0001] This application relates to the field of data processing technology, specifically to a method, apparatus, and device for processing tag text. Background Technology

[0002] Scene recognition involves deep semantic description of multimedia information such as videos and images. Through structured parsing of multimedia information, it identifies the context and events that occur within the multimedia content. Scene recognition requires establishing a tag library to store tag text and corresponding multimedia features; the tag text can be scene labels. Therefore, when scene recognition is needed, this tag library can be used to determine the tag text corresponding to the multimedia information to be identified.

[0003] However, with the development of the times, there will be many new tag texts. How to quickly process the new tag texts is a technical problem that urgently needs to be solved. Summary of the Invention

[0004] In view of this, embodiments of this application provide a method, apparatus, and device for processing tag text, so as to achieve rapid processing of newly added tag text.

[0005] To address the above problems, the technical solutions provided in this application are as follows:

[0006] In a first aspect, embodiments of this application provide a label text processing method, the method comprising:

[0007] Get the text of the newly added tag;

[0008] Obtain the first multimedia feature of the multimedia information corresponding to the newly added tag text;

[0009] Calculate the first similarity between the first multimedia feature and the second multimedia feature corresponding to the existing tag text in the tag library;

[0010] If the first similarity between the first multimedia feature and each of the second multimedia features is less than the first similarity threshold, the newly added tag text and the first multimedia feature corresponding to the newly added tag text are added to the tag library.

[0011] Secondly, embodiments of this application provide a label text processing apparatus, the apparatus comprising:

[0012] The first acquisition unit is used to acquire the text of the newly added tag;

[0013] The second acquisition unit is used to acquire the first multimedia feature of the multimedia information corresponding to the newly added tag text;

[0014] The first calculation unit is used to calculate the first similarity between the first multimedia feature and the second multimedia feature corresponding to the existing tag text in the tag library;

[0015] The first adding unit is configured to add the newly added tag text and the first multimedia feature corresponding to the newly added tag text to the tag library if the first similarity between the first multimedia feature and each of the second multimedia features is less than the first similarity threshold.

[0016] Thirdly, embodiments of this application provide an electronic device, including:

[0017] One or more processors;

[0018] Storage device, on which one or more programs are stored,

[0019] When the one or more programs are executed by the one or more processors, the one or more processors implement the tag text processing method as described above.

[0020] Fourthly, embodiments of this application provide a computer-readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the tag text processing method as described above.

[0021] Therefore, the embodiments of this application have the following beneficial effects:

[0022] In this embodiment, after obtaining the newly added tag text, the first multimedia feature of the multimedia information corresponding to the newly added tag text is obtained. A first similarity is calculated between the first multimedia feature and the second multimedia features corresponding to existing tag texts in the tag library. If the first similarity between the first multimedia feature and each of the second multimedia features is less than a first similarity threshold, it indicates that the newly added tag text is significantly different from the existing tag texts in the tag library and is not a similar tag to existing tag texts. In this case, the newly added tag text and its corresponding first multimedia feature can be added to the tag library, enabling rapid input of the newly added tag text and its corresponding multimedia features into the library. Attached Figure Description

[0023] Figure 1 A schematic diagram illustrating an exemplary application scenario provided in the embodiments of this application;

[0024] Figure 2 A flowchart illustrating a label text processing method provided in this application embodiment;

[0025] Figure 3 A flowchart illustrating another label text processing method provided in this application embodiment;

[0026] Figure 4A flowchart illustrating a scenario of a tag text processing method provided in this application embodiment;

[0027] Figure 5 A schematic diagram of a label text processing device provided in an embodiment of this application;

[0028] Figure 6 This is a schematic diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0029] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the embodiments of this application will be further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0030] To facilitate understanding and explanation of the technical solutions provided in the embodiments of this application, the background technology of this application will be described first.

[0031] Scene recognition is a deep semantic description of multimedia information such as videos and images. Through structured parsing of multimedia information, it identifies the context and events that occurred within the multimedia content. Scene recognition requires establishing a tag library, which stores tag text and its corresponding multimedia features. The tag text can be scene labels.

[0032] When scene recognition is required, the multimedia features of the multimedia information to be recognized are extracted and matched with the multimedia features in the tag library. The tag text corresponding to the matched multimedia features can be used as the tag text of the multimedia information to be recognized, thereby realizing the scene recognition of the multimedia information to be recognized.

[0033] With the development of the times, a lot of new tag text will be generated. We face problems such as how to determine the relationship between the new tag text and the existing tag text in the tag library, and whether to add the new tag text to the tag library. In other words, how to quickly process the new tag text is an urgent technical problem to be solved.

[0034] Based on this, embodiments of this application provide a method, apparatus, and device for processing tagged text. The method, after obtaining newly added tagged text, acquires a first multimedia feature of the multimedia information corresponding to the newly added tagged text. It calculates a first similarity between the first multimedia feature and second multimedia features corresponding to existing tagged text in a tag library. If the first similarity between the first multimedia feature and each of the second multimedia features is less than a first similarity threshold, it indicates that the newly added tagged text is significantly different from existing tagged text in the tag library and is not a similar tag to existing tagged text. In this case, the newly added tagged text and its corresponding first multimedia feature can be added to the tag library, achieving rapid input of the newly added tagged text and its corresponding multimedia features into the library.

[0035] To facilitate understanding of the label text processing method provided in the embodiments of this application, the following is combined with... Figure 1 The example scenario is shown below. See also... Figure 1 As shown in the figure, this figure is a schematic diagram of an exemplary application scenario provided in the embodiments of this application.

[0036] First, a tag library is pre-built, which includes existing tag texts and their corresponding multimedia features. In this embodiment, the multimedia feature corresponding to each existing tag text is a second multimedia feature, and each existing tag text corresponds to a second multimedia feature. In practical applications, the tag library can be stored in key-value format, where the key represents the existing tag text and the value represents the corresponding second multimedia feature. All existing tag texts in the tag library can serve as a tag system.

[0037] When there is a need to add a new tag to the tag library, the text of the new tag is obtained, and the first multimedia feature of the corresponding multimedia information can be obtained from the new tag text. That is, there is a correspondence between the new tag text and the first multimedia feature.

[0038] Calculate the first similarity between the first multimedia feature and each of the second multimedia features in the tag library. If the first similarity between the first multimedia feature and each of the second multimedia features is less than the first similarity threshold, the new tag text and its corresponding first multimedia feature can be added to the tag library. In practical applications, the new tag text and its corresponding first multimedia feature can be added to the tag library in key-value pairs.

[0039] After a new tag text and its corresponding first multimedia feature are added to the tag library, the new tag text can be converted into an existing tag text in the tag library. The first multimedia feature corresponding to the new tag text becomes the second multimedia feature corresponding to the converted existing tag text. This achieves the updating of the tag library.

[0040] Those skilled in the art will understand that Figure 1 The schematic diagram shown is merely one example in which the embodiments of this application can be implemented. The scope of application of the embodiments of this application is not limited by any aspect of this framework.

[0041] To facilitate understanding of this application, the following description, in conjunction with the accompanying drawings, illustrates a label text processing method provided by an embodiment of this application.

[0042] S201: Get the text of the newly added tag.

[0043] When there is a need to add a new tag to the tag library, the first step is to retrieve the text of the new tag. The text of the new tag consists of one or more characters.

[0044] S202: Obtain the first multimedia feature of the multimedia information corresponding to the newly added tag text.

[0045] Multimedia information corresponding to the semantics of newly added tag text can be obtained from the newly added tag text. The multimedia features corresponding to the newly added tag text are extracted as the first multimedia feature, that is, there is a correspondence between the newly added tag text and the first multimedia feature. Multimedia information can be video, image, audio, etc., and correspondingly, multimedia features can be video features, image features, audio features, etc., which can be represented by feature vectors.

[0046] In practical applications, multimedia information corresponding to newly added tag text can be retrieved using a general search engine. Therefore, in one possible implementation, the specific implementation of obtaining the first multimedia feature of the multimedia information corresponding to the newly added tag text in step S202 may include:

[0047] Using the newly added tag text as search keywords, at least one multimedia information corresponding to the newly added tag text is retrieved; the mean value of the multimedia features of at least one multimedia information is calculated as the first multimedia feature.

[0048] In this embodiment, by using the newly added tag text for retrieval, at least one multimedia information corresponding to the semantics of the newly added tag text can be obtained; multimedia features of each of the at least one multimedia information are extracted, and the average feature value of these multimedia features is calculated as the first multimedia feature. That is, the first multimedia feature can comprehensively represent at least one multimedia information corresponding to the newly added tag text.

[0049] In some possible implementations, after extracting at least one multimedia feature of multimedia information, discrete multimedia features can be removed by clustering, and then the feature mean of these multimedia features can be calculated as the first multimedia feature.

[0050] The number of multimedia information corresponding to the newly retrieved tag text can be set according to the actual situation, and this application embodiment does not limit this. For example, by using the newly retrieved tag text to retrieve 100 videos that correspond to the semantics of the newly retrieved tag text, the video features of each of the 100 videos are extracted, and the average feature value of the 100 video features is calculated as the first multimedia feature corresponding to the newly retrieved tag text.

[0051] S203: Calculate the first similarity between the first multimedia feature and the second multimedia feature corresponding to the existing tag text in the tag library.

[0052] The tag library stores existing tag texts and their corresponding multimedia features. In practical applications, the tag library can be stored in a key-value format, where the key represents the existing tag text and the value represents the corresponding multimedia feature.

[0053] Calculate the first similarity between the first multimedia feature and the second multimedia feature corresponding to existing tag text in the tag library. By comparing the first similarity between the first multimedia feature and the second multimedia feature, it can be determined whether the newly added tag text is similar to existing tag text in the tag library.

[0054] S204: If the first similarity between the first multimedia feature and each of the second multimedia features is less than the first similarity threshold, add the newly added label text and the first multimedia feature corresponding to the newly added label text to the label library.

[0055] If the first similarity between the first multimedia feature and each of the second multimedia features in the tag library is less than the first similarity threshold, it indicates that the newly added tag text is dissimilar to all existing tag texts, and the newly added tag text and its corresponding first multimedia feature can be added to the tag library. The first similarity threshold can be set according to actual circumstances, and this embodiment does not limit it. In practical applications, the newly added tag text and its corresponding first multimedia feature can be added to the tag library in key-value pairs.

[0056] Based on the content of S201-S204, this embodiment of the application, after obtaining the newly added tag text, obtains the first multimedia feature of the multimedia information corresponding to the newly added tag text. It then calculates the first similarity between the first multimedia feature and the second multimedia features corresponding to existing tag texts in the tag library. If the first similarity between the first multimedia feature and each of the second multimedia features is less than the first similarity threshold, it indicates that the newly added tag text is significantly different from the existing tag texts in the tag library and is not a similar tag to existing tag texts. In this case, the newly added tag text and its corresponding first multimedia feature can be added to the tag library, achieving rapid entry of the newly added tag text and its corresponding multimedia features into the library.

[0057] Based on the above embodiments, some possible implementations of this application may further include:

[0058] A1: If the first similarity between the first multimedia feature and any second multimedia feature is greater than or equal to the second similarity threshold, the existing labeled text corresponding to the target second multimedia feature is determined as the target existing labeled text; the target second multimedia feature is the second multimedia feature corresponding to the first similarity being greater than or equal to the second similarity threshold.

[0059] A2: Add the new tag text to the tag library and establish similar tag relationships between the new tag text and the target existing tag text.

[0060] In this embodiment of the application, if the first similarity between the first multimedia feature and any one or more second multimedia features in the tag library is greater than or equal to the second similarity threshold, it indicates that the newly added tag text is similar to the existing tag text corresponding to these second multimedia features.

[0061] First, the second multimedia feature corresponding to a first similarity greater than or equal to a second similarity threshold is obtained as the target second multimedia feature. The existing tag text corresponding to the target second multimedia feature is determined as the target existing tag text. Then, the newly added tag text and the target existing tag text are similar tags. At this time, only the newly added tag text is added to the tag library, and the similarity tag relationship between the newly added tag text and the target existing tag text is bound. The second similarity threshold can be set according to the actual situation. The second similarity threshold can be the same as or different from the first similarity threshold. This application embodiment does not limit this.

[0062] Since the multimedia features of the newly added tag text are similar to those of the target existing tag text, there is no need to repeatedly add the multimedia features of the newly added tag text to the tag library. Instead, a similar tag relationship between the newly added tag text and the target existing tag text is directly established, so that the newly added tag text and the target existing tag text simultaneously correspond to the multimedia features of the target existing tag text. This saves storage resources of the tag library and realizes the binding of the similar tag relationship between the newly added tag text and the target existing tag text.

[0063] Since newly added tag text may have multiple semantics or the semantics are relatively abstract, it may not be suitable to add the newly added tag text to the tag library. Therefore, in this embodiment of the application, it is also possible to first determine whether the newly added tag text has ambiguity.

[0064] See Figure 3 As shown, this figure is a flowchart of another label text processing method provided in an embodiment of this application. Figure 3 As shown, the method may include S301-S311:

[0065] S301: Get the text of the newly added tag.

[0066] After obtaining the newly added tag text, before obtaining the first multimedia feature of the multimedia information corresponding to the newly added tag text, you can first execute S302-S308.

[0067] S302: Using the information tag text as a search keyword, retrieve at least one multimedia information corresponding to the newly added tag text, and sort the at least one multimedia information according to its relevance to the newly added tag text.

[0068] Generally, search engines rank search results according to their relevance to the search keywords. Therefore, when searching using newly added tag text in a general search engine, at least one multimedia message will be obtained, ranked according to its relevance to the newly added tag text. The number of multimedia messages corresponding to the newly added tag text can be set according to actual needs, and this application embodiment does not limit this. For example, searching using newly added tag text may yield the top 100 videos that correspond to the semantics of the newly added tag text.

[0069] S303: Obtain multimedia information within a first relevance ranking range and multimedia information within a second relevance ranking range from at least one set of multimedia information; the first relevance ranking range precedes the second relevance ranking range.

[0070] In this embodiment, at least one retrieved multimedia information can be divided into two parts: multimedia information within a first relevance ranking range and multimedia information within a second relevance ranking range. For example, the top 100 videos corresponding to the semantics of the newly added tag text are obtained, from which videos ranked 1-50 and videos ranked 51-100 are extracted, where ranked 1-50 is the first relevance ranking range and ranked 51-100 is the second relevance ranking range.

[0071] S304: Calculate the mean value of the multimedia features of multimedia information within the first relevance ranking range as the third multimedia feature.

[0072] Multimedia features are extracted from the multimedia information within the first relevance ranking range, and the mean of these multimedia features is calculated as the third multimedia feature. That is, the third multimedia feature can comprehensively represent the multimedia information within the first relevance ranking range.

[0073] In some possible implementations, after extracting the multimedia features of multimedia information within the first relevance ranking range, discrete multimedia features can be removed by clustering, and then the feature mean of these multimedia features can be calculated as the third multimedia feature.

[0074] S305: Calculate the mean value of the multimedia features of multimedia information within the second relevance ranking range as the fourth multimedia feature.

[0075] Multimedia features are extracted from the multimedia information within the second relevance ranking range, and the mean of these multimedia features is calculated as the fourth multimedia feature. That is, the fourth multimedia feature can comprehensively represent the multimedia information within the second relevance ranking range.

[0076] In some possible implementations, after extracting the multimedia features of multimedia information within the second relevance ranking range, discrete multimedia features can be removed by clustering, and then the feature mean of these multimedia features can be calculated as the fourth multimedia feature.

[0077] S306: Calculate the second similarity between the third multimedia feature and the fourth multimedia feature.

[0078] If the multimedia information retrieved using the same newly added tag text within the first relevance ranking range is dissimilar to the multimedia information within the second relevance ranking range, it indicates that the newly added tag text has multiple semantics or its semantics are relatively abstract. Therefore, the ambiguity of the newly added tag text can be determined by calculating the second similarity between the third and fourth multimedia features.

[0079] S307: If the second similarity is greater than or equal to the third similarity threshold, obtain the first multimedia feature of the multimedia information corresponding to the newly added tag text.

[0080] If the second similarity is greater than or equal to the third similarity threshold, it means that the multimedia information within the first relevance ranking range is relatively similar to the multimedia information within the second relevance ranking range. It is then possible to further determine whether the newly added tag text needs to be added to the tag library, i.e., to execute the first multimedia feature acquisition process for the multimedia information corresponding to the newly added tag text, as well as steps S309-S311. The third similarity threshold can be set according to actual circumstances, and this embodiment does not limit its implementation.

[0081] In this embodiment, since the third multimedia feature has already been calculated, and the third multimedia feature comprehensively represents the multimedia information within the first relevance ranking range, the third multimedia feature can be directly determined as the first multimedia feature of the multimedia information corresponding to the newly added tag text.

[0082] S308: If the second similarity is less than the third similarity threshold, the newly added label text is determined to be ambiguous.

[0083] If the second similarity threshold is less than the third similarity threshold, it indicates a significant difference between the multimedia information within the first relevance ranking range and the multimedia information within the second relevance ranking range, thus confirming that the newly added tag text is ambiguous. This newly added tag text is unsuitable for inclusion in the tag library and can trigger a manual correction process to adjust it.

[0084] S309: Calculate the first similarity between the first multimedia feature and the second multimedia feature corresponding to the existing tag text in the tag library.

[0085] S310: If the first similarity between the first multimedia feature and each of the second multimedia features is less than the first similarity threshold, add the newly added label text and the first multimedia feature corresponding to the newly added label text to the label library.

[0086] S311: If the first similarity between the first multimedia feature and any second multimedia feature is greater than or equal to the second similarity threshold, the existing tag text corresponding to the target second multimedia feature is determined as the target existing tag text; the target second multimedia feature is the second multimedia feature corresponding to the first similarity greater than or equal to the second similarity threshold; the newly added tag text is added to the tag library, and a similarity tag relationship is established between the newly added tag text and the target existing tag text.

[0087] S309-S311 are similar to S203-S204 and A1-A2 in the above embodiments. For related descriptions, please refer to the above embodiments, and they will not be repeated here.

[0088] Based on S301-S311, in this embodiment of the application, after obtaining the newly added tag text, at least one multimedia information sorted according to its relevance to the newly added tag text is retrieved. By comparing whether the multimedia features of multimedia information in different relevance sorting ranges are similar, it is determined whether the newly added tag text has ambiguity, thereby timely discovering newly added tag texts that are not suitable for being entered into the tag text database.

[0089] See Figure 4 As shown, taking video as an example of multimedia information, the tag text processing method provided in this application embodiment will be further explained.

[0090] After obtaining the new tag text, the top 100 videos corresponding to the semantics of the new tag text are retrieved through a video search engine. Videos ranked 1-50 and 51-100 are then extracted. The third multimedia feature corresponding to the videos ranked 1-50 is obtained, and the fourth multimedia feature corresponding to the videos ranked 51-100 is obtained. A second similarity is calculated between the third and fourth multimedia features. If the second similarity is greater than or equal to the third similarity threshold, a first similarity is calculated between the first multimedia feature corresponding to the new tag text (which is the third multimedia feature in this embodiment) and the second multimedia feature corresponding to existing tag texts in the tag library. If the similarity is less than the third similarity threshold, the new tag text is determined to be ambiguous.

[0091] The first similarity score can also be used to determine whether to add the new tagged text and its corresponding first multimedia feature to the tag library. Specifically, if the first similarity between the first multimedia feature and each of the second multimedia features is less than the first similarity threshold, the new tagged text and its corresponding first multimedia feature are added to the tag library. If the first similarity between the first multimedia feature and any of the second multimedia features is greater than or equal to the second similarity threshold, the new tagged text is added to the tag library, and a similarity tag relationship is established between the existing tagged text corresponding to that second multimedia feature and the new tagged text. This enables rapid processing of the new tagged text.

[0092] In some possible implementations, multimedia features can include multi-dimensional multimedia features, such as visual features, multimodal features, and atmospheric features. Visual features can characterize the visual aspects of multimedia information, multimodal features can characterize specific visual characteristics of multimedia information, such as which abstract activity the current multimedia information represents, and atmospheric features can characterize various complex visual features of multimedia information. Therefore, when adding multimedia features to the tag library, the multimedia features can be categorized and stored according to their dimensions.

[0093] Specifically, in this embodiment, the first multimedia feature and the second multimedia feature may include multimedia features of multiple dimensions; therefore, calculating the first similarity between the first multimedia feature and the second multimedia feature corresponding to existing tag text in the tag library may specifically include:

[0094] Calculate the first similarity between the first multimedia feature of the same dimension and the second multimedia feature corresponding to the existing tag text in the tag library.

[0095] That is, when the first multimedia feature and the second multimedia feature can include multimedia features of multiple dimensions, the first similarity can be calculated using any one of the multimedia features. For example, the first similarity can be calculated between the multimodal features in the first multimedia feature and the multimodal features in the second multimedia feature.

[0096] Based on this, a specific implementation of calculating the feature mean of at least one multimedia feature of multimedia information as the first multimedia feature may include:

[0097] Extract multimedia features from at least one multimedia information from multiple dimensions, and calculate the feature mean of each dimension of multimedia features as the first multimedia feature.

[0098] To obtain a first multimedia feature that includes multiple dimensions, multimedia features of at least one multimedia information can be extracted, such as visual features, multimodal features, and atmospheric features of at least one multimedia information. Then, the mean value of each dimension of the multimedia feature is calculated, and these mean values ​​are collectively used as the first multimedia feature. Thus, the first multimedia feature includes multimedia features of multiple dimensions.

[0099] Similarly, the specific implementation of calculating the mean value of multimedia features of multimedia information within the first relevance ranking range as the third multimedia feature may include:

[0100] Multimedia features of multiple dimensions of multimedia information within the first relevance ranking range are extracted, and the feature mean of multimedia features of each dimension is calculated as the third multimedia feature.

[0101] The specific implementation of calculating the mean value of multimedia features of multimedia information within the second relevance ranking range as the fourth multimedia feature may include:

[0102] Multimedia features of multiple dimensions of multimedia information within the second relevance ranking range are extracted, and the feature mean of each dimension of multimedia features is calculated as the fourth multimedia feature.

[0103] The specific implementation for calculating the second similarity between the third multimedia feature and the fourth multimedia feature may include:

[0104] Calculate the second similarity between the third and fourth multimedia features of the same dimension.

[0105] In other words, to obtain a third multimedia feature that includes multiple dimensions, multimedia features of multiple dimensions can be extracted from the multimedia information within the first relevance ranking range. Then, the mean value of the multimedia feature for each dimension is calculated, and these mean values ​​are collectively used as the third multimedia feature. In this way, the third multimedia feature includes multimedia features of multiple dimensions.

[0106] To obtain a fourth multimedia feature that includes multiple dimensions, multimedia features of multiple dimensions can be extracted from the multimedia information within the second relevance ranking range. Then, the mean value of each dimension's multimedia feature is calculated, and these mean values ​​are collectively used as the fourth multimedia feature. Thus, the fourth multimedia feature includes multimedia features across multiple dimensions.

[0107] When calculating the second similarity, any dimension of multimedia features can be selected for calculation. For example, the second similarity can be calculated between the multimodal features in the third multimedia feature and the multimodal features in the fourth multimedia feature.

[0108] In some possible implementations, the specific implementation of adding the new tag text and the corresponding first multimedia feature to the tag library may include:

[0109] New tag text and the first multimedia features of each dimension corresponding to the new tag text are added to the tag library.

[0110] In this embodiment of the application, when it is necessary to add the newly added tag text and the corresponding first multimedia feature to the tag library, the first multimedia feature can be added to the tag library according to different dimensions, so as to enrich the multimedia features in the tag library and facilitate the use of the tag library.

[0111] Based on the label text processing method provided in the above-described embodiments, this application also provides a label text processing device, which will be described below with reference to the accompanying drawings.

[0112] See Figure 5 As shown in the figure, this is a schematic diagram of the structure of a label text processing device provided in an embodiment of this application. Figure 5 As shown, the label text processing device includes:

[0113] The first acquisition unit 501 is used to acquire the newly added tag text;

[0114] The second acquisition unit 502 is used to acquire the first multimedia feature of the multimedia information corresponding to the newly added tag text;

[0115] The first calculation unit 503 is used to calculate the first similarity between the first multimedia feature and the second multimedia feature corresponding to the existing tag text in the tag library;

[0116] The first adding unit 504 is used to add the newly added tag text and the first multimedia feature corresponding to the newly added tag text to the tag library if the first similarity between the first multimedia feature and each of the second multimedia features is less than the first similarity threshold.

[0117] In one possible implementation, the device further includes:

[0118] The determining unit is configured to determine the existing labeled text corresponding to the target second multimedia feature as the target existing labeled text if the first similarity between the first multimedia feature and any second multimedia feature is greater than or equal to a second similarity threshold; the target second multimedia feature is the second multimedia feature corresponding to the first similarity being greater than or equal to the second similarity threshold.

[0119] The second adding unit is used to add the newly added tag text to the tag library and establish a similar tag relationship between the newly added tag text and the target existing tag text.

[0120] In one possible implementation, the second acquiring unit includes:

[0121] The retrieval subunit is used to retrieve at least one multimedia information corresponding to the newly added tag text by using the newly added tag text as a retrieval keyword.

[0122] A calculation subunit is used to calculate the feature mean of the multimedia features of the at least one multimedia information as a first multimedia feature.

[0123] In one possible implementation, the device further includes:

[0124] The retrieval unit is used to retrieve at least one multimedia information corresponding to the newly added tag text by using the information tag text as a retrieval keyword, wherein the at least one multimedia information is sorted according to its relevance to the newly added tag text.

[0125] The third acquisition unit is configured to acquire multimedia information within a first relevance ranking range and multimedia information within a second relevance ranking range from the at least one multimedia information; the first relevance ranking range precedes the second relevance ranking range.

[0126] The second calculation unit is used to calculate the feature mean of the multimedia features of the multimedia information within the first relevance ranking range as the third multimedia feature.

[0127] The third calculation unit is used to calculate the feature mean of the multimedia features of the multimedia information within the second relevance ranking range as the fourth multimedia feature.

[0128] The fourth calculation unit is used to calculate the second similarity between the third multimedia feature and the fourth multimedia feature;

[0129] The triggering unit is configured to trigger the second acquisition unit to execute the acquisition of the first multimedia feature corresponding to the multimedia information of the newly added tag text if the second similarity is greater than or equal to the third similarity threshold.

[0130] In one possible implementation, the second acquisition unit is specifically used for:

[0131] The third multimedia feature is determined as the first multimedia feature of the multimedia information corresponding to the newly added tag text.

[0132] In one possible implementation, the device further includes:

[0133] The determining unit is configured to determine that the newly added label text is ambiguous if the second similarity is less than a third similarity threshold.

[0134] In one possible implementation, the first multimedia feature and the second multimedia feature include multimedia features of multiple dimensions; the first computing unit is specifically used for:

[0135] Calculate the first similarity between the first multimedia feature of the same dimension and the second multimedia feature corresponding to the existing tag text in the tag library.

[0136] In one possible implementation, the first adding unit is specifically used for:

[0137] The newly added tag text and the first multimedia features of each dimension corresponding to the newly added tag text are added to the tag library.

[0138] In one possible implementation, the first multimedia feature includes multimedia features of multiple dimensions; the computational subunit is specifically used for:

[0139] Multimedia features of multiple dimensions of the at least one multimedia information are extracted, and the feature mean of each dimension of the multimedia feature is calculated as the first multimedia feature.

[0140] In one possible implementation, the third multimedia feature and the fourth multimedia feature include multimedia features of multiple dimensions;

[0141] The second computing unit is specifically used for:

[0142] Multimedia features of multiple dimensions of multimedia information within the first relevance ranking range are extracted, and the feature mean of each dimension of multimedia features is calculated as the third multimedia feature.

[0143] The third computing unit is specifically used for:

[0144] Multimedia features of multiple dimensions of multimedia information within the second relevance ranking range are extracted, and the feature mean of each dimension of multimedia features is calculated as the fourth multimedia feature.

[0145] The fourth calculation unit is specifically used for:

[0146] Calculate the second similarity between the third and fourth multimedia features of the same dimension.

[0147] Based on the label text processing method provided in the above-described method embodiments, this application also provides an electronic device, including: one or more processors; a storage device storing one or more programs thereon, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the label text processing method described in any of the above embodiments.

[0148] The following is for reference. Figure 6 The diagram illustrates a structural schematic of an electronic device 1300 suitable for implementing embodiments of this application. The terminal devices in these embodiments may include, but are not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Android Devices), PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs (televisions), desktop computers, etc. Figure 6 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0149] like Figure 6 As shown, electronic device 1300 may include a processing device (e.g., a central processing unit, a graphics processor, etc.) 1301, which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 1302 or a program loaded from storage device 1306 into random access memory (RAM) 1303. RAM 1303 also stores various programs and data required for the operation of electronic device 1300. Processing device 1301, ROM 1302, and RAM 1303 are interconnected via bus 1304. Input / output (I / O) interface 1305 is also connected to bus 1304.

[0150] Typically, the following devices can be connected to I / O interface 1305: input devices 1306 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 1307 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1306 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1309. Communication device 1309 allows electronic device 1300 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 6 An electronic device 1300 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively.

[0151] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication device 1309, or installed from storage device 1306, or installed from ROM 1302. When the computer program is executed by processing device 1301, it performs the functions defined in the methods of embodiments of this application.

[0152] The electronic device provided in this application embodiment belongs to the same inventive concept as the label text processing method provided in the above embodiment. Technical details not described in detail in this embodiment can be found in the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0153] Based on the label text processing method provided in the above method embodiments, this application provides a computer-readable medium storing a computer program thereon, wherein the program, when executed by a processor, implements the label text processing method as described in any of the above embodiments.

[0154] It should be noted that the computer-readable medium described above in this application can be a computer-readable signal medium, a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0155] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0156] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device.

[0157] The aforementioned computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to perform the aforementioned tag text processing method.

[0158] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof. These programming languages ​​include, but are not limited to, object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0159] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0160] The units described in the embodiments of this application can be implemented in software or in hardware. The name of the unit / module does not necessarily limit the unit itself; for example, a voice data acquisition module can also be described as a "data acquisition module".

[0161] The functions described above in this document can be performed, at least in part, by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: Field Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), Application Standard Products (ASSPs), System-on-Chip (SoCs), Complex Programmable Logic Devices (CPLDs), and so on.

[0162] In the context of this application, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0163] According to one or more embodiments of this application, [Example 1] provides a label text processing method, the method comprising:

[0164] Get the text of the newly added tag;

[0165] Obtain the first multimedia feature of the multimedia information corresponding to the newly added tag text;

[0166] Calculate the first similarity between the first multimedia feature and the second multimedia feature corresponding to the existing tag text in the tag library;

[0167] If the first similarity between the first multimedia feature and each of the second multimedia features is less than the first similarity threshold, the newly added tag text and the first multimedia feature corresponding to the newly added tag text are added to the tag library.

[0168] According to one or more embodiments of this application, [Example 2] provides a label text processing method, the method further comprising:

[0169] If the first similarity between the first multimedia feature and any of the second multimedia features is greater than or equal to the second similarity threshold, the existing labeled text corresponding to the target second multimedia feature is determined as the target existing labeled text; the target second multimedia feature is the second multimedia feature corresponding to the first similarity being greater than or equal to the second similarity threshold;

[0170] The newly added tag text is added to the tag library, and a similar tag relationship is established between the newly added tag text and the target existing tag text.

[0171] According to one or more embodiments of this application, [Example 3] provides a tag text processing method, wherein obtaining the first multimedia feature of the multimedia information corresponding to the newly added tag text includes:

[0172] Using the newly added tag text as a search keyword, at least one multimedia information corresponding to the newly added tag text can be retrieved.

[0173] The mean value of the multimedia features of the at least one multimedia information is calculated as the first multimedia feature.

[0174] According to one or more embodiments of this application, [Example 4] provides a tag text processing method, which further includes, before obtaining the first multimedia feature of the multimedia information corresponding to the newly added tag text:

[0175] Using the information tag text as search keywords, at least one multimedia information corresponding to the newly added tag text is retrieved, and the at least one multimedia information is sorted according to its relevance to the newly added tag text;

[0176] Multimedia information within a first relevance ranking range and multimedia information within a second relevance ranking range are obtained from the at least one multimedia information; the first relevance ranking range precedes the second relevance ranking range.

[0177] The mean value of the multimedia features of multimedia information within the first relevance ranking range is used as the third multimedia feature.

[0178] The mean value of the multimedia features of multimedia information within the second relevance ranking range is calculated as the fourth multimedia feature.

[0179] Calculate the second similarity between the third multimedia feature and the fourth multimedia feature;

[0180] If the second similarity is greater than or equal to the third similarity threshold, the first multimedia feature for obtaining the multimedia information corresponding to the newly added tag text is executed.

[0181] According to one or more embodiments of this application, [Example 5] provides a tag text processing method, wherein obtaining the first multimedia feature of the multimedia information corresponding to the newly added tag text includes:

[0182] The third multimedia feature is determined as the first multimedia feature of the multimedia information corresponding to the newly added tag text.

[0183] According to one or more embodiments of this application, [Example Six] provides a label text processing method, the method further comprising:

[0184] If the second similarity is less than the third similarity threshold, the newly added label text is determined to be ambiguous.

[0185] According to one or more embodiments of this application, [Example Seven] provides a tag text processing method, wherein the first multimedia feature and the second multimedia feature include multimedia features of multiple dimensions; the step of calculating a first similarity between the first multimedia feature and the second multimedia feature corresponding to existing tag text in the tag library includes:

[0186] Calculate the first similarity between the first multimedia feature of the same dimension and the second multimedia feature corresponding to the existing tag text in the tag library.

[0187] According to one or more embodiments of this application, [Example Eight] provides a tag text processing method, wherein adding the newly added tag text and the first multimedia feature corresponding to the newly added tag text to the tag library includes:

[0188] The newly added tag text and the first multimedia features of each dimension corresponding to the newly added tag text are added to the tag library.

[0189] According to one or more embodiments of this application, [Example Nine] provides a tag text processing method, wherein the first multimedia feature includes multimedia features of multiple dimensions; the step of calculating the feature mean of the multimedia features of the at least one multimedia information as the first multimedia feature includes:

[0190] Multimedia features of multiple dimensions of the at least one multimedia information are extracted, and the feature mean of each dimension of the multimedia feature is calculated as the first multimedia feature.

[0191] According to one or more embodiments of this application, [Example 10] provides a tag text processing method, wherein the third multimedia feature and the fourth multimedia feature include multimedia features of multiple dimensions;

[0192] The step of calculating the mean value of the multimedia features of multimedia information within the first relevance ranking range as the third multimedia feature includes:

[0193] Multimedia features of multiple dimensions of multimedia information within the first relevance ranking range are extracted, and the feature mean of each dimension of multimedia features is calculated as the third multimedia feature.

[0194] The calculation of the mean value of the multimedia features of multimedia information within the second relevance ranking range as the fourth multimedia feature includes:

[0195] Multimedia features of multiple dimensions of multimedia information within the second relevance ranking range are extracted, and the feature mean of each dimension of multimedia features is calculated as the fourth multimedia feature.

[0196] The calculation of the second similarity between the third multimedia feature and the fourth multimedia feature includes:

[0197] Calculate the second similarity between the third and fourth multimedia features of the same dimension.

[0198] According to one or more embodiments of this application, [Example 11] provides a tag text processing apparatus, the apparatus comprising:

[0199] The first acquisition unit is used to acquire the text of the newly added tag;

[0200] The second acquisition unit is used to acquire the first multimedia feature of the multimedia information corresponding to the newly added tag text;

[0201] The first calculation unit is used to calculate the first similarity between the first multimedia feature and the second multimedia feature corresponding to the existing tag text in the tag library;

[0202] The first adding unit is configured to add the newly added tag text and the first multimedia feature corresponding to the newly added tag text to the tag library if the first similarity between the first multimedia feature and each of the second multimedia features is less than the first similarity threshold.

[0203] According to one or more embodiments of this application, [Example Twelve] provides a tag text processing apparatus, the apparatus further comprising:

[0204] The determining unit is configured to determine the existing labeled text corresponding to the target second multimedia feature as the target existing labeled text if the first similarity between the first multimedia feature and any second multimedia feature is greater than or equal to a second similarity threshold; the target second multimedia feature is the second multimedia feature corresponding to the first similarity being greater than or equal to the second similarity threshold.

[0205] The second adding unit is used to add the newly added tag text to the tag library and establish a similar tag relationship between the newly added tag text and the target existing tag text.

[0206] According to one or more embodiments of this application, [Example Thirteen] provides a tag text processing apparatus, wherein the second acquisition unit includes:

[0207] The retrieval subunit is used to retrieve at least one multimedia information corresponding to the newly added tag text by using the newly added tag text as a retrieval keyword.

[0208] A calculation subunit is used to calculate the feature mean of the multimedia features of the at least one multimedia information as a first multimedia feature.

[0209] According to one or more embodiments of this application, [Example Fourteen] provides a tag text processing apparatus, the apparatus further comprising:

[0210] The retrieval unit is used to retrieve at least one multimedia information corresponding to the newly added tag text by using the information tag text as a retrieval keyword, wherein the at least one multimedia information is sorted according to its relevance to the newly added tag text.

[0211] The third acquisition unit is configured to acquire multimedia information within a first relevance ranking range and multimedia information within a second relevance ranking range from the at least one multimedia information; the first relevance ranking range precedes the second relevance ranking range.

[0212] The second calculation unit is used to calculate the feature mean of the multimedia features of the multimedia information within the first relevance ranking range as the third multimedia feature.

[0213] The third calculation unit is used to calculate the feature mean of the multimedia features of the multimedia information within the second relevance ranking range as the fourth multimedia feature.

[0214] The fourth calculation unit is used to calculate the second similarity between the third multimedia feature and the fourth multimedia feature;

[0215] The triggering unit is configured to trigger the second acquisition unit to execute the acquisition of the first multimedia feature corresponding to the multimedia information of the newly added tag text if the second similarity is greater than or equal to the third similarity threshold.

[0216] According to one or more embodiments of this application, [Example Fifteen] provides a tag text processing apparatus, wherein the second acquisition unit is specifically used for:

[0217] The third multimedia feature is determined as the first multimedia feature of the multimedia information corresponding to the newly added tag text.

[0218] According to one or more embodiments of this application, [Example Sixteen] provides a tag text processing apparatus, the apparatus further comprising:

[0219] The determining unit is configured to determine that the newly added label text is ambiguous if the second similarity is less than a third similarity threshold.

[0220] According to one or more embodiments of this application, [Example Seventeen] provides a tag text processing device, wherein the first multimedia feature and the second multimedia feature include multimedia features of multiple dimensions; the first computing unit is specifically used for:

[0221] Calculate the first similarity between the first multimedia feature of the same dimension and the second multimedia feature corresponding to the existing tag text in the tag library.

[0222] According to one or more embodiments of this application, [Example 18] provides a tag text processing apparatus, wherein the first adding unit is specifically used for:

[0223] The newly added tag text and the first multimedia features of each dimension corresponding to the newly added tag text are added to the tag library.

[0224] According to one or more embodiments of this application, [Example Nineteen] provides a tag text processing device, wherein the first multimedia feature includes multimedia features of multiple dimensions; the computing subunit is specifically used for:

[0225] Multimedia features of multiple dimensions of the at least one multimedia information are extracted, and the feature mean of each dimension of the multimedia feature is calculated as the first multimedia feature.

[0226] According to one or more embodiments of this application, [Example 20] provides a tag text processing device, wherein the third multimedia feature and the fourth multimedia feature include multimedia features of multiple dimensions;

[0227] The second computing unit is specifically used for:

[0228] Multimedia features of multiple dimensions of multimedia information within the first relevance ranking range are extracted, and the feature mean of each dimension of multimedia features is calculated as the third multimedia feature.

[0229] The third computing unit is specifically used for:

[0230] Multimedia features of multiple dimensions of multimedia information within the second relevance ranking range are extracted, and the feature mean of each dimension of multimedia features is calculated as the fourth multimedia feature.

[0231] The fourth calculation unit is specifically used for:

[0232] Calculate the second similarity between the third and fourth multimedia features of the same dimension.

[0233] According to one or more embodiments of this application, [Example Twenty-One] provides an electronic device, including:

[0234] One or more processors;

[0235] Storage device, on which one or more programs are stored,

[0236] When the one or more programs are executed by the one or more processors, the one or more processors implement the tag text processing method as described in any of [Example 1] to [Example 10].

[0237] According to one or more embodiments of this application, [Example 22] a computer-readable medium is characterized in that it stores a computer program thereon, wherein the program, when executed by a processor, implements the tag text processing method as described in any one of [Example 1] to [Example 10].

[0238] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems or apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple, and relevant parts can be referred to the method section.

[0239] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0240] It should also be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus 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 apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0241] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0242] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method of processing label text, characterized by, The method comprises: acquiring a new label text; retrieving at least one multimedia information corresponding to the new label text by taking the information label text as a retrieval keyword, and sorting the at least one multimedia information according to the relevance to the new label text; acquiring multimedia information within a first relevance sorting range and multimedia information within a second relevance sorting range from the at least one multimedia information; the first relevance sorting range is before the second relevance sorting range; calculating a feature mean of multimedia features of the multimedia information within the first relevance sorting range as a third multimedia feature; calculating a feature mean of multimedia features of the multimedia information within the second relevance sorting range as a fourth multimedia feature; calculating a second similarity between the third multimedia feature and the fourth multimedia feature; if the second similarity is greater than or equal to a third similarity threshold, acquiring a first multimedia feature of the multimedia information corresponding to the new label text; calculating a first similarity between the first multimedia feature and a second multimedia feature corresponding to an existing label text in a label library; if the first similarity between the first multimedia feature and each of the second multimedia features is less than a first similarity threshold, adding the new label text and the first multimedia feature corresponding to the new label text to the label library.

2. The method of claim 1, wherein, The method further comprises: if the first similarity between the first multimedia feature and any of the second multimedia features is greater than or equal to a second similarity threshold, determining an existing label text corresponding to a target second multimedia feature as a target existing label text; the target second multimedia feature is the second multimedia feature corresponding to the first similarity greater than or equal to the second similarity threshold; adding the new label text to the label library and establishing a similar label relationship between the new label text and the target existing label text.

3. The method of claim 1, wherein, The acquisition of the first multimedia feature of the multimedia information corresponding to the new label text comprises: retrieving at least one multimedia information corresponding to the new label text by taking the new label text as a retrieval keyword; calculating a feature mean of multimedia features of the at least one multimedia information as a first multimedia feature.

4. The method of claim 1, wherein, The acquisition of the first multimedia feature of the multimedia information corresponding to the new label text comprises: determining the third multimedia feature as the first multimedia feature of the multimedia information corresponding to the new label text.

5. The method of claim 1, wherein, The method further comprises: if the second similarity is less than the third similarity threshold, determining that the new label text is polysemous.

6. The method of claim 1, wherein, The first multimedia feature and the second multimedia feature comprise multimedia features of multiple dimensions; the calculation of the first similarity between the first multimedia feature and the second multimedia feature corresponding to the existing label text in the label library comprises: calculating the first similarity between the first multimedia feature and the second multimedia feature corresponding to the existing label text in the label library in the same dimension.

7. The method of claim 6, wherein, The adding the new label text and the first multimedia feature corresponding to the new label text into the label library comprises: The new label text and the first multimedia feature corresponding to each dimension of the new label text are added into the label library.

8. The method of claim 3, wherein, The first multimedia feature comprises multimedia features of multiple dimensions; and the calculating the feature mean of the multimedia feature of the at least one multimedia information as the first multimedia feature comprises: The multimedia features of multiple dimensions of the at least one multimedia information are extracted, and the feature mean of the multimedia feature of each dimension is calculated as the first multimedia feature.

9. The method of claim 1, wherein, The third multimedia feature and the fourth multimedia feature comprise multimedia features of multiple dimensions; The calculating the feature mean of the multimedia feature of the multimedia information in the first relevance sorting range as the third multimedia feature comprises: The multimedia features of multiple dimensions of the multimedia information in the first relevance sorting range are extracted, and the feature mean of the multimedia feature of each dimension is calculated as the third multimedia feature; The calculating the feature mean of the multimedia feature of the multimedia information in the second relevance sorting range as the fourth multimedia feature comprises: The multimedia features of multiple dimensions of the multimedia information in the second relevance sorting range are extracted, and the feature mean of the multimedia feature of each dimension is calculated as the fourth multimedia feature; The calculating the second similarity between the third multimedia feature and the fourth multimedia feature comprises: The second similarity between the third multimedia feature and the fourth multimedia feature of the same dimension is calculated.

10. A label text processing apparatus characterized by comprising: The apparatus comprises: A first obtaining unit configured to obtain a new label text; A retrieving unit configured to retrieve at least one multimedia information corresponding to the new label text by taking the information label text as a retrieval keyword, and sort the at least one multimedia information according to the relevance to the new label text; A third obtaining unit configured to obtain multimedia information in a first relevance sorting range and multimedia information in a second relevance sorting range from the at least one multimedia information; the first relevance sorting range is before the second relevance sorting range; A second calculating unit configured to calculate the feature mean of the multimedia feature of the multimedia information in the first relevance sorting range as a third multimedia feature; A third calculating unit configured to calculate the feature mean of the multimedia feature of the multimedia information in the second relevance sorting range as a fourth multimedia feature; A fourth calculating unit configured to calculate the second similarity between the third multimedia feature and the fourth multimedia feature; A second obtaining unit configured to, if the second similarity is greater than or equal to a third similarity threshold, obtain the first multimedia feature of the multimedia information corresponding to the new label text; A first calculating unit configured to calculate the first similarity between the first multimedia feature and the second multimedia feature corresponding to the existing label text in the label library; and A first obtaining unit configured to obtain a new label text; A retrieving unit configured to retrieve at least one multimedia information corresponding to the new label text by taking the information label text as a retrieval keyword, and sort the at least one multimedia information according to the relevance to the new label text; A third obtaining unit configured to obtain multimedia information in a first relevance sorting range and multimedia information in a second relevance sorting range from the at least one multimedia information; the first relevance sorting range is before the second relevance sorting range; A second calculating unit configured to calculate the feature mean of the multimedia feature of the multimedia information in the first relevance sorting range as a third multimedia feature; A third calculating unit configured to calculate the feature mean of the multimedia feature of the multimedia information in the second relevance sorting range as a fourth multimedia feature; A fourth calculating unit configured to calculate the second similarity between the third multimedia feature and the fourth multimedia feature; A second obtaining unit configured to, if the second similarity is greater than or equal to a third similarity threshold, obtain the first multimedia feature of the multimedia information corresponding to the new label text; A first calculating unit configured to calculate the first similarity between the first multimedia feature and the second multimedia feature corresponding to the existing label text in the label library; and The first adding unit is configured to add the new label text and the first multimedia feature corresponding to the new label text into the label library if the first similarity between the first multimedia feature and each of the second multimedia features is less than a first similarity threshold.

11. An electronic device, comprising: The method comprises: one or more processors; a memory device having stored thereon 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 label text processing method according to any one of claims 1-9.

12. A computer readable medium characterized by a computer program is stored thereon, and the program is executed by a processor to implement the label text processing method according to any one of claims 1-9.

Images