Material library supplementing method and device

Abstract

The embodiment of the invention provides a material library supplementing method and device, computer equipment, a computer readable storage medium and a computer program product, and belongs to the technical field of material libraries. The method comprises the following steps: acquiring a target media file; obtaining one or more keywords based on the target media file; according to a preset rule, determining a hotspot keyword and popularity information of the hotspot keyword from the one or more keywords; based on the hotspot keyword, the popularity information of the hotspot keyword, the target media file and existing materials in a material library, obtaining a to-be-supplemented material; and supplementing the to-be-supplemented materials into the material library. The real-time hot content in the target media file is identified, and the to-be-supplemented material is automatically acquired and supplemented into the material library according to the real-time hot content, so that the real-time updating of the material library is realized, and the timeliness of the material library is improved. Through the high-timeliness material library, the content creator can obtain the materials related to the real-time hot content in time, so that the use experience of the content creator is improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of material libraries, and in particular, to a method, device, computer device, computer-readable storage medium, and computer program product for supplementing a material library. Background Art

[0002] Content creators can enrich their creative content by obtaining materials from the material library. However, the existing material library is not updated in a timely manner, and there is a lack of real-time hot materials in the material library, resulting in a poor user experience for content creators.

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

[0004] The embodiments of the present application provide a method, device, computer device, computer-readable storage medium, and computer program product for supplementing a material library to solve or alleviate one or more of the above technical problems.

[0005] One aspect of the embodiments of the present application provides a method for supplementing a material library, the method comprising: Obtaining a target media file; Based on the target media file, obtaining one or more keywords; According to a preset rule, determining a hot keyword and heat information of the hot keyword from one or more of the keywords; Based on the hot keyword, the heat information of the hot keyword, the target media file, and the existing materials in the material library, obtaining materials to be supplemented; and Supplementing the materials to be supplemented into the material library.

[0006] Optionally, the target media file includes audio content; Based on the target media file, obtaining one or more keywords, including: Converting the audio content into text content; Inputting the text content into a pre-trained text analysis model to obtain each of the keywords and corresponding current word frequency data.

[0007] Optionally, according to a preset rule, determining a hot keyword and heat information of the hot keyword from one or more of the keywords, including: Obtaining historical word frequency data of each of the keywords; Based on the current word frequency data and historical word frequency data of each of the keywords, determining the long-term heat and short-term heat of the keyword; Determine the hot keywords and the heat information of the hot keywords according to the long-term heat and short-term heat of each of the keywords.

[0008] Optionally, each of the existing materials corresponds to a material keyword. Based on the hot keywords, the heat information of the hot keywords, the target media file, and the existing materials in the material library, obtain the materials to be supplemented, including: Based on the heat information of the hot keywords, determine a first value of the number of materials required for the hot keywords. Based on the hot keywords, the existing materials, and the material keywords of the existing materials, determine the existing materials that match the hot keywords, where the number of the matching existing materials is recorded as a second value. In the case where the second value is less than the first value, determine the hot keyword as the target hot keyword; and Based on the target hot keyword, obtain the materials to be supplemented from the target media file.

[0009] Optionally, the target media file includes target video content; based on the target hot keyword, obtaining the materials to be supplemented from the target media file includes: Obtain the target semantic vector of the target hot keyword. Obtain the image feature vectors corresponding to multiple frames of the target video content; where one frame corresponds to one timestamp. According to the target semantic vector and the multiple image feature vectors, determine the target frame and the corresponding target timestamp from the multiple frames; and Based on the target timestamp, determine a target video slice from the target video content, and the target video slice is used as the materials to be supplemented.

[0010] Optionally, supplement the materials to be supplemented into the material library, including: Based on the heat information of the hot keywords, determine a supplement order list for the materials to be supplemented; where the supplement order list includes multiple hot keywords and the materials to be supplemented corresponding to each hot keyword. According to the supplement order list, sequentially supplement the multiple materials to be supplemented into the material library in the order of priority.

[0011] Optionally, the method further includes: Check the materials to be supplemented, and determine whether the materials to be supplemented are added to the material library according to the check result.

[0012] Another aspect of the embodiments of the present application provides a material library supplementing device, which includes: An obtaining module, configured to obtain a target media file; and based on the target media file, obtain one or more keywords; A determining module, configured to determine a hot keyword and heat information of the hot keyword from one or more of the keywords according to a preset rule; The obtaining module is further configured to obtain to-be-supplemented materials based on the hot keyword, the heat information of the hot keyword, the target media file, and existing materials in the material library; A supplementing module, configured to supplement the to-be-supplemented materials into the material library.

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

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

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

[0016] The embodiments of the present application adopting the above technical solutions may include the following advantages: By obtaining the keywords of the target media file and determining the hot keywords and their heat information from the keywords according to the preset rule, the real-time hot content in the target media file can be identified. According to the real-time hot content, the to-be-supplemented materials are automatically obtained and supplemented into the material library, realizing the real-time update of the material library, thereby improving the timeliness of the material library. It can be seen that through the material library with high timeliness, content creators can obtain materials related to real-time hot content in a timely manner, thereby improving the use experience of content creators. BRIEF DESCRIPTION OF THE DRAWINGS

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

[0018] Figure 1 Schematically shows a flowchart of a material library supplement method according to Embodiment 1 of the present application; Figure 2 Schematically shows Figure 1 a sub-step flowchart of step S102 in Figure 3 Schematically shows Figure 1 a sub-step flowchart of step S104 in Figure 4 Schematically shows Figure 1 a sub-step flowchart of step S106 in Figure 5 Schematically shows Figure 1 another sub-step flowchart of step S106 in Figure 6 Schematically shows Figure 1 a sub-step flowchart of step S108 in Figure 7 Schematically shows an application example diagram of the material library supplement method according to Embodiment 1 of the present application; Figure 8 Schematically shows a block diagram of a material library supplement device according to Embodiment 2 of the present application; and Figure 9 Schematically shows a schematic diagram of the hardware architecture of a computer device according to Embodiment 3 of the present application. Detailed implementation manners

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

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

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

[0022] First, the following are the term explanations involved in the present application: ASR (Automatic Speech Recognition): It is a technology that converts speech signals into text data.

[0023] Entity word: It refers to a word with a specific meaning and reference object in the text, such as a person's name, product name, location, event, etc.

[0024] Hotspot monitoring: It refers to the real-time or regular monitoring of the popularity of specific entity words on channels such as social media and news platforms.

[0025] Material library: It refers to a database or storage system that stores various materials (such as pictures, videos, audio, etc.). The material library is the source for content creators to obtain creative materials.

[0026] UGC (User-Generated Content): It refers to materials (such as pictures, videos, text, etc.) created and uploaded by users themselves.

[0027] AIGC (AI-Generated Content): It refers to materials automatically generated through artificial intelligence technologies (such as deep learning models).

[0028] Long-term popularity: It refers to the average popularity value of a certain entity word over a relatively long period of time, which can reflect the continuous influence of the entity word.

[0029] Short-term popularity: It refers to the popularity change of a certain entity word in the near future (such as the past few days or weeks), which can reflect the immediate attention of the entity word.

[0030] Secondly, to facilitate those skilled in the art to understand the technical solutions provided by the embodiments of the present application, the related technologies are described below: Hot events and topics in the Internet emerge continuously, and the demand of content creators for real-time hot material is increasing day by day. However, the applicant has learned that there are many deficiencies in the supplement of hot material in the material library: the material update is lagging behind. The material library cannot monitor hot content in real time, resulting in a slow material update speed and difficulty in meeting the immediate needs of creators for hot material. The material sources are limited, mainly relying on existing material resources, lacking the ability of dynamic collection and generation of hot events, and it is difficult to cover diverse hot demands. The quality of materials is uneven. When supplementing hot materials, due to the lack of effective screening and review mechanisms, the quality of materials is not high or there is a risk of infringement. Relying on manual supplement of the material library, lacking intelligent hot monitoring and intelligent supplement functions.

[0031] For this reason, the embodiment of the present application provides a technical solution for supplementing the material library. In this technical solution, by obtaining the keywords of the target media file and determining the hot keywords and their popularity information from the keywords according to the preset rules, the real-time hot content in the target media file is identified. According to the real-time hot content, the material to be supplemented is automatically obtained and supplemented into the material library, realizing the real-time update of the material library, thereby improving the timeliness of the material library. Through the highly time-sensitive material library, content creators can obtain materials related to real-time hot content in a timely manner, thereby improving the usage experience of content creators. See the following for details.

[0032] Embodiment 1 Figure 1 The flowchart of the method for supplementing the material library according to Embodiment 1 of the present application is schematically shown.

[0033] As Figure 1 shown, the method for supplementing the material library may include steps S100 to S108, where: Step S100, obtain the target media file.

[0034] Step S102, based on the target media file, obtain one or more keywords.

[0035] Step S104, according to the preset rules, determine the hot keywords and the popularity information of the hot keywords from one or more of the keywords.

[0036] Step S106, based on the hot keywords, the popularity information of the hot keywords, the target media file and the existing materials in the material library, obtain the materials to be supplemented.

[0037] Step S108, supplement the materials to be supplemented into the material library.

[0038] The material library supplement method provided in this embodiment obtains the keywords of the target media file, determines the hot keywords and their popularity information from the keywords according to preset rules, so as to identify the real-time hot content in the target media file. According to the real-time hot content, the materials to be supplemented are automatically obtained and supplemented into the material library, realizing the real-time update of the material library, thereby improving the timeliness of the material library. Through the material library with high timeliness, content creators can obtain materials related to real-time hot content in a timely manner, thereby improving the usage experience of content creators.

[0039] The following combines Figure 1 to elaborate in detail on each step in steps S100 to S108 and other optional steps.

[0040] Step S100 , obtain the target media file.

[0041] The target media file can be obtained through various network platforms, social media websites, video sharing platforms, news websites and other platforms. It can include news videos, user-generated content (UGC), advertisements or live streams and other files.

[0042] In some embodiments, in order to improve the quality and efficiency of subsequent material extraction, before obtaining the target media file, the quality of the target media file can be screened first. For example, when the target media file is a video file generated by a user on a video sharing platform, multiple recently submitted video files on the video sharing platform can be obtained first. With authorization, the historical data of the submitter (such as the number of fans, interaction rate, creation quality) and the video data of the submitter (such as the number of views, like rate, comment volume) are analyzed through a machine learning model or preset rules to evaluate the quality, so as to screen out high-quality video files as the target media file.

[0043] Step S102 , based on the target media file, obtain one or more keywords.

[0044] In some embodiments, keywords can be extracted by means of natural language processing (NLP) technology, computer vision (CV) technology, speech recognition (ASR) technology, etc. For video files, the dialogue in the video can be converted into text through speech recognition, and then keywords can be extracted through text analysis technology. The obtained keywords can be entity words directly related to the video content, such as "person name", "geographical location", "brand name", etc. For example, if the target media file is an explanatory video about "metaverse", after speech recognition and text analysis, keywords such as "metaverse", "technology", "virtual reality" can be extracted.

[0045] Step S104, determine the hot keywords and the popularity information of the hot keywords from one or more of the said keywords according to preset rules.

[0046] The preset rules are used to screen and identify hot keywords according to certain criteria to obtain the keywords with high discussion and popularity in the current media content, so as to timely supplement relevant hot materials. The preset rules can cover various factors and can be specifically set according to the actual situation. For example: the frequency of the keyword appearing in the target media file, the growth trend within a specific time period, etc. For example, if a keyword appears multiple times in different target media files in a short period of time and the relevant discussions are increasing, then this keyword can be determined as a hot keyword.

[0047] After determining the hot keywords, it is also necessary to determine the popularity information of the hot keywords. According to the popularity information of the hot keywords, the quantity of relevant materials required for the hot keywords or the order of supplementing relevant materials can be determined. The popularity information can reflect the attention, popularity of the hot keywords and the heat of relevant discussions, and can be quantified through various data sources. For example, the popularity information can include the search volume of the hot keywords, the discussion volume on social platforms, the interaction of relevant users (such as comments, likes, shares, etc.), the number of occurrences of the hot keywords or the growth rate of the number of occurrences at different granularities such as days, weeks, months, years, etc.

[0048] Step S106 , obtain the materials to be supplemented based on the hot keywords, the popularity information of the hot keywords, the target media file and the existing materials in the material library.

[0049] By comprehensively analyzing hot keywords and their popularity information, the content of target media files, and the existing materials in the material library, the missing materials in the material library can be identified and the replenishment process can be initiated. Specifically, the quantity of materials required for a hot keyword can be determined based on the hot keyword and the corresponding popularity information. By checking the quantity of existing materials corresponding to the hot keyword in the material library, it can be determined whether it is necessary to supplement the materials related to the hot keyword into the material library. Further, if the quantity of materials related to the hot keyword in the material library does not meet the quantity of materials required by the hot keyword, it can be determined that there are missing materials related to the hot keyword in the material library, thereby triggering the material replenishment process. Relevant slice content can be obtained from the target media file using the hot keyword as the to-be-supplemented material, thus realizing the timely replenishment of hot materials. In some embodiments, a timed script can be set to check whether the material library needs to supplement materials related to hot content. Specifically, it can be set to trigger a retrieval at 3 am to retrieve the quantity of materials corresponding to all hot keywords in the material library and whether each hot keyword reaches the preset material supply volume. If the quantity of materials for a hot keyword (such as "artificial intelligence") is insufficient, a list of replenishment requirements for missing hot materials can be output, thereby initiating the material replenishment process.

[0050] Step S108 , supplement the to-be-supplemented material to the material library.

[0051] The to-be-supplemented material can be obtained from target media files with secondary creation authorization. Specifically, video understanding technology can be used to analyze the target media file, determine the region corresponding to the hot keyword, and then perform intelligent slicing on this region to extract relevant pictures and generate new materials that meet the requirements. For target media files without secondary creation authorization, a secondary creation solicitation activity can be launched on the video platform to encourage creators to submit original content that conforms to the hot keyword and provide corresponding incentive mechanisms (such as revenue sharing, exposure improvement, etc.). In some embodiments, suitable suppliers (such as news agencies, film and television copyright libraries, etc.) can be screened to purchase materials that conform to the hot keyword from the outside. Materials such as videos, pictures, or audio corresponding to the hot keyword can also be generated by calling an AIGC (AI-Generated Content) model. Through multiple material replenishment methods, the content coverage of the material library can be enriched, thus meeting the needs of content creators for various types of hot materials.

[0052] In an alternative embodiment, the target media file includes audio content, such as Figure 2 shown, step S102 includes: Step S200, convert the audio content into text content.

[0053] Step S202: Input the text content into a pre-trained text analysis model to obtain each of the keywords and their corresponding current word frequency data.

[0054] Through technologies such as Automatic Speech Recognition (ASR), the audio content in the target media file can be converted into text data, which serves as the data basis for extracting hot keywords. The ASR technology for obtaining the text content of the target media file is applicable to target media files containing audio content such as voice narration, conversations, interviews, speeches, etc., such as video manuscripts, podcasts, news reports, etc. Before performing ASR processing, noise reduction, echo cancellation, volume normalization, etc. can be performed on the audio content to improve the recognition accuracy of ASR. For example, for an interview video containing background noise, a neural network noise reduction algorithm can be applied to enhance speech clarity.

[0055] A text analysis model can refer to a machine learning or deep learning model that has been trained on a large-scale corpus and can perform natural language processing (NLP) tasks. By deeply analyzing the text content through the text analysis model, keywords can be extracted, and the number of occurrences of each keyword in the text content (i.e., word frequency) can be counted. The word frequency data can be used to measure the popularity of keywords, thereby identifying the keywords with relatively high current attention.

[0056] In this embodiment, by converting the audio content of the target media file into text content and inputting it into the text analysis model to extract keywords and their word frequency data, the core information in the target media file can be automatically and accurately identified, and the current hot content can be determined through the word frequency data, improving the intelligence level of hot content monitoring and the efficiency of obtaining hot content.

[0057] In an alternative embodiment, as Figure 3 shown, step S104 includes: Step S300: Obtain the historical word frequency data of each of the keywords.

[0058] The historical word frequency data can be obtained by querying a word frequency database. It can be queried based on a preset time granularity, such as weekly, monthly, or yearly granularity, to count the historical word frequency data of keywords in the past week, past month, or past year. According to the historical word frequency data, the specific change in the popularity of keywords can be judged more carefully.

[0059] Step S302: Based on the current word frequency data and historical word frequency data of each of the keywords, determine the long-term popularity and short-term popularity of the keywords.

[0060] Long-term popularity can reflect the overall attention level of keywords over a relatively long time range (such as several months or a year), while short-term popularity can reflect the popularity of keywords in the most recent time period (such as the most recent week or day). By comparing the long-term and short-term popularity, calculating the short-term popularity growth rate of keywords (for example, comparing the popularity of the most recent day with the average popularity of the past week) and the long-term popularity stability (for example, the fluctuation of the average popularity in the most recent three months), to distinguish "long-term hot events with steady growth" from "short-term burst hot events", so as to more reasonably plan the supplementary strategy of the materials required for keywords.

[0061] Exemplarily, both the long-term popularity and the short-term popularity can be calculated by the following formula:

[0062] represents the popularity value, represents the word frequency data at time (unit: day), represents the preset decay rate value, represents the total time span. Among them, can change with the total time span. For example, when the total time span is 1 year, when the total time span is 30 days, .

[0063] Step S304, determine the hot keywords and the popularity information of the hot keywords according to the long-term popularity and short-term popularity of each of the keywords.

[0064] Based on the data of the long-term popularity and short-term popularity of keywords, the determination conditions of hot keywords can be set according to the actual situation. For example, when a keyword simultaneously satisfies that the long-term popularity is higher than the average long-term popularity value of all keywords, and the short-term popularity growth rate exceeds a preset threshold (such as a growth of more than 200% in the past 3 days), then this keyword can be determined as a hot keyword. The popularity information of hot keywords can include information such as the name of the hot keyword, the category to which the hot keyword belongs, the long-term popularity value, the short-term popularity value, the word frequency data or popularity growth rate of the hot keyword at the day granularity, week granularity, month granularity, and year granularity.

[0065] In this embodiment, by obtaining historical word frequency data, current word frequency data and calculating long-term and short-term popularity, the current popularity and long-term trend of each keyword can be evaluated, so as to accurately identify hot keywords. According to the hot keywords and their popularity information, the materials corresponding to the current hot content can be accurately obtained, further improving the quality and efficiency of material supplementation and the usage experience of content creators.

[0066] In an alternative embodiment, each of the existing materials corresponds to a material keyword. For example, Figure 4 as shown, step S106 includes: Step S400, based on the popularity information of the hot keyword, determine a first value of the number of materials required for the hot keyword.

[0067] Step S402, based on the hot keyword, the existing materials, and the material keywords of the existing materials, determine the existing materials that match the hot keyword, where the number of the matching existing materials is denoted as a second value.

[0068] Step S404, in the case where the second value is less than the first value, determine the hot keyword as the target hot keyword.

[0069] Step S406, based on the target hot keyword, obtain the to-be-supplemented materials from the target media files.

[0070] In some embodiments, the material keywords corresponding to the existing materials can be obtained by understanding the content of each existing material in the material library. Specifically, technologies such as computer vision and natural language processing can be used to understand information such as the theme, scene, and characters in the materials. For video materials, the images, audio, and subtitle content in the video can be analyzed. For picture materials, the visual elements of the pictures can be analyzed. Further, the material keywords are extracted from the existing materials, and the categories and attributes of the material keywords are marked. The material keywords can include information such as names of people, place names, and events. For example, for a material library containing news hot material, the material keywords can include "automatic driving", "driverless vehicle", etc. After obtaining the material keywords corresponding to the existing materials, a correspondence table between the existing materials and the material keywords can be constructed to determine the existing materials corresponding to the hot keyword and the corresponding second value through the correspondence table.

[0071] By comparing the second value with a first threshold of the number of materials required for the hot keyword, it can be determined whether there is a lack of relevant hot materials for the hot keyword in the material library. For example, if there are 100 existing materials with the material keyword of "iron flower hitting" in the material library, but according to the popularity information of "iron flower hitting", it can be judged that at least 200 materials are required for this hot keyword, then "iron flower hitting" can be determined as the target hot keyword. Based on this target hot keyword, at least 100 slices related to "iron flower hitting" can be obtained from multiple target media files as the to-be-supplemented materials.

[0072] In this embodiment, by comparing the quantity of materials required for the hot keywords with the existing materials, the missing materials can be accurately identified, thereby ensuring that the quantity of materials in the material library can match the quantity of materials required for the current hot keywords. By promptly identifying and supplementing the missing materials, it is ensured that the high-heat keywords have sufficient material support, improving the timeliness and richness of the material content.

[0073] In an alternative embodiment, the target media file includes target video content, such as Figure 5 As shown, step S106 includes: Step S500, obtaining the target semantic vector of the target hot keyword.

[0074] In some embodiments, the target hot keyword can be encoded through word embedding techniques (such as Word2Vec, GloVe, BERT, etc.) to obtain a numerical representation of the target hot keyword.

[0075] Step S502, obtaining image feature vectors corresponding to multiple frames of the target video content; wherein, one frame corresponds to one timestamp.

[0076] The target video consists of consecutive frames, and each frame represents the image information at a specific time point in the video. In this step, the video is decomposed into multiple frames, and each frame corresponds to a specific timestamp, which can ensure that the images at each time point can be independently represented. Further, the visual content of each video frame can be described by calculating the image feature vector. The image feature vector can be obtained by processing the video frame through a deep learning model (such as a convolutional neural network). The image feature vector can reflect the key features in the image, such as information about color, shape, items, people, or scenes.

[0077] In some embodiments, the extraction range of the image feature vector can be determined according to the type of the target hot keyword. Especially when dealing with large-scale video data, the computational amount can be effectively reduced and the efficiency can be improved. Exemplarily, when the target hot keyword is a specific item, at this time, the image feature extraction only needs to focus on the image area related to the item. For example, when the target hot keyword is "cat", through object detection techniques (such as YOLO, Faster R-CNN, etc.), the area about the cat in the video frame can be located and only the image feature vector of this area is extracted, thus avoiding extracting all the image features of the entire frame and greatly saving computing power.

[0078] Step S504, determining the target frame and the corresponding target timestamp from the multiple frames according to the target semantic vector and the multiple image feature vectors.

[0079] In some embodiments, the target semantic vector of the target hot keyword can be compared with the semantic vectors of other modality data (such as images and audio) through a multimodal large language model, so as to achieve image-text matching or cross-modal retrieval. For example, the semantic vectors of an image and its description are used to determine the correlation between them. Specifically, in all frames, the frame with the highest matching degree with the target keyword can be identified as the target frame by calculating the similarity between the target semantic vector and the image feature vector.

[0080] Step S506: Based on the target timestamp, determine a target video slice from the target video content, and the target video slice serves as the material to be supplemented.

[0081] The target timestamp can point to the segment in the target video content with the highest matching degree with the target keyword, and the target video slice obtained according to the target timestamp can reflect the current hot content. Corresponding target video slices can be automatically extracted based on the target timestamp through various automatic slicing tools or pre-trained automatic slicing models to improve the efficiency of obtaining the material to be supplemented.

[0082] In this embodiment, by combining the target semantic vector of the target hot keyword with the image feature vector of the target video content, the target frame related to the target hot keyword can be more accurately determined from multiple frames, and the corresponding target timestamp can be extracted, so as to efficiently locate the segment in the target video content related to the hot keyword. Based on the target timestamp, the target video slice can be further extracted as the material to be supplemented, ensuring the relevance and quality of the material to be supplemented with the target hot keyword, and improving the efficiency and accuracy of obtaining the material to be supplemented.

[0083] In an alternative embodiment, as Figure 6 shown, step S108 includes: Step S600: Based on the popularity information of the hot keyword, determine the supplement order list of the material to be supplemented; wherein, the supplement order list includes multiple hot keywords and the materials to be supplemented corresponding to each hot keyword.

[0084] Step S602: According to the supplement order list, supplement multiple materials to be supplemented into the material library in sequence.

[0085] The supplementary order list of materials to be supplemented is sorted according to the popularity information of hot keywords, which can ensure that materials with higher comprehensive popularity are supplemented first. For example, hot keywords with higher popularity, such as "popular movies" or "major news events", will be arranged at the forefront of the supplementary order list, while hot keywords with relatively lower popularity will be postponed for supplementation. In addition, a dynamic adjustment mechanism can be set for the supplementary order list. When the popularity of a certain hot keyword rises rapidly, the corresponding materials will also be supplemented in advance.

[0086] In this embodiment, by analyzing and sorting the popularity information of hot keywords, the supplementary order list of elements to be supplemented can be used to supplement hot materials more accurately and quickly, thus better meeting the needs of content creators for real-time hot materials.

[0087] In an alternative embodiment, the method further includes: Checking the materials to be supplemented and determining whether the materials to be supplemented are added to the material library according to the check results.

[0088] In some embodiments, the checks performed on the materials to be supplemented may include clarity checks, relevance checks, copyright information checks, content compliance checks, and watermark checks, etc. Exemplarily, through an image processing model or a deep learning model (such as a convolutional neural network CNN), the resolution, clarity, and noise of the materials to be supplemented can be detected. If the resolution of the materials to be supplemented is lower than the preset standard or contains too much noise, the materials will be determined to not meet the quality requirements and thus will not be added to the material library. The relevance check can be performed by comparing the materials to be supplemented with the hot keywords through text matching or natural language processing (NLP) models to evaluate the relevance between the materials to be supplemented and the hot keywords. The copyright information check can be performed by a copyright detection model or a database related to copyright information to conduct a copyright review of the materials to be supplemented to ensure that the materials have been authorized or are public domain resources. Additionally, by combining the immutable, transparent, and decentralized characteristics of blockchain technology, the source, authorization status, and copyright information of the materials to be supplemented can be recorded and verified to ensure the legality and clear copyright ownership of the materials to be supplemented. The content compliance check can use natural language processing and image recognition algorithms to identify sensitive information in the materials, such as inappropriate remarks, images, or video content, to ensure that the materials to be supplemented do not contain content that violates platform regulations or laws and regulations. The watermark check can use an image recognition model to detect whether there are watermarks or unwanted marks in the materials, such as watermarks, copyright logos, etc. In addition, through image processing techniques (such as image denoising, edge detection), non-compliant watermarks can be automatically identified and removed.

[0089] In some embodiments, after supplementing the materials to be supplemented into the material library, deduplication processing can be performed on the material library. Specifically, a hash algorithm or a feature extraction method can be used to identify and remove duplicate materials in the material library, ensuring the efficient management of the material library, avoiding duplicate storage, and improving resource utilization.

[0090] In this embodiment, by automatically checking the quality of the materials to be supplemented and then deciding whether to add the materials to be supplemented according to the inspection results, the situation of low-quality materials or infringing materials in the material library is reduced, thereby improving the quality of the materials in the material library. It does not rely on manual inspection, further improving the intelligence level of supplementing the material library.

[0091] To make the present application easier to understand, the following is combined with Figure 7 Provide an exemplary application.

[0092] S11, Obtain high-quality manuscripts.

[0093] S12, Input the video manuscript into the speech-to-text model to obtain ASR data.

[0094] S13, According to the ASR data, extract the entity words therein, and analyze the frequency and importance of the entity words, thereby outputting a list of entity words and the corresponding popularity information.

[0095] S14, Monitor hotspots. Specifically, the long-term popularity and short-term popularity of entity words can be calculated through the daily statistical list of entity words and the corresponding popularity information, thereby identifying the current hot entity words.

[0096] S15, Understand the existing material content in the material library. Specifically, content understanding can be performed on materials such as videos and pictures in the material library, entity words in the materials are extracted and their categories and attributes are marked, thereby constructing the corresponding relationship between the materials and the entity words.

[0097] S16, Compare the number of materials corresponding to the entity words in the material library that are the same as the current hot entity words with the required amount of materials for the current hot entity words to determine the supplementary requirement list for the missing hot materials.

[0098] S17, Material supplementation. Specifically, according to the entity words corresponding to the missing hot materials, the corresponding video areas can be sliced from the second-creation authorized manuscripts to form materials.

[0099] S18, Review the copyright of the materials and screen out the materials that are the same as the existing materials in the material library.

[0100] S19, Add the materials that pass the review to the material library for users to use.

[0101] In this exemplary application, by obtaining the keywords of the target media file and determining the hot keywords and their popularity information from the keywords according to preset rules, the real-time hot content in the target media file is identified. According to the real-time hot content, the supplementary materials to be supplemented are automatically obtained and added to the material library, realizing the real-time update of the material library, thereby improving the timeliness of the material library. Through the highly time-sensitive material library, content creators can obtain materials related to real-time hot content in a timely manner, thus improving the usage experience of content creators.

[0102] Embodiment 2 Figure 8 The block diagram of the material library supplement device according to Embodiment 2 of the present application is schematically shown. The device can be divided into one or more program modules. One or more program modules are stored in a storage medium and executed by one or more processors to complete the embodiments of the present application. The program modules referred to in the embodiments of the present application refer to a series of computer program instruction segments that can complete specific functions. The following description will specifically introduce the functions of each program module in this embodiment. As Figure 8 shown, the device 1600 may include: an acquisition module 1610, a determination module 1620, and a supplement module 1630, where: The acquisition module 1610 is configured to acquire a target media file; based on the target media file, acquire one or more keywords; The determination module 1620 is configured to determine hot keywords and the popularity information of the hot keywords from one or more of the keywords according to preset rules; The acquisition module 1610 is further configured to acquire supplementary materials to be supplemented based on the hot keywords, the popularity information of the hot keywords, the target media file, and the existing materials in the material library; The supplement module 1630 is configured to supplement the supplementary materials to be supplemented into the material library.

[0103] In an optional embodiment, the target media file includes audio content; Based on the target media file, acquiring one or more keywords includes: Converting the audio content into text content; Inputting the text content into a pre-trained text analysis model to obtain each of the keywords and the corresponding current word frequency data.

[0104] In an optional embodiment, determining hot keywords and the popularity information of the hot keywords from one or more of the keywords according to preset rules includes: Obtaining the historical word frequency data of each of the keywords; Determine the long-term popularity and short-term popularity of each keyword based on the current word frequency data and historical word frequency data of each keyword; Determine the hot keywords and the popularity information of the hot keywords according to the long-term popularity and short-term popularity of each keyword.

[0105] In an alternative embodiment, each of the existing materials corresponds to a material keyword; Obtain the materials to be supplemented based on the hot keywords, the popularity information of the hot keywords, the target media file, and the existing materials in the material library, including: Determine a first value of the number of materials required for the hot keyword based on the popularity information of the hot keyword; Determine the existing materials that match the hot keyword based on the hot keyword, the existing materials, and the material keywords of the existing materials, where the number of the matching existing materials is denoted as a second value; In the case where the second value is less than the first value, determine the hot keyword as the target hot keyword; and Obtain the materials to be supplemented from the target media file based on the target hot keyword.

[0106] In an alternative embodiment, the target media file includes target video content; obtaining the materials to be supplemented from the target media file based on the target hot keyword includes: Obtain the target semantic vector of the target hot keyword; Obtain the image feature vectors corresponding to multiple frames of the target video content; where one frame corresponds to one timestamp; Determine the target frame and the corresponding target timestamp from multiple frames according to the target semantic vector and multiple image feature vectors; and Determine a target video slice from the target video content based on the target timestamp, and the target video slice serves as the materials to be supplemented.

[0107] In an alternative embodiment, supplementing the materials to be supplemented into the material library includes: Determine the supplement order list of the materials to be supplemented based on the popularity information of the hot keyword; where the supplement order list includes multiple hot keywords and the materials to be supplemented corresponding to each hot keyword; Supplement multiple materials to be supplemented into the material library in sequence according to the supplement order list.

[0108] In an alternative embodiment, the apparatus 1600 further includes an inspection module, and the inspection module is used for: Check the material to be supplemented, and determine whether the material to be supplemented is added to the material library according to the inspection result.

[0109] Embodiment III Figure 9 Schematically shows a hardware architecture diagram of a computer device 10000 suitable for implementing the material library supplementation method according to Embodiment III of the present application. In some embodiments, the computer device 10000 may be a terminal device such as a smart phone, a wearable device, a tablet computer, a personal computer, a vehicle-mounted terminal, a game console, a virtual device, a workbench, a digital assistant, a set-top box, a robot, etc. In other embodiments, the computer device 10000 may be a rack-mounted server, a blade server, a tower server, or a cabinet server (including an independent server or a server cluster composed of multiple servers). As Figure 9 shown, the computer device 10000 includes, but is not limited to: a memory 10010, a processor 10020, and a network interface 10030 that can be communicatively linked to each other through a system bus. Among them: The memory 10010 includes at least one type of computer-readable storage medium. The readable storage medium includes flash memory, a hard disk, a multimedia card, a card-type memory (such as an SD or DX memory), a random access memory (RAM), a static random access memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, an optical disk, etc. In some embodiments, the memory 10010 may be an internal storage module of the computer device 10000, such as the hard disk or memory of the computer device 10000. In other embodiments, the memory 10010 may also be an external storage device of the computer device 10000, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the computer device 10000. Of course, the memory 10010 may also include both the internal storage module and the external storage device of the computer device 10000. In this embodiment, the memory 10010 is generally used to store the operating system and various application software installed on the computer device 10000, such as the program code of the material library supplementation method. In addition, the memory 10010 may also be used to temporarily store various data that have been output or will be output.

[0110] The processor 10020 can be a Central Processing Unit (CPU), a controller, a microcontroller, a microprocessor, or other chips in some embodiments. The processor 10020 is generally used to control the overall operation of the computer device 10000, such as performing control and processing related to data interaction or communication with the computer device 10000. In this embodiment, the processor 10020 is used to run the program code stored in the memory 10010 or process data.

[0111] The network interface 10030 may include a wireless network interface or a wired network interface. The network interface 10030 is generally used to establish a communication link between the computer device 10000 and other computer devices. For example, the network interface 10030 is used to connect the computer device 10000 to an external terminal through a network, and establish a data transmission channel and a communication link between the computer device 10000 and the external terminal. The network can be an enterprise intranet (Intranet), the Internet, the Global System of Mobile communication (abbreviated as GSM), Wideband Code Division Multiple Access (abbreviated as WCDMA), 4G network, 5G network, Bluetooth, Wi-Fi and other wireless or wired networks.

[0112] It should be noted that Figure 9 Only the computer device with components 10010 - 10030 is shown, but it should be understood that it is not required to implement all the shown components, and more or fewer components can be implemented alternatively.

[0113] In this embodiment, the material library supplement method stored in the memory 10010 can also be divided into one or more program modules and executed by one or more processors (such as the processor 10020) to complete the embodiments of the present application.

[0114] Embodiment 4 The embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the material library supplement method in the embodiment are implemented.

[0115] In this embodiment, the computer-readable storage medium includes flash memory, hard disks, multimedia cards, card-type memories (e.g., SD or DX memories, etc.), random access memories (RAM), static random access memories (SRAM), read-only memories (ROM), electrically erasable programmable read-only memories (EEPROM), programmable read-only memories (PROM), magnetic memories, magnetic disks, optical disks, etc. In some embodiments, the computer-readable storage medium may be an internal storage unit of a computer device, such as the hard disk or memory of the computer device. In other embodiments, the computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc., equipped on the computer device. Of course, the computer-readable storage medium may also include both the internal storage unit and the external storage device of the computer device. In this embodiment, the computer-readable storage medium is generally used to store the operating system installed on the computer device and various application software, such as the program code of the material library supplement method in the embodiment. In addition, the computer-readable storage medium can also be used to temporarily store various data that have been output or will be output.

[0116] Embodiment 5 The embodiment of the present application also provides a computer program product, including a computer program, which implements the method in the above embodiment when executed by a processor.

[0117] Obviously, those skilled in the art should understand that the above-mentioned modules or steps of the embodiments of the present application can be implemented by a general computer device. They can be concentrated on a single computer device or distributed on a network composed of multiple computer devices. Optionally, they can be implemented by program codes executable by the computer device, so that they can be stored in a storage device and executed by the computer device. And in some cases, the steps shown or described can be executed in a different order from here, or they can be separately fabricated into individual integrated circuit modules, or multiple modules or steps among them can be fabricated into a single integrated circuit module to be implemented. In this way, the embodiments of the present application are not limited to any specific combination of hardware and software.

[0118] It should be noted that the above are only the preferred embodiments of the present application, and do not limit the patent protection scope of the present application. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of the present application, or directly or indirectly applied in other related technical fields, shall be equally included in the patent protection scope of the present application.

Claims

1. A material library supplement method, characterized in that: The method comprises: Get the target media file; Based on the target media file, obtaining one or more keywords; According to a preset rule, determining a hot keyword and the popularity information of the hot keyword from one or more of the keywords; Based on the hot keywords, the popularity information of the hot keywords, the target media files and the existing materials in the material library, obtaining the materials to be supplemented; and The material to be supplemented is supplemented into the material library.

2. The method according to claim 1, characterized in that The target media file includes audio content; Based on the target media file, one or more keywords are obtained, including: Converting the audio content into text content; The text content is input into a pre-trained text analysis model to obtain each of the keywords and the corresponding current word frequency data.

3. The method according to claim 2, characterized in that Determining hot keywords and popularity information of the hot keywords from one or more of the keywords according to a preset rule includes: Obtaining historical frequency data of each of the keywords; Determine the long-term popularity and short-term popularity of the keyword based on the current word frequency data and the historical word frequency data of each of the keywords; The hot keywords and the popularity information of the hot keywords are determined according to the long-term popularity and short-term popularity of each of the keywords.

4. The method according to claim 1, characterized in that: Each of the existing materials corresponds to a material keyword; Based on the hot keywords, the popularity information of the hot keywords, the target media files and the existing materials in the material library, the materials to be supplemented are obtained, including: Based on the popularity information of the hot keyword, determine a first value of the amount of materials required for the hot keyword; Based on the hot keyword, the existing material and the material keyword of the existing material, determining the existing material matching the hot keyword, wherein the number of the matching existing materials is recorded as a second value; When the second value is less than the first value, determining the hot keyword as a target hot keyword; and Based on the target hot keyword, the material to be supplemented is obtained from the target media file.

5. The method according to claim 4, characterized in that The target media file includes target video content; Based on the target hot keyword, obtaining the material to be supplemented from the target media file includes: Obtaining a target semantic vector of the target hot keyword; Obtaining image feature vectors corresponding to a plurality of frames of the target video content; wherein one frame corresponds to one timestamp; Determining a target frame and a corresponding target timestamp from the multiple frames according to the target semantic vector and the multiple image feature vectors; and Based on the target timestamp, a target video slice is determined from the target video content, and the target video slice serves as the material to be supplemented.

6. The method according to any one of claims 1 to 5, characterized in that: Supplementing the to-be-supplemented material into the material library includes: Based on the popularity information of the hot keywords, determining a supplementary sequence table of the materials to be supplemented; wherein the supplementary sequence table includes a plurality of the hot keywords and the materials to be supplemented corresponding to each of the hot keywords; According to the supplement sequence table, the plurality of materials to be supplemented are supplemented into the material library in sequence.

7. The method according to any one of claims 1 to 5, characterized in that: The method further comprises: The material to be supplemented is checked, and it is determined whether the material to be supplemented is added to the material library according to the checking result.

8. A material library supplement device, characterized in that: The device comprises: An acquisition module, used to acquire a target media file; based on the target media file, acquire one or more keywords; A determination module, used to determine hot keywords and popularity information of the hot keywords from one or more of the keywords according to a preset rule; The acquisition module is further used to acquire the material to be supplemented based on the hot keywords, the popularity information of the hot keywords, the target media files and the existing materials in the material library; A supplementing module is used to supplement the material to be supplemented into the material library.

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

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

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

Images