Short video element analysis method, system and terminal

Abstract

The invention relates to a short video element analysis method, system and terminal, and relates to the technical field of short video analysis, and the method comprises the steps: obtaining sound detection information; obtaining an initial text according to the sound detection information; when music appears in the sound detection information, background melody is determined according to the sound detection information; matching background music information based on the background melody; determining music characters according to the background music information; determining a new initial text according to the background melody and the music characters; determining a combined vocabulary based on the new initial text; determining whether the combined vocabulary is consistent with the reference vocabulary; if yes, pushing the short video to a short video platform; if not, taking the short video as a marked short video; acquiring browsing related information; and determining marked content according to the browsing related information and the marked short video, pushing the marked short video to a short video platform, and adding the combined vocabularies to a pushing database. The short video pushing method and device have the effect of improving the short video pushing efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of short video analysis, and in particular to a short video element analysis method, system and terminal. Background Art

[0002] Short video analysis is a process of processing video data and extracting the information required by the operator from the video.

[0003] In people's daily lives, short videos have become an indispensable part. Short videos can inspire users' creativity and imagination, enabling users to create corresponding short videos and publish them to short video platforms. When short videos shot by users' creativity and imagination are liked by the public, short videos are manually analyzed to extract pictures, texts, music and other content presented in the short videos, and the extracted content is used as the elements of the hit. When other users watch the hit short videos, they will also create and publish videos around the hit elements, and continue to manually analyze the elements of each short video and classify them to facilitate the subsequent push of videos with hit elements.

[0004] Since short video analysis is generally performed manually, when there are too many elements in a short video, manual analysis of the elements becomes time-consuming and laborious, thereby reducing the efficiency of short video push. Summary of the invention

[0005] In order to improve the efficiency of short video push, the present invention provides a short video element analysis method, system and terminal.

[0006] In a first aspect, the present invention provides a short video element analysis method, which adopts the following technical solution: A short video element analysis method, comprising: Get the sound detection information of the short video; Inputting the sound detection information into a preset audio analysis system to obtain an initial text; When music appears in the sound detection information, determine the background music according to the sound detection information; Matching background music information from a preset music library based on background music; Determine the music text according to the background music information; Eliminate the corresponding words in the initial text according to the background rhythm and the music words, and use the initial text after elimination as the new initial text; Splitting based on the characters in the new initial text, and combining the split characters adjacently to determine combined words; Determine whether the combined vocabulary is consistent with the preset benchmark vocabulary; If the combined vocabulary is consistent with the preset benchmark vocabulary, the short video corresponding to the combined vocabulary consistent with the preset benchmark vocabulary will be pushed to the short video platform first; If the combined vocabulary is inconsistent with the preset reference vocabulary, the short video corresponding to the combined vocabulary inconsistent with the preset reference vocabulary is used as the marked short video; Obtain browsing related information of the marked short video; The marked content is determined based on browsing related information and marking short videos, and the marked short videos are pushed to the short video platform corresponding to the marked content and the marked content is added to a preset push database.

[0007] By adopting the above technical scheme, the sound of the short video is analyzed to obtain the initial text, and the new initial text is obtained after the music text is eliminated. The vocabulary in the new initial text is understood to obtain the combined vocabulary and compared with the benchmark vocabulary, and then the corresponding short videos are pushed with priority based on the comparison results, so that the short videos related to the benchmark vocabulary can be pushed directly to improve the efficiency of short video push, and the marked content is determined by the traffic situation of the marked short video, and the marked content of the marked short video is added to the push database, so that it is convenient to continue to push the short video related to the marked content in the future, thereby improving the efficiency of short video push.

[0008] Optionally, the method for determining the combined vocabulary includes: Selecting uncombined characters from the updated initial text according to the combined vocabulary; Determine uncombined syllables based on uncombined characters; Combining the adjacent uncombined syllables, and taking the combined syllables as the selected syllables; Selecting different selected words from a preset word database based on the selected syllable; Determine the domain of combined vocabulary based on combined vocabulary; Determine the selected vocabulary area based on the selected vocabulary; Determine whether the selected vocabulary domain is consistent with the combined vocabulary domain; When the selected vocabulary field is consistent with the combined vocabulary field, the selected vocabulary corresponding to the selected vocabulary field consistent with the combined vocabulary field is added to the combined vocabulary; When the selected vocabulary field is inconsistent with the combined vocabulary field, the characters corresponding to the selected vocabulary and the combined vocabulary adjacent to the selected vocabulary are input into the vocabulary database for recombination, and when the fields corresponding to the recombined vocabulary are consistent, the recombined vocabulary is added to the combined vocabulary.

[0009] By adopting the above technical scheme, the syllables corresponding to the uncombined characters are understood to obtain different selected words, and the combined words are increased by comparing the selected word fields corresponding to the selected words with the combined word fields, so that it is not easy for words with the same syllables to affect the element analysis of the short video, thereby improving the accuracy of short video push.

[0010] Optionally, the method of using the short video corresponding to the combined vocabulary inconsistent with the preset reference vocabulary as a tag short video includes: Determining a benchmark matching vocabulary based on the benchmark vocabulary input into a vocabulary database; Determine whether the combined vocabulary is consistent with the benchmark matching vocabulary; If the combined vocabulary is consistent with the benchmark matching vocabulary, the short video corresponding to the combined vocabulary consistent with the benchmark matching vocabulary will be pushed to the short video platform first; If the combined vocabulary is inconsistent with the benchmark matching vocabulary, the benchmark matching field is determined according to the benchmark matching vocabulary; When the combined vocabulary field is inconsistent with the reference matching field, the short video corresponding to the combined vocabulary field inconsistent with the reference matching field is used as the marked short video.

[0011] By adopting the above technical solution, the combined vocabulary is compared with the benchmark vocabulary, and the comparison results are used to obtain the situation of giving priority to short videos, so that short videos related to the benchmark vocabulary can be identified and given priority, thereby improving the efficiency of short video push.

[0012] Optionally, the method of using the short video corresponding to the combined vocabulary field inconsistent with the reference matching field as a tag short video includes: Obtain the display image information corresponding to the short video; Determine the feature type according to the displayed image information and the preset reference features; Determine different feature areas based on feature types; Determine whether the feature domain falls into the combined vocabulary domain; If the feature domain does not fall into the combined vocabulary domain, the short video corresponding to the combined vocabulary domain that is inconsistent with the benchmark matching domain continues to be used as the marked short video; If the feature domain falls into the combined vocabulary domain, the maximum number of overlaps of a single feature domain is determined based on the combined vocabulary domain; According to the feature domain, the occurrence frequency of different feature domains in the combined vocabulary domain is determined; When the number of occurrences is greater than the maximum number of overlaps, the short videos corresponding to the feature fields that fall into the combined vocabulary fields are pushed to the short video platform first; When the number of occurrences is less than the maximum number of overlaps, the short video corresponding to the combined vocabulary field that is inconsistent with the reference matching field continues to be used as the marked short video.

[0013] By adopting the above technical solution, the image of the short video is understood to obtain the characteristic field, and the priority push of the short video is obtained by the falling of the characteristic field in the combined vocabulary field, so that the key points expressed by the short video can be classified, and the short video with the focus on the combined vocabulary field can be pushed preferentially.

[0014] Optionally, the method for determining the marked content includes: When the amount of relevant information browsed is greater than a preset benchmark, trigger information of the user on the short video platform is obtained; Determine the marking duration of the short video according to the trigger information; Determine the tag type based on the tagged short video and the tag duration, and the tag type includes text, image, music, etc.; When the tag type is music, the background music information is used as the tag content; When the mark type is a picture or text, the mark image information and the mark text information corresponding to the mark duration are determined according to the displayed image information and the updated initial text; When the number of occurrences is greater than the maximum number of overlaps, the feature area corresponding to the marked image information is used as the marked content; When the number of occurrences is less than the maximum number of overlaps, a combination of words in the tag text information is selected as a tag word according to the tag text information, and the tag word is used as the tag content.

[0015] By adopting the above technical solution, the user's trigger information is understood to obtain the tag type, and different tag contents are obtained through the tag type, so that the tag content can be known more accurately to improve the efficiency of subsequent short video push.

[0016] Optional verification methods for browsing related information include: Obtain the user's control scope on the short video platform and the video update type of the short video platform; Determine the estimated control position based on the control range; According to the video update type to call the control direction; Determine the user's holding state based on the estimated control position and control direction; Matching a reference error distance from a preset error database based on the grip state; Calculate the error distance based on the estimated control position and the preset trigger position; When the error distance is greater than the reference error distance, the total video duration of the marked short video and the viewing duration of the marked short video by the user are obtained; Determine baseline watch time based on total video length; When the viewing time is not greater than the benchmark viewing time, the browsing related information is updated according to the viewing time.

[0017] By adopting the above technical solution, the user's holding state is understood to determine whether the user inadvertently outputs trigger information, and further analysis is performed through comparison of the viewing time with the benchmark viewing time to update browsing related information, so that it is not easy for users to inadvertently like or swipe likes, causing the short video to become a hit, thereby improving the accuracy of pushing hit short videos.

[0018] Optionally, the verification method for browsing related information also includes: Obtaining a short video of the user in a preset video storage repository and using it as a selected short video; Determine the selected content based on the sound detection information and the displayed image information corresponding to the selected short video; Retrieving a selected video area based on the selected content; When the user outputs the trigger information, the marked short video corresponding to the received trigger information is defined as the trigger short video, and the trigger short video is recorded into the preset video storage library, and the trigger content is determined according to the sound detection information and the display image information corresponding to the trigger short video; Retrieving a trigger video area based on the trigger content; determining whether the trigger video field is consistent with the selected video field; When the trigger video domain is inconsistent with the selected video domain, the browse related information is updated according to the selected video domain.

[0019] By adopting the above technical solution, by comparing the selected short video with the field corresponding to the triggering short video, the browsing related information is updated according to the comparison result, so that it is not easy for users to randomly like or accidentally like and generate false traffic.

[0020] Optionally, the verification method for browsing related information also includes: When the selected video area is consistent with the triggering video area, obtaining the comment area corresponding to the triggering short video in the short video platform; Determine whether the comment area receives trigger information; When no trigger information is received in the comment area, the relevant information is updated and browsed according to the selected video area; When receiving trigger information in the comment area, obtain the user's comment information; When text appears in the comment information, the comment text is determined based on the comment information input into a preset text recognition component, and the remaining comment text in the comment area is used as the remaining comment text; According to the remaining comment text and the comment text input into the vocabulary database to determine the degree of vocabulary overlap; When the vocabulary overlap is less than a preset benchmark overlap, the comment area is determined based on the comment text; When the comment field and the trigger video field are inconsistent, the browsing related information is updated according to the comment field and the trigger video field.

[0021] By adopting the above technical solution, the comment information of the user is understood to obtain the comment text and the remaining comment text, and then the comment text and the remaining comment text are used to obtain the degree of vocabulary overlap and compared with the benchmark overlap degree. The browsing related information is updated according to the comparison result, so that it is not easy for users to make random comments and generate false traffic.

[0022] In the second aspect, the present application provides a short video element analysis system, which adopts the following technical solution: A short video element analysis system, comprising: Acquisition module, used for sound detection information, browsing related information, displaying image information, trigger information, control range, video update type, total video duration, viewing duration, selected short video, comment area and comment information; A memory for storing a program of a short video element analysis method; The processor is used to load, execute and implement the program stored in the memory.

[0023] In a third aspect, the present application provides a smart terminal, which adopts the following technical solution: An intelligent terminal includes a memory and a processor, wherein the memory stores a computer program that can be loaded by the processor and execute a short video element analysis method.

[0024] In summary, the present application includes at least one of the following beneficial technical effects: 1. Analyze the sound of the short video to obtain the initial text, and then remove the music lyrics to obtain a new initial text. After understanding the vocabulary in the new initial text to obtain the combined vocabulary and compare it with the benchmark vocabulary, the corresponding short video is pushed first according to the comparison results, so that the short video related to the benchmark vocabulary can be pushed directly to improve the efficiency of short video push, and the marked content is determined by the traffic situation of the marked short video, and the marked content of the marked short video is added to the push database, so that it is convenient to continue to push the short video related to the marked content in the future, thereby improving the efficiency of short video push; 2. By understanding the user's holding state to determine whether the user accidentally outputs the trigger information, and further analyzing the comparison results of the viewing time with the benchmark viewing time to update the browsing related information, it is less likely that the user will accidentally like or like the short video, causing it to become a hit, thereby improving the accuracy of the push of hit short videos; 3. By understanding the user's comment information to obtain the comment text and the remaining comment text, and then comparing the comment text with the remaining comment text to obtain the degree of vocabulary overlap and compare it with the benchmark overlap degree, the browsing related information is updated according to the comparison results, so that it is not easy for users to make random comments and generate false traffic. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 is a flow chart of a method for analyzing short video elements according to an embodiment of the present invention; Figure 2 is a flow chart of a method after determining a combined vocabulary according to an embodiment of the present invention; Figure 3 is a flow chart of a method according to an embodiment of the present invention before using a short video corresponding to a combined vocabulary field inconsistent with a reference matching field as a marked short video; Figure 4 It is a method flow chart of a verification method for browsing related information according to an embodiment of the present invention. DETAILED DESCRIPTION

[0026] The present invention is further described in detail below in conjunction with the accompanying drawings and embodiments.

[0027] Reference Figure 1 The present application embodiment discloses a short video element analysis method, comprising the following steps: Step S100: Obtaining sound detection information of the short video.

[0028] The sound detection information refers to the sound information of each short video in the short video platform, and the sound parameters obtained by detecting each short video in the short video platform through a preset audio analysis system are used as the sound detection information. The audio analysis system refers to a system for analyzing the sound of a short video and retrieving the text and rhythm information expressed by the sound. In this embodiment, the audio analysis system includes a speech-to-text module and a music audio analysis module.

[0029] Step S101: inputting sound detection information into a preset audio analysis system to obtain an initial text.

[0030] The initial text refers to the text that is obtained by converting the sound in the short video and has not been processed. The text information converted by inputting the sound detection information into the audio analysis system is used as the initial text.

[0031] Step S102: when music appears in the sound detection information, determine the background music according to the sound detection information.

[0032] Background music refers to the musical melody of the background music corresponding to the short video. When music appears in the sound detection information, it means that there is text corresponding to the music or title in the conversion process of the initial text. Therefore, the corresponding music is retrieved from the sound detection information through the audio analysis system as the background music.

[0033] Step S103: matching background music information in a preset music library based on the background melody.

[0034] Background music information refers to the name and lyrics of the background music used in the short video, and the background music information is matched by inputting the background music into a preset music library. The music library stores the music, name, lyrics and other information corresponding to different music. The music library is pre-set by those skilled in the art and will not be described in detail here.

[0035] Step S104: Determine the music text according to the background music information.

[0036] The music text refers to the lyrics corresponding to the background music, and the corresponding lyrics are retrieved from the background music information as the music text.

[0037] Step S105: Eliminate the corresponding characters in the initial text according to the background music and the musical characters, and use the initial text after elimination as a new initial text.

[0038] The background music is used to match the musical text at each time point, and then the musical text at the corresponding time point in the initial text is removed, and the removed text is integrated to form a new initial text.

[0039] Step S106: splitting the characters in the new initial text, and combining the split characters adjacently to determine combined words.

[0040] A combined vocabulary refers to a vocabulary composed of adjacent and non-repeating characters in an initial text. The characters in the new initial text are split and the split characters are combined adjacently. Each combined character is input into a vocabulary database for corresponding matching, and the vocabulary corresponding to the matched character combination is used as a combined vocabulary.

[0041] Step S107: When no music appears in the sound detection information, determine the combined vocabulary according to the initial text.

[0042] When no music appears in the sound detection information, it means that there are no music lyrics in the conversion process of the initial text. Therefore, the text in the initial text is directly split, and the split text is combined adjacently. Each combined text is input into the vocabulary database for corresponding matching, and the vocabulary corresponding to the matched text combination is used as the combined vocabulary.

[0043] Step S108: Determine whether the combined vocabulary is consistent with the preset reference vocabulary. If yes, jump to step S109. If no, jump to step S110.

[0044] The benchmark vocabulary refers to the vocabulary of the short video whose page views exceed the minimum page views corresponding to the popular short video. The benchmark vocabulary is set in advance by those skilled in the art and will not be described in detail here. By judging the combined vocabulary and the benchmark vocabulary, it is determined whether the short video contains the content elements of the popular short video.

[0045] Step S109: The short videos corresponding to the combined vocabulary consistent with the preset benchmark vocabulary are preferentially pushed to the short video platform.

[0046] When the combined vocabulary is consistent with the benchmark vocabulary, it indicates that there are explosive content elements in the short video, so the short video corresponding to the combined vocabulary consistent with the benchmark vocabulary will be pushed to the short video platform first.

[0047] Step S110: taking the short video corresponding to the combined vocabulary that is inconsistent with the preset reference vocabulary as the marked short video.

[0048] A marked short video refers to a short video that does not contain the corresponding elements of the popular content. When the combined vocabulary is inconsistent with the base vocabulary, it means that there are no popular content elements in the short video. Therefore, the short video corresponding to the combined vocabulary that is inconsistent with the preset base vocabulary is used as a marked short video.

[0049] Step S111: Obtain browsing related information of the marked short video.

[0050] Browsing related information refers to the number of views, likes, favorites, etc. obtained by the marked short video on the short video platform. The browsing related information is obtained by retrieving the browsing, likes, favorites, etc. corresponding to the marked short video from the system corresponding to the short video platform and integrating them to obtain the browsing related information.

[0051] Step S112: Determine the marked content based on browsing related information and marking the short video, push the marked short video to the short video platform corresponding to the marked content, and add the marked content to a preset push database.

[0052] Tagged content refers to the elements of the corresponding hit in the tagged short video. The tagged content is obtained by browsing the relevant information and analyzing the tagged short video. The push database refers to the database that stores the hit content and is used to facilitate the direct push of short videos related to the hit content. The push database is initially set in advance by a person skilled in the art and is updated in real time by adding the tagged content, which will not be elaborated here. By pushing the tagged short video to the short video platform corresponding to the tagged content, the short video platform logged in by the user who is interested in the tagged content can receive the tagged short video first, and add the tagged content to the preset push database, so that the subsequent short video corresponding to the tagged content can be pushed conveniently, thereby improving the efficiency of short video push.

[0053] Reference Figure 2 , the methods after determining the combined vocabulary include: Step S200: selecting uncombined characters from the updated initial text according to the combined vocabulary.

[0054] Uncombined characters refer to adjacent characters that cannot be combined into words in the updated initial text. Combination words are removed from the updated initial text, and the remaining characters are regarded as uncombined characters.

[0055] Step S201: determining uncombined syllables according to uncombined characters.

[0056] Uncombined syllables refer to the pronunciation syllables corresponding to uncombined characters, and the pronunciation syllables corresponding to each character are retrieved from the uncombined characters as uncombined syllables.

[0057] Step S202: Combining adjacent uncombined syllables, and taking the combined syllables as selected syllables.

[0058] The selected syllable refers to the lexical syllable selected after combination, by combining adjacent uncombined syllables and taking the combined syllable as the selected syllable.

[0059] Step S203: Select different selected words from a preset vocabulary database based on the selected syllable.

[0060] The selected vocabulary refers to the vocabulary corresponding to the selected syllable, and the selected syllable is input into a preset vocabulary database to match different selected vocabulary. The vocabulary database stores vocabulary corresponding to different combinations of syllables. The vocabulary database is a database set by humans and will not be described in detail here.

[0061] Step S204: Determine the combined vocabulary domain according to the combined vocabulary.

[0062] The combined vocabulary field refers to the text field corresponding to the combined vocabulary. The combined vocabulary field is matched from the vocabulary database through the combined vocabulary. The vocabulary database stores text fields corresponding to different vocabulary, which will not be described in detail here.

[0063] Step S205: Determine the selected vocabulary domain according to the selected vocabulary.

[0064] The selected vocabulary field refers to the text field corresponding to the selected vocabulary. The selected vocabulary field is matched from the vocabulary database through the selected vocabulary. The vocabulary database stores text fields corresponding to different selected vocabulary, which will not be described in detail here.

[0065] Step S206: Determine whether the selected vocabulary domain is consistent with the combined vocabulary domain. If yes, jump to step S207. If no, jump to step S208.

[0066] By determining whether the selected vocabulary field is consistent with the combined vocabulary field, it is determined whether words with the same syllables are likely to affect the analysis of short videos.

[0067] Step S207: adding the selected vocabulary corresponding to the selected vocabulary field consistent with the combined vocabulary field to the combined vocabulary.

[0068] When the selected vocabulary field is consistent with the combined vocabulary field, it means that the words with the same syllables are not likely to affect the analysis of the short video, so the selected vocabulary corresponding to the selected vocabulary field that is consistent with the combined vocabulary field is added to the combined vocabulary.

[0069] Step S208: Inputting characters corresponding to the selected word and the combined words adjacent to the selected word into the word database for recombining, and when the fields corresponding to the recombined words are consistent, adding the recombined words to the combined words.

[0070] When the selected vocabulary field is inconsistent with the combined vocabulary field, it means that words with the same syllables can easily affect the combination of words and the analysis of short videos. Therefore, the text corresponding to the selected vocabulary and the combined vocabulary adjacent to the selected vocabulary is re-split, and the split text is re-matched with the vocabulary and the field corresponding to the vocabulary through the vocabulary database until the fields of each vocabulary are consistent. The recombined vocabulary is added to the combined vocabulary, so that it is not easy for words with the same syllables to affect the analysis of short video elements, so as to improve the accuracy of short video platforms in pushing short videos of popular content.

[0071] The method of using the short video corresponding to the combined vocabulary inconsistent with the preset reference vocabulary as a tag short video includes: Step S300: Determine a reference matching vocabulary based on the reference vocabulary input into the vocabulary database.

[0072] The reference matching vocabulary refers to the vocabulary used in conjunction with the reference vocabulary in the text, and the reference matching vocabulary is matched by inputting the reference vocabulary into the vocabulary database. The vocabulary database also stores the vocabulary used in conjunction with different vocabulary, which will not be described in detail here.

[0073] Step S301: Determine whether the combined vocabulary is consistent with the reference matching vocabulary. If yes, jump to step S302. If no, jump to step S303.

[0074] By determining whether the combined vocabulary is consistent with the benchmark matching vocabulary, it is possible to determine whether the short video corresponding to the combined vocabulary is related to the popular content.

[0075] Step S302: The short videos corresponding to the combined words that are consistent with the reference matching words are preferentially pushed to the short video platform.

[0076] When the combined vocabulary is consistent with the benchmark matching vocabulary, it means that the short video corresponding to the combined vocabulary is highly relevant to the popular content, so the short video corresponding to the combined vocabulary consistent with the benchmark matching vocabulary will be pushed to the short video platform first.

[0077] Step S303: Determine a reference matching domain according to the reference matching vocabulary.

[0078] The benchmark matching field refers to the text field corresponding to the benchmark matching vocabulary. When the combined vocabulary is inconsistent with the benchmark matching vocabulary, it means that the short video corresponding to the combined vocabulary has a low degree of relevance to the hot content. Therefore, the benchmark matching field is matched from the vocabulary database through the benchmark matching vocabulary. The vocabulary database stores text fields corresponding to different vocabularies, which will not be elaborated here.

[0079] Step S304: when the combined vocabulary field is inconsistent with the reference matching field, the short video corresponding to the combined vocabulary field inconsistent with the reference matching field is used as a marked short video.

[0080] When the domain of the combined vocabulary is consistent with the benchmark matching domain, it means that the short video corresponding to the combined vocabulary is related to the domain of the hit content, so the short video corresponding to the combined vocabulary domain will continue to be pushed first. When the domain of the combined vocabulary is inconsistent with the benchmark matching domain, it means that the short video corresponding to the combined vocabulary is not related to the domain of the hit content, so the short video corresponding to the combined vocabulary domain inconsistent with the benchmark matching domain is used as the marked short video, thereby improving the accuracy of the short video platform in pushing short videos of hit content.

[0081] Reference Figure 3 The method of marking the short video corresponding to the combined vocabulary field inconsistent with the reference matching field as the short video includes: Step S400: Obtain display image information corresponding to the short video.

[0082] Display image information refers to the image displayed by each short video, and the image corresponding to each short video is retrieved from the system of the short video platform as the display image information.

[0083] Step S401: determining a feature type according to displayed image information and preset reference features.

[0084] The reference features refer to features such as text, people, objects, and trees in the image. The reference features are set in advance by those skilled in the art and are not described in detail here. The feature type refers to the type of text, people, objects, and trees corresponding to the features identified in the image. The images corresponding to the reference features are selected from the displayed image information, and the types corresponding to the selected images are used as the feature types.

[0085] Step S402: Determine different feature areas according to feature types.

[0086] The feature field refers to the text field corresponding to the feature type, and the corresponding feature field is matched from the preset feature database through the feature field. The feature database stores text fields corresponding to different feature types. The feature database is a database set by humans and will not be described in detail here.

[0087] Step S403: Determine whether the feature domain falls into the combined vocabulary domain. If not, jump to step S404. If yes, jump to step S405.

[0088] By determining whether the feature field falls into the combined vocabulary field, it is determined whether the short video is the type corresponding to the combination of pictures and texts.

[0089] Step S404: continue to use the short videos corresponding to the combined vocabulary fields that are inconsistent with the reference matching fields as marked short videos.

[0090] When the feature domain does not fall into the combined vocabulary domain, it means that the short video is not the type corresponding to the text-image combination, so the short video corresponding to the combined vocabulary domain inconsistent with the benchmark combination domain continues to be used as the marked short video.

[0091] Step S405: Determine the maximum number of overlaps of a single feature field based on the combined vocabulary field.

[0092] The maximum number of overlaps refers to the maximum number of times a single feature field overlaps in the combined vocabulary field. When the feature field falls into the combined vocabulary field, it means that the short video is a type corresponding to the combination of pictures and texts. Therefore, the number of fields in the combined vocabulary field that are the same as the feature field is counted as the maximum number of overlaps.

[0093] Step S406: Determine the occurrence times of different characteristic fields in the combined vocabulary field according to the characteristic fields.

[0094] The number of occurrences refers to the number of times different feature fields appear in the combined vocabulary field. By comparing each feature field with the combined vocabulary field, when a single feature field overlaps in the combined vocabulary field, it means that a single feature field appears in the combined vocabulary field, so the number is increased by one, and the comparison and analysis of the next feature field is continued until all different feature fields are matched, and the number obtained is taken as the number of occurrences.

[0095] Step S407: When the number of occurrences is greater than the maximum number of overlaps, the short video corresponding to the feature field falling into the combined vocabulary field is pushed to the short video platform first.

[0096] When the number of occurrences is greater than the maximum number of overlaps, it means that the short video uses text to express and describe the picture, so the short video whose feature field falls into the combined vocabulary field will be pushed to the short video platform first.

[0097] Step S408: When the number of occurrences is less than the maximum number of overlaps, continue to use the short videos corresponding to the combined vocabulary fields that are inconsistent with the reference matching fields as marked short videos.

[0098] When the number of occurrences is less than the maximum number of overlaps, it means that the short video is expressed by a combination of pictures and texts. Therefore, the short video corresponding to the combined vocabulary field that is inconsistent with the benchmark matching field will continue to be used as a marked short video, so that the key points expressed by the short video can be classified, and the short videos with the focus on explosive content will be pushed first, thereby improving the accuracy of the short video platform in pushing explosive short videos.

[0099] Methods for determining tagged content include: Step S500: When the amount of browsing related information is greater than a preset benchmark amount of browsing related information, trigger information of the user on the short video platform is obtained.

[0100] Trigger information refers to information such as users liking, collecting, and commenting on short videos on the short video platform. The response status corresponding to the likes, collections, and comments is retrieved from the short video platform, and the information corresponding to the response status is used as trigger information.

[0101] Step S501: Determine the marking duration of the short video according to the trigger information.

[0102] The marked duration refers to the time length of the marked short video played when the user outputs the trigger information. When the user outputs the trigger information, the current duration of the marked short video is marked, and the marked duration is used as the marked duration.

[0103] Step S502: Determine the marking type according to the marked short video and the marking duration, where the marking type includes text, image, music, etc.

[0104] The tag type refers to the type of text, image and music expressed by the tagged short video. When the tagged short video has a marked duration, the type of the field obtained by combining the combined vocabulary field and the feature field corresponding to the tagged short video is used as the tag type.

[0105] Step S503: When the tag type is music, background music information is used as tag content.

[0106] When the tag type is music, it means that the user is interested in the music played in the tagged short video, so the background music information is used as the tag content.

[0107] Step S504: when the mark type is picture and text, determine the mark image information and mark text information corresponding to the mark duration according to the displayed image information and the updated initial text.

[0108] The marked image information refers to the displayed image information of the marked short video within the marked duration, and the marked text information refers to the updated initial text corresponding to the marked short video within the marked duration. When the marking type is picture and text, it means that the user is interested in the picture and text played by the marked short video, so the displayed image information of the marked short video within the marked duration is retrieved as the marked image information, and the updated initial text corresponding to the marked short video within the marked duration is retrieved as the marked text information.

[0109] Step S505: When the number of occurrences is greater than the maximum number of overlaps, the feature area corresponding to the marked image information is used as the marked content.

[0110] When the number of occurrences is greater than the maximum number of overlaps, it indicates that the marked short video uses text to express and describe the image, so the characteristic area corresponding to the marked image information is used as the marked content.

[0111] Step S506: When the number of occurrences is less than the maximum number of overlaps, a combination of words in the marked text information is selected as a marked word according to the marked text information, and the marked word is used as the marked content.

[0112] The marked vocabulary refers to the corresponding vocabulary in the marked text information. When the number of occurrences is less than the maximum number of overlaps, it means that the marked short video is expressed by pictures and texts. Therefore, the corresponding combination of vocabulary in the marked text information is selected from the combination of vocabulary corresponding to the updated initial text as the marked vocabulary, and the marked vocabulary is used as the marked content.

[0113] Reference Figure 4 , the verification methods for browsing related information include: Step S600: Obtain the user's control range on the short video platform and the video update type of the short video platform.

[0114] The control range refers to the range of control operations such as sliding short videos and clicking by users on the short video platform. The range obtained by combining the positions corresponding to the operations such as sliding short videos and clicking by users from the system of the short video platform is used as the control range. The video update type refers to the operation type of short video update performed by the short video platform. The video update type can be sliding from bottom to top to update the short video. The information type of the short video update is retrieved from the system as the video update type.

[0115] Step S601: Determine an estimated control position according to the control range.

[0116] The estimated control position refers to the estimated center position of the control range on the short video platform, and the center position of the shape corresponding to the control range is used as the estimated control position.

[0117] Step S602: Retrieve the control direction according to the video update type.

[0118] The control direction refers to the direction in which different fingers of the user slide when controlling the short video update on the short video platform, and the control direction is retrieved from the video update type.

[0119] Step S603: Determine the user's holding state according to the estimated control position and control direction.

[0120] The holding state refers to the palm state of the user when holding the mobile terminal. The holding state includes the state of the thumb operating when holding with one hand or the state of the fingers of the other hand operating when holding with one hand. The holding state is matched from the preset holding database by estimating the control position and control direction. The holding database stores holding states corresponding to different estimated control positions and control directions. The holding database is a manually set database and will not be described in detail here.

[0121] Step S604: matching a reference error distance from a preset error database based on the holding state.

[0122] The reference error distance refers to the error distance of the control range expansion corresponding to different holding states. The reference error distance is matched from a preset error database according to the holding state. The error database stores the reference error distances corresponding to different holding states. The error database is a manually set database and will not be described in detail here.

[0123] Step S605: Calculate the error distance according to the estimated control position and the preset trigger position.

[0124] The trigger position refers to the position corresponding to the user's output of trigger information on the short video platform. The error distance refers to the distance between the estimated control position and the trigger position. The straight-line distance between the estimated control position and the trigger position is calculated, and then the point on the control range where the straight-line distance is located is marked as the intersection point, and the distance between the intersection point and the trigger position is calculated as the error distance.

[0125] Step S606: When the error distance is greater than the reference error distance, the total video duration of the marked short video and the viewing duration of the marked short video by the user are obtained.

[0126] When the error distance is not greater than the reference error distance, it indicates that the probability that the user accidentally touches the trigger position to output the trigger information is low, so the browsing related information continues to be output.

[0127] The total video duration refers to the total time length corresponding to the marked short video, and the viewing time refers to the time the marked short video is viewed on the user's short video platform. When the error distance is greater than the benchmark error distance, it indicates that the probability that the user accidentally touches the trigger position to output the trigger information is high. Therefore, the total duration of the marked short video is retrieved from the system of the short video platform as the total video duration, and the time the marked short video stays on the user's short video platform is used as the viewing time.

[0128] Step S607: Determine a benchmark viewing duration based on the total video duration.

[0129] The benchmark viewing time refers to the minimum time for a user to play a short video that he or she is interested in. The benchmark viewing time is matched from a preset playback database according to the total duration of the video. The playback database stores the benchmark viewing time corresponding to different total durations of videos. The playback database is a manually set database and will not be described in detail here.

[0130] Step S608: When the viewing time is not greater than the reference viewing time, the browsing related information is updated according to the viewing time.

[0131] When the viewing time is longer than the benchmark viewing time, it means that the user is interested in the marked short video and the trigger information is output, so the browsing related information continues to be output. When the viewing time is not longer than the benchmark viewing time, it means that the user is not interested in the marked short video and the trigger information is output, and there is a situation where the trigger position is touched unintentionally, so the traffic of the marked short videos that output trigger information and the viewing time is not longer than the benchmark viewing time is counted from the short video system, and the difference between the data corresponding to the browsing related information and the statistical traffic data is calculated, and the data corresponding to the difference is integrated to obtain new browsing related information, so that it is not easy for users to accidentally like or swipe likes to cause the short video to become a hit, thereby improving the accuracy of the push of hit short videos.

[0132] Further, the method after determining the estimated control position includes: Step S6010: Obtain the number of control ranges as the control number.

[0133] The control quantity refers to the number of control ranges that appear in the short video platform. It is calculated by counting the non-overlapping control ranges in the short video platform and using the counting result as the control quantity.

[0134] Step S6011: When the control quantity is greater than a preset reference control quantity, update the browse related information according to the control quantity.

[0135] The benchmark control number refers to the benchmark number corresponding to the control range when the user normally uses the short video platform. The benchmark control number is set in advance by a person skilled in the art and will not be described in detail here. When the control number is greater than the preset benchmark control number, it means that abnormal conditions such as water droplets appear on the user's terminal to affect the normal use of the short video platform. Therefore, the browsing-related information of the marked short videos corresponding to the control number is counted from the short video system, and the difference between the data corresponding to the browsing-related information and the statistical traffic data is calculated, and the data corresponding to the difference is integrated to obtain new browsing-related information.

[0136] Step S6012: When the control quantity is not greater than a preset reference control quantity, the interval control distance is calculated according to the estimated control position corresponding to each control range.

[0137] The interval control distance refers to the straight-line distance between each control range. When the control number is not greater than the benchmark control number, it means that there is no abnormal situation of water droplets on the user's terminal to affect the normal use of the short video platform. Therefore, the straight-line distance between the estimated control positions corresponding to each control range is calculated, and the calculated straight-line distance is used as the interval control distance.

[0138] Step S6013: When the interval control distance is not greater than the preset reference control distance, the control interval time between each control range is obtained.

[0139] The benchmark control distance refers to the maximum distance corresponding to abnormal situations such as disconnection and accidental touch due to sweat when the user uses the short video platform. The benchmark control distance is set in advance by technicians in this field and will not be elaborated here.

[0140] The control interval time refers to the time parameter value corresponding to the interval of each control range. When the interval control distance is greater than the benchmark control distance, it means that there are no abnormal situations such as disconnection and accidental touch due to sweat when the user uses the short video platform, so the browsing related information continues to be output. When the interval control distance is not greater than the benchmark control distance, it means that there is a high probability of abnormal situations such as disconnection and accidental touch due to sweat when the user uses the short video platform, so the time points of each range are marked and the difference between each time point is calculated, and the difference is used as the control interval time.

[0141] Step S6014: Update browsing related information according to the control interval time.

[0142] When the control interval is less than the preset benchmark interval, it means that abnormal situations such as disconnection and accidental touch due to sweat occur when the user uses the short video platform. Therefore, the traffic of short videos marked with the corresponding control quantity and control interval is counted from the short video system, and the difference between the data corresponding to the browsing related information and the statistical traffic data is calculated, and the data corresponding to the difference is integrated to obtain new browsing related information.

[0143] The verification method for browsing related information also includes: Step S700: Obtain a short video of the user in a preset video storage library and use it as a selected short video.

[0144] The video repository refers to a database for storing short videos corresponding to the receipt of trigger information. The video repository is set in advance by a person skilled in the art and will not be described in detail here. The selected short video refers to a short video in the video repository. The system retrieves a short video in the video repository as the selected short video.

[0145] Step S701: Determine the selected content based on the sound detection information and display image information corresponding to the selected short video.

[0146] The selected content refers to the marked content corresponding to the selected short video. The selected content is obtained by analyzing the images, texts and music in the selected short video through the audio analysis system and the benchmark features.

[0147] Step S702: Retrieve a selected video area based on the selected content.

[0148] The selected video field refers to the text field to which the selected short video belongs, and the selected video field is retrieved from the selected content.

[0149] Step S703: When the user outputs trigger information, the marked short video corresponding to the received trigger information is defined as a trigger short video, and the trigger short video is recorded into a preset video storage library, and the trigger content is determined based on the sound detection information and display image information corresponding to the trigger short video.

[0150] The triggered short video refers to the marked short video corresponding to the current time when the trigger information is received. When the user outputs the trigger information, the marked short video corresponding to the time when the trigger information is received is defined as the triggered short video, and the triggered short video is recorded into the preset video storage library. The trigger content refers to the marked content corresponding to the triggered short video, and the trigger content is obtained by analyzing the sound detection information and display image information corresponding to the trigger short video.

[0151] Step S704: Retrieve a trigger video area based on the trigger content.

[0152] The trigger video domain refers to the domain to which the text in the trigger short video belongs, and the trigger video domain is retrieved from the trigger content.

[0153] Step S705: Determine whether the trigger video area is consistent with the selected video area.

[0154] By determining whether the field of the triggering video is consistent with the field of the selected video, it is determined whether the triggering short video is in the field that the user is interested in.

[0155] Step S706: When the triggering video domain is inconsistent with the selected video domain, the browsing related information is updated according to the selected video domain.

[0156] When the selected video field is inconsistent with the trigger video field, it means that the probability that the triggered short video is the field of interest to the user is low. Therefore, the traffic of marked short videos in the trigger video field and the selected video field is counted from the short video system, and the difference between the data corresponding to the browsing related information and the counted traffic data is calculated, and the data corresponding to the difference is integrated to obtain new browsing related information.

[0157] The verification method for browsing related information also includes: Step S800: When the selected video area is consistent with the triggering video area, obtain the comment area corresponding to the triggered short video in the short video platform.

[0158] The comment area refers to the area where users comment on the triggered short video. When the field of the triggered video is consistent with the field of the selected video, it means that the probability that the triggered short video is the field that the user is interested in is high, so the area where users comment on the triggered short video is retrieved from the short video platform as the comment area.

[0159] Step S801: Determine whether the comment area receives trigger information. If not, jump to step S802. If yes, jump to step S803.

[0160] By determining whether the comment area receives trigger information, it is determined whether the user comments on the triggered short video.

[0161] Step S802: Continue to update and browse related information according to the selected video field.

[0162] When the comment area does not receive trigger information, it means that the user has not commented on the triggered short video, so step S706 is continued.

[0163] Step S803: Obtain the user's comment information.

[0164] Comment information refers to the text, pictures, expressions and other information that the current user comments on the triggered short video. When the comment area does not receive trigger information, it means that the user has commented on the triggered short video, so the text, pictures, expressions and other information that the current user comments on the triggered short video is retrieved from the system as comment information.

[0165] Step S804: When text appears in the comment information, the comment text is determined based on the comment information input into a preset text recognition component, and the remaining comment text in the comment area is used as the remaining comment text.

[0166] The text recognition component refers to a component used to recognize text, the comment text refers to the text that the current user comments on the triggered short video, and the remaining comment text refers to the text that other users comment on the triggered short video. When non-text information such as pictures and emoticons appear in the comment information, it means that the user makes a simple comment, so continue to execute step S706. When text appears in the comment information, the text recognition component is used to scan the comment area, and the recognition result corresponding to the comment information is used as the comment text, and the remaining text in the comment area is used as the remaining comment text.

[0167] Step S805: Determine the degree of vocabulary overlap based on the remaining comment text and the comment text input into the vocabulary database.

[0168] The degree of vocabulary overlap refers to the degree of vocabulary overlap between the comment text and the remaining comment text. The remaining comment text and the comment text are input into the vocabulary database to match the corresponding vocabulary, and the remaining comment text and the text matched by the comment text are compared for overlap, and the result of the overlap comparison is used as the degree of vocabulary overlap.

[0169] Step S806: When the vocabulary overlap degree is less than a preset reference overlap degree, determining the comment area according to the comment text.

[0170] The benchmark overlap degree refers to the maximum overlap degree of the corresponding comment text after the current user copies other users and modifies them. The benchmark overlap degree is set in advance by those skilled in the art and will not be elaborated here.

[0171] The comment field refers to the field corresponding to the comment text. When the vocabulary overlap degree is not less than the reference overlap degree, it means that the current user has copied the comment text of other users and slightly modified it, so step S706 is continued. When the vocabulary overlap degree is less than the reference overlap degree, it means that the current user has not copied the comment text of other users and slightly modified it, so the comment text is split to combine into various words, and the comment field is matched from the vocabulary database through the words. The vocabulary database stores the fields corresponding to different words, which will not be described in detail here.

[0172] Step S807: When the comment field and the trigger video field are inconsistent, the browsing related information is updated according to the comment field and the trigger video field.

[0173] When the comment field is consistent with the trigger video field, it means that the current user is making normal comments, so the browsing related information continues to be output; when the comment field is consistent with the trigger video field, it means that the current user is making random comments, so the traffic of the marked short videos corresponding to the comment field is counted from the short video system, and the difference between the data corresponding to the browsing related information and the counted traffic data is calculated, and the data corresponding to the difference is integrated to obtain new browsing related information.

[0174] Based on the same inventive concept, an embodiment of the present invention provides a short video element analysis system, including: The acquisition module is used for sound detection information, browsing related information, displaying image information, triggering information, control range, video update type, total video length, viewing time, selected short videos, comment area and comment information.

[0175] A memory for storing a program for a short video element analysis method.

[0176] The processor is used to load, execute and implement the program stored in the memory.

[0177] Based on the same inventive concept, an embodiment of the present invention provides an intelligent terminal, including a memory and a processor, wherein the memory stores a computer program that can be loaded by the processor and execute a short video element analysis method.

[0178] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, device and unit described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0179] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A short video element analysis method, characterized in that: include: Get the sound detection information of the short video; Inputting the sound detection information into a preset audio analysis system to obtain an initial text; When music appears in the sound detection information, determine the background music according to the sound detection information; Matching background music information through a preset music library based on background music; Determine the music text according to the background music information; Eliminate the corresponding words in the initial text according to the background rhythm and the music words, and use the initial text after elimination as the new initial text; Splitting based on the characters in the new initial text, and combining the split characters adjacently to determine combined words; Determine whether the combined vocabulary is consistent with the preset benchmark vocabulary; If the combined vocabulary is consistent with the preset benchmark vocabulary, the short video corresponding to the combined vocabulary consistent with the preset benchmark vocabulary will be pushed to the short video platform first; If the combined vocabulary is inconsistent with the preset reference vocabulary, the short video corresponding to the combined vocabulary inconsistent with the preset reference vocabulary is used as the marked short video; Obtain browsing related information of the marked short video; The marked content is determined based on browsing related information and marking short videos, and the marked short videos are pushed to the short video platform corresponding to the marked content and the marked content is added to a preset push database.

2. A short video element analysis method according to claim 1, characterized in that: Methods for determining the combination of words include: Selecting uncombined characters from the updated initial text according to the combined vocabulary; Determine uncombined syllables based on uncombined characters; Combining the adjacent uncombined syllables, and taking the combined syllables as the selected syllables; Selecting different selected words from a preset word database based on the selected syllable; Determine the domain of combined vocabulary based on combined vocabulary; Determine the selected vocabulary area based on the selected vocabulary; Determine whether the selected vocabulary domain is consistent with the combined vocabulary domain; When the selected vocabulary field is consistent with the combined vocabulary field, the selected vocabulary corresponding to the selected vocabulary field consistent with the combined vocabulary field is added to the combined vocabulary; When the selected vocabulary field and the combined vocabulary field are inconsistent, the characters corresponding to the selected vocabulary and the combined vocabulary adjacent to the selected vocabulary are input into the vocabulary database for recombination, and when the fields corresponding to the recombined vocabulary are consistent, the recombined vocabulary is added to the combined vocabulary.

3. A short video element analysis method according to claim 2, characterized in that: The method of using the short video corresponding to the combined vocabulary inconsistent with the preset reference vocabulary as a tag short video includes: Determining a benchmark matching vocabulary based on the benchmark vocabulary input into a vocabulary database; Determine whether the combined vocabulary is consistent with the benchmark matching vocabulary; If the combined vocabulary is consistent with the benchmark matching vocabulary, the short video corresponding to the combined vocabulary consistent with the benchmark matching vocabulary will be pushed to the short video platform first; If the combined vocabulary is inconsistent with the benchmark matching vocabulary, the benchmark matching field is determined according to the benchmark matching vocabulary; When the combined vocabulary field is inconsistent with the reference matching field, the short video corresponding to the combined vocabulary field inconsistent with the reference matching field is used as the marked short video.

4. A short video element analysis method according to claim 3, characterized in that: Methods for marking short videos corresponding to the combined vocabulary field that is inconsistent with the benchmark matching field include: Obtain the display image information corresponding to the short video; Determine the feature type according to the displayed image information and the preset reference features; Determine different feature areas based on feature types; Determine whether the feature domain falls into the combined vocabulary domain; If the feature domain does not fall into the combined vocabulary domain, the short video corresponding to the combined vocabulary domain that is inconsistent with the benchmark matching domain continues to be used as the marked short video; If the feature domain falls into the combined vocabulary domain, the maximum number of overlaps of a single feature domain is determined based on the combined vocabulary domain; According to the feature domain, the occurrence frequency of different feature domains in the combined vocabulary domain is determined; When the number of occurrences is greater than the maximum number of overlaps, the short videos corresponding to the feature fields that fall into the combined vocabulary fields are pushed to the short video platform first; When the number of occurrences is less than the maximum number of overlaps, the short video corresponding to the combined vocabulary field that is inconsistent with the reference matching field continues to be used as the marked short video.

5. A short video element analysis method according to claim 4, characterized in that: Methods for determining tagged content include: When the amount of relevant information browsed is greater than a preset benchmark, trigger information of the user on the short video platform is obtained; Determine the marking duration of the short video according to the trigger information; Determine the tag type based on the tagged short video and the tag duration, and the tag type includes text, image, music, etc.; When the tag type is music, the background music information is used as the tag content; When the mark type is a picture or text, the mark image information and the mark text information corresponding to the mark duration are determined according to the displayed image information and the updated initial text; When the number of occurrences is greater than the maximum number of overlaps, the feature area corresponding to the marked image information is used as the marked content; When the number of occurrences is less than the maximum number of overlaps, a combination of words in the tag text information is selected as a tag word according to the tag text information, and the tag word is used as the tag content.

6. A short video element analysis method according to claim 5, characterized in that: Verification methods for browsing related information include: Obtain the user's control scope on the short video platform and the video update type of the short video platform; Determine the estimated control position based on the control range; According to the video update type to call the control direction; Determine the user's holding state based on the estimated control position and control direction; Matching a reference error distance from a preset error database based on the grip state; Calculate the error distance based on the estimated control position and the preset trigger position; When the error distance is greater than the reference error distance, the total video duration of the marked short video and the viewing duration of the marked short video by the user are obtained; Determine baseline watch time based on total video length; When the viewing time is not greater than the benchmark viewing time, the browsing related information is updated according to the viewing time.

7. A short video element analysis method according to claim 5, characterized in that: The verification method for browsing related information also includes: Obtaining a short video of the user in a preset video storage repository and using it as a selected short video; Determine the selected content based on the sound detection information and the displayed image information corresponding to the selected short video; Retrieving a selected video area based on the selected content; When the user outputs the trigger information, the marked short video corresponding to the received trigger information is defined as the trigger short video, and the trigger short video is recorded into the preset video storage library, and the trigger content is determined according to the sound detection information and the display image information corresponding to the trigger short video; Retrieving a trigger video area based on the trigger content; determining whether the trigger video field is consistent with the selected video field; When the trigger video domain is inconsistent with the selected video domain, the browse related information is updated according to the selected video domain.

8. A short video element analysis method according to claim 7, characterized in that: The verification method for browsing related information also includes: When the selected video area is consistent with the triggering video area, obtaining the comment area corresponding to the triggering short video in the short video platform; Determine whether the comment area receives trigger information; When no trigger information is received in the comment area, continue to execute to update the browsing related information according to the selected video area; When receiving trigger information in the comment area, obtain the user's comment information; When text appears in the comment information, the comment text is determined based on the comment information input into a preset text recognition component, and the remaining comment text in the comment area is used as the remaining comment text; According to the remaining comment text and the comment text input into the vocabulary database to determine the degree of vocabulary overlap; When the vocabulary overlap is less than a preset benchmark overlap, the comment area is determined based on the comment text; When the comment field and the trigger video field are inconsistent, the browsing related information is updated according to the comment field and the trigger video field.

9. A short video element analysis system, characterized in that: include: An acquisition module is used to obtain sound detection information, browse related information, display image information, trigger information, control range, video update type, total video duration, viewing duration, selected short videos, comment area, and comment information; A memory, used to store a program of a short video element analysis method according to any one of claims 1 to 8; The processor is used to load, execute and implement the program stored in the memory.

10. An intelligent terminal, characterized in that: It comprises a memory and a processor, wherein the memory stores a computer program that can be loaded by the processor and execute a short video element analysis method as described in any one of claims 1 to 8.

Images