A method and system for generating interpretation of element information in application video conference
By semantic analysis and retrieval of the first draft of the conference and real-time voice in the video conference, relevant information fragments were selected, and the problem of difficulty in explaining professional knowledge in professional video conferences was solved and the understanding ability of conference participants was improved.
Patent Information
- Application Number
- CN202411167069.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-23
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-08-23
AI Technical Summary
In professional video conferences, due to the complex topics of discussion, it is difficult for participants to fully understand the content of the conference, and the existing technology fails to effectively explain the expertise in the conference in real time.
By collecting the information on the first draft of the meeting for semantic analysis, the conference keyword group is determined, and relevant information fragments are selected based on the network and self-used databases; combined with real-time speech recognition technology, keywords are extracted and information fragments are screened to provide explanatory recommendation materials.
It has achieved timely explanation of professional elements in professional video conferences, and improved the meeting awareness and understanding of participants.
Smart Images

Figure CN118694886B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of video conferencing, and in particular to a method and system for interpreting and generating element information in a video conference. Background Art
[0002] Video conferencing is a technology based on network communication technology that allows people in different locations to share real-time video, audio and data. With the popularization of the Internet and the development of communication technology, video conferencing has become an indispensable part of modern remote work. In professional video conferencing, since the topics discussed often involve complex knowledge and terminology, in order to ensure that all participants can fully understand the content of the meeting, real-time explanation of the professional knowledge mentioned in the video conference is required. Summary of the invention
[0003] The purpose of the present invention is to provide a generation method and system capable of interpreting professional element information in a meeting in real time.
[0004] In order to achieve the above object, the present invention adopts the following technical solutions:
[0005] A method for generating and interpreting element information in a video conference, comprising:
[0006] Collect the meeting draft information of different meeting participants, perform semantic analysis on each meeting draft information, determine several meeting keywords, and freely combine several meeting keywords belonging to the same meeting draft information to obtain several meeting keyword groups;
[0007] Conducting a network search for the conference keyword group, and determining the relevance of the conference keyword group based on the search results, and recording a number of conference keyword groups whose relevance is greater than or equal to a preset value as reference conference keyword groups;
[0008] Search the reference conference keyword phrases on the Internet and in-house databases, and based on information value analysis, extract information fragments to obtain several information fragments, and record the collection of information fragments as an information fragment set;
[0009] Real-time speech in video conferences is collected in real time, and speech recognition technology is used to convert real-time speech into real-time quotation information. Semantic analysis is performed on the real-time quotation information to extract several real-time keywords. Based on the occurrence frequency of real-time keywords and the order correlation characteristics of their mutual occurrence, information fragments in the information fragment set are screened, and the screened information fragments are displayed as explanatory recommendation materials.
[0010] In some embodiments disclosed in the present invention, the method for determining a number of conference keywords includes:
[0011] Acquire the conference theme, and set a number of reference keywords for the conference theme, and perform synonym analysis on the determined conference keywords based on the reference keywords, and filter the keywords in the conference draft information based on the result of the synonym analysis, and divide the keywords into focus keywords and general keywords based on the filtering result;
[0012] Based on the semantic relationship between the focus keywords and the regular keywords in the sentence, the regular keywords are screened, and the screened regular keywords and focus keywords are simultaneously identified as conference keywords.
[0013] In some embodiments disclosed in the present invention, the method for determining the relevance of the conference keyword group includes:
[0014] Construct query statements based on conference keyword groups, submit the query statements to search engines and academic databases, and select the retrieved document information according to the preset demand volume;
[0015] Performing semantic analysis on the selected document information, including determining expression features of conference keywords in the conference keyword group relative to the document information, and determining expression features of synonyms of the conference keywords relative to the document information;
[0016] Based on the first expression feature and the second expression feature, the relevance of the conference keyword group is determined.
[0017] In some embodiments disclosed in the present invention, the expression features include the proportion of keywords in each document information, and the complete mapping proportion of the conference keyword group to the document information;
[0018] The keyword occupancy ratio is the proportion of conference keywords and synonyms of conference keywords in the conference keyword group relative to the keywords in the document information;
[0019] Among them, the method for determining the complete mapping ratio is to mark the conference keywords and synonyms of the conference keywords that can be found mapped in the same document information, and determine the ratio of the number of marked keywords in the conference keyword group to the total number, and recognize the ratio as the complete mapping ratio. Among them, if the synonyms of the conference keywords are marked, the corresponding conference keywords in the conference keyword group are also marked.
[0020] In some embodiments disclosed in the present invention, the expression for calculating the relevance of the conference keyword group is:
[0021] ;
[0022] in, For the degree of relevance, is the fractional expansion adjustment coefficient, n is the number of document information, is the expression feature analysis function of the i-th document information, if is greater than or equal to the preset value, then Output 1, otherwise output 0. is the proportion of keywords in the i-th document information, is the occupancy ratio adjustment constant, is the complete mapping ratio of the i-th document information, and b is the relevance adjustment constant;
[0023] Among them, the expression of keyword occupancy ratio is:
[0024] ;
[0025] in, is the total number of conference keywords and their synonyms in the i-th document information, is the number of all keywords in the i-th document information;
[0026] Among them, the expression of the complete mapping ratio is:
[0027] ;
[0028] in, is the number of marked conference keywords in the conference keyword group, is the number of all conference keywords in the conference keyword group, and the marked conference keywords are conference keywords that can find mapping keywords or synonyms in the document information.
[0029] In some embodiments disclosed in the present invention, based on information value analysis, the method for intercepting information fragments includes:
[0030] The retrieved information content is analyzed in terms of content relevance, timeliness, source credibility, number of citations, and information relevance, and weight coefficients that need attention are configured for content relevance, timeliness, source credibility, number of citations, and information relevance, and content-related value conversion operators, timeliness value conversion operators, source credibility value conversion operators, number of citations value conversion operators, and information relevance value conversion operators are constructed respectively, and the information value is obtained through comprehensive calculation;
[0031] Real-time evaluation of the information value of the fragments in the information content is performed. If the information value is greater than or equal to a preset value, the complete part of the fragment of the information content is intercepted to obtain the information fragment;
[0032] The expression for calculating information value is:
[0033] ;
[0034] in, is the information value, is the first weight coefficient that needs attention, is the content relevance parameter of the information content, is the second weight coefficient that needs attention, is a parameter related to the timeliness of information content, is the third weight coefficient that needs attention, is a parameter related to source credibility, is the fourth weight coefficient that needs attention, is the parameter related to the number of citations, is the fifth weight coefficient that needs attention, It is information-related parameter.
[0035] In some embodiments disclosed in the present invention, based on the occurrence frequency of real-time keywords and the order correlation characteristics of their occurrence, the method for screening information fragments in the information fragment set includes:
[0036] Performing semantic analysis on the information fragments in the information fragment set, determining the comparison keywords in the information fragments, and determining the occurrence frequency and the order of occurrence correlation features of the comparison keywords;
[0037] The occurrence frequency and order-related features of the comparison keywords are compared with the real-time keywords, and based on the comparison results, the degree of consistency between the information fragment and the real-time quotation information is determined, and based on the relationship between the degree of consistency, the information fragment is displayed preferentially.
[0038] In some embodiments disclosed in the present invention, the method of comparing the occurrence frequency and occurrence order correlation features of the comparison keywords with the real-time keywords includes:
[0039] A plane text grid template is constructed, wherein the plane text grid is a plurality of text grids arranged in a uniform array;
[0040] Fill the real-time quotation information into the plane text grid template, mark the real-time keywords therein, and connect the text grid groups corresponding to the real-time keywords to obtain the real-time quotation text grid expression diagram based on the semantic connection relationship between different real-time keywords;
[0041] Fill the information fragments into the flat text grid template, mark the comparison keywords therein, and connect the text grid groups corresponding to the comparison keywords to obtain the information fragment text grid expression diagram based on the semantic connection relationship between different comparison keywords;
[0042] All the text grids in the real-time quotation text grid expression diagram and the information fragment text grid expression diagram are sequentially translated and adjusted, and the real-time quotation text grid expression diagram and the information fragment text grid expression diagram are compared, and based on the comparison result, the degree of consistency between the information fragment and the real-time quotation information is determined.
[0043] In some embodiments disclosed in the present invention, the expression for calculating the degree of conformity is:
[0044] ;
[0045] in, For the degree of relevance, The conformity adjustment factor, The number of matching keyword pairs in the real-time quotation text expression graph and the information fragment text expression graph. is the number of real-time keywords in the real-time quotation text expression graph, The number of keywords compared in the information fragment text grid expression diagram, is the connection quantity adjustment factor, is the number of matching lines between text grid groups, is the number of lines in the real-time quotation text expression graph, The information fragment text grid expresses the number of connections in the graph, and d is the constant for adjusting the number of connections;
[0046] Among them, the method for judging the matching keyword pair is to set range circles for both the real-time keyword and the comparison keyword. If the range circles of the two keywords overlap, the two keywords are judged to be the matching keyword pair.
[0047] In some embodiments disclosed in the present invention, there is also disclosed a system for generating interpretation of element information in a video conference, including:
[0048] The first module is used to collect the meeting draft information of different meeting participants, and perform semantic analysis on each meeting draft information to determine a number of meeting keywords, and freely combine a number of meeting keywords belonging to the same meeting draft information to obtain a number of meeting keyword groups;
[0049] The second module is used to perform a network search for the conference keyword group, and based on the search results, determine the relevance of the conference keyword group, and record a number of conference keyword groups with a relevance greater than or equal to a preset value as reference conference keyword groups;
[0050] The third module is used to search the reference conference keyword phrases on the Internet and in a self-used database, and based on the information value analysis, to intercept the information fragments to obtain a number of information fragments, and record the collection of information fragments as an information fragment set;
[0051] The fourth module is used to collect real-time voice in video conferences in real time, and use voice recognition technology to convert real-time voice into real-time quotation information, and perform semantic analysis on the real-time quotation information to extract several real-time keywords. Based on the occurrence frequency of real-time keywords and the order correlation characteristics of their mutual occurrence, the information fragments in the information fragment set are screened, and the screened information fragments are displayed as explanatory recommendation materials.
[0052] The present invention discloses a method and system for generating interpretation of element information in an applied video conference, and relates to the field of video conference. Specifically, the present invention discloses freely combining a number of conference keywords belonging to the same conference draft information to obtain a number of conference keyword groups, performing a network search on the conference keyword groups, and recording a number of conference keyword groups with a correlation degree greater than or equal to a preset value as reference conference keyword groups, performing a network and a self-use database search on the reference conference keyword groups to obtain a number of information fragments, and recording the collection of information fragments as an information fragment set, screening the information fragments in the information fragment set based on the occurrence frequency of real-time keywords and the order correlation characteristics of their occurrence, and displaying the screened information fragments as explanatory recommendation materials. The present invention realizes timely interpretation of professional elements in professional video conferences through the above technical scheme, and improves the participants' cognition and understanding of the conference. BRIEF DESCRIPTION OF THE DRAWINGS
[0053] Figure 1 A method step diagram of a method for interpreting element information in a video conference disclosed in an embodiment of the present invention. DETAILED DESCRIPTION
[0054] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.
[0055] A method for generating and interpreting element information in a video conference, comprising:
[0056] Step S100, collecting meeting draft information of different meeting participants, and performing semantic analysis on each meeting draft information to determine a number of meeting keywords, and freely combining a number of meeting keywords belonging to the same meeting draft information to obtain a number of meeting keyword groups.
[0057] In this step, the system collects preliminary speeches or documents submitted by meeting participants at the beginning of the meeting; these drafts can include meeting agendas, proposals, lists of questions, or preliminary ideas. The system then uses natural language processing (NLP) technology to perform semantic analysis on these texts. NLP technologies include part-of-speech tagging, syntactic analysis, entity recognition, and sentiment analysis, etc., with the goal of extracting meaningful information and key concepts from the text. Through this process, the system can identify several meeting keywords that represent the core content and discussion focus of the meeting. Next, the system will freely combine multiple keywords belonging to the same meeting draft to form multiple meeting keyword groups, which provide the basis for subsequent information retrieval and recommendation.
[0058] In some embodiments disclosed in the present invention, the method for determining a number of conference keywords includes:
[0059] Step S101, obtain the conference theme, and set a number of reference keywords for the conference theme, and perform synonym analysis on the determined conference keywords based on the reference keywords, and based on the results of the synonym analysis, filter the keywords in the conference draft information, and based on the filtering results, divide the keywords into focus keywords and general keywords.
[0060] Step S102: based on the semantic relationship between the focus keywords and the regular keywords in the sentence, the regular keywords are screened, and the screened regular keywords and the focus keywords are simultaneously identified as conference keywords.
[0061] Step S200: perform a network search for conference keyword groups, determine the relevance of conference keyword groups based on the search results, and record several conference keyword groups with relevance greater than or equal to a preset value as reference conference keyword groups.
[0062] In this step, the system will search the conference keyword phrases generated in the previous step on the Internet to find relevant information resources. Internet search can include the use of search engines, professional databases, or knowledge graphs; the system will evaluate the relevance of each keyword phrase based on the retrieved results, which may involve calculating the frequency of occurrence of the keyword phrase, the number of related links, the authority of the resource, and other factors; those keyword phrases whose relevance exceeds the preset threshold will be marked as reference conference keyword phrases, which will be used for more in-depth information retrieval in the next step.
[0063] In some embodiments disclosed in the present invention, the method for determining the relevance of the conference keyword group includes:
[0064] Step S201, construct a query statement based on the conference keyword group, submit the query statement to the search engine and the academic database, and select the retrieved document information according to the preset demand volume.
[0065] Step S202 , performing semantic analysis on the selected document information, including determining expression features of conference keywords in the conference keyword group relative to the document information, and determining expression features of synonyms of the conference keywords relative to the document information.
[0066] Step S203: determining the relevance of the conference keyword group based on the first expression feature and the second expression feature.
[0067] In some embodiments disclosed in the present invention, the expression features include the proportion of keywords in each document information, and the complete mapping proportion of the conference keyword group to the document information;
[0068] The keyword occupancy ratio is the proportion of conference keywords and synonyms of conference keywords in the conference keyword group relative to the keywords in the document information;
[0069] Among them, the method for determining the complete mapping ratio is to mark the conference keywords and synonyms of the conference keywords that can be found mapped in the same document information, and determine the ratio of the number of marked keywords in the conference keyword group to the total number, and recognize the ratio as the complete mapping ratio. Among them, if the synonyms of the conference keywords are marked, the corresponding conference keywords in the conference keyword group are also marked.
[0070] In some embodiments disclosed in the present invention, the expression for calculating the relevance of the conference keyword group is:
[0071] .
[0072] in, For the degree of relevance, is the fractional expansion adjustment coefficient, n is the number of document information, is the expression feature analysis function of the i-th document information, if is greater than or equal to the preset value, then Output 1, otherwise output 0. is the proportion of keywords in the i-th document information, is the occupancy ratio adjustment constant, is the complete mapping ratio of the i-th document information, and b is the relevance adjustment constant.
[0073] Among them, the expression of keyword occupancy ratio is:
[0074] .
[0075] in, is the total number of conference keywords and their synonyms in the i-th document information, is the number of all keywords in the i-th document information.
[0076] Among them, the expression of the complete mapping ratio is:
[0077] .
[0078] in, is the number of conference keywords marked in the conference keyword group, is the number of all conference keywords in the conference keyword group, and the marked conference keywords are conference keywords that can find mapping keywords or synonyms in the document information.
[0079] Fractional amplification adjustment coefficient: The determination of the fractional amplification adjustment coefficient is mainly based on the system's overall amplification or reduction requirements for the calculation of the relevance of the conference keyword group. This coefficient is used to adjust the calculation results of the relevance to better adapt to different application scenarios and needs.
[0080] Occupancy ratio adjustment constant: The determination of the occupancy ratio adjustment constant is mainly based on the importance of the keyword in the document information; this constant is used to adjust the calculation of the keyword occupancy ratio to ensure that the importance of the keyword is reasonably reflected.
[0081] Relevance adjustment constant: When calculating the relevance of conference keyword groups, the complete mapping ratio is an important indicator. However, in some cases, even if the complete mapping ratio is high, the relevance may still be low due to other factors (such as the keyword occupancy ratio). By adjusting the relevance adjustment constant, you can ensure that the complete mapping ratio is properly weighted when calculating the relevance, thereby more accurately reflecting the relevance between the conference keyword group and the document information.
[0082] Step S300, searching the reference conference keyword group on the Internet and in-house database, and based on information value analysis, intercepting information fragments to obtain a number of information fragments, and recording the collection of information fragments as an information fragment set.
[0083] In this step, the system will conduct a more comprehensive web and in-house database search for conference keyword phrases marked as references; the goal of this search is to find detailed information fragments, which can be articles, reports, research or any valuable content related to the conference topic; the system will use information value analysis to evaluate the value of the retrieved information fragments, which may include factors such as accuracy, timeliness, completeness and relevance of the content; valuable information fragments will be extracted and combined into an information fragment set; this collection will provide participants with additional information resources during the conference.
[0084] In some embodiments disclosed in the present invention, based on information value analysis, the method for intercepting information fragments includes:
[0085] Step S301, analyze the retrieved information content in terms of content relevance, timeliness, source credibility, number of citations and information association, and configure weight coefficients that require attention for content relevance, timeliness, source credibility, number of citations and information association respectively, and construct content-related value conversion operators, timeliness value conversion operators, source credibility value conversion operators, number of citations value conversion operators and information association value conversion operators respectively, and obtain the information value through comprehensive calculation.
[0086] Step S302, real-time evaluation of the information value of the fragment in the information content is performed. If the information value is greater than or equal to a preset value, the complete part of the fragment of the information content is intercepted to obtain the information fragment.
[0087] The expression for calculating information value is:
[0088] .
[0089] in, is the information value, is the first weight coefficient that needs attention, is the content relevance parameter of the information content, is the second weight coefficient that needs attention, is a parameter related to the timeliness of information content, is the third weight coefficient that needs attention, is a parameter related to source credibility, is the fourth weight coefficient that needs attention, is the parameter related to the number of citations, is the fifth weight coefficient that needs attention, It is information-related parameter.
[0090] Among them, the method for determining the parameters related to content relevance may include: keyword matching, calculating the matching degree between the keywords in the information fragment and the conference theme, which is achieved by calculating the matching degree between the keywords in the information fragment and the conference keyword group. Topic consistency, analyzing the consistency between the theme of the information fragment and the conference theme, which is achieved through a topic model (such as LDA), and comparing the topic distribution of the information fragment with the conference topic distribution. Context relevance, evaluating the degree of relevance between the context information in the information fragment and the conference content, which is achieved by analyzing the phrases and sentence structures in the information fragment, and the semantic relationship with the conference keyword group.
[0091] Among them, the method for determining timeliness-related parameters may include: release date, the release date of the information fragment, the closer the date is, the newer the information is; update frequency, the frequency of updating the information fragment, the frequently updated information fragment has higher timeliness; timeliness score, assigning a timeliness score to the information fragment based on the release date and update frequency of the information fragment, as well as comparison with other information fragments.
[0092] Among them, methods for determining credibility-related parameters may include: source credibility, analyzing the credibility of the source of the information fragment, such as news websites, academic journals, professional blogs, etc.; author authority: evaluating the authority of the author of the information fragment, such as the author's academic background, professional experience, etc.; source scoring, assigning a credibility score to the source of the information fragment based on the credibility of the source and the authority of the author.
[0093] Among them, the method for determining the parameters related to the number of citations may include: the number of citations, which is the number of citations of the information fragment on the Internet. The more citations, the higher the value of the information fragment; citation distribution, which analyzes the distribution of citations of the information fragment, such as the citation distribution of different sources, the citation distribution of different fields, etc.; citation score, which assigns a citation score to the citation of the information fragment based on the number of citations and citation distribution.
[0094] Among them, the method for determining information-related parameters may include: information relevance, analyzing the degree of relevance between information fragments and other information fragments, such as subject relevance, content relevance, etc.; information clustering, organizing information fragments with high relevance into information clusters, and the information fragments within the information cluster have higher value; information network, constructing an association network between information fragments, and analyzing the position and role of information fragments in the network; association scoring, assigning an association score to the association of information fragments based on information relevance and the formation of information clusters.
[0095] Weight coefficient: Assigning different weight coefficients to different factors or indicators can quantify their importance in the overall evaluation or calculation; this enables a more accurate reflection of the actual impact of each factor or indicator on the final result; the weight coefficient can be flexibly adjusted according to the actual application scenario and needs; this means that different weights can be assigned to different factors or indicators according to different situations or goals to adapt to different evaluation or calculation requirements.
[0096] First of all, it should be made clear that the expression of the present invention is for the purpose of protecting the relational logic of calculating the value of information. The significance of quantifying different reference factors in the information value expression is that quantification can make the originally vague and subjective evaluation objective and measurable; through specific numerical values, the performance of information fragments in various aspects, such as content relevance, timeliness, credibility, etc., can be more accurately evaluated, and through the adjustment of weight coefficients, the influence and contribution of different relevant parameters on the information value can be aligned.
[0097] (1) Quantification of content relevance parameters:
[0098] The content relevance parameter mainly evaluates the degree of association between the information fragment and the conference theme. When quantifying this parameter, the keyword matching degree can be considered, that is, calculating the matching degree between the keywords in the information fragment and the conference theme keyword group. The higher the matching degree, the stronger the content relevance. At the same time, the consistency between the theme of the information fragment and the conference theme can be analyzed. The topic distribution of the information fragment and the conference theme distribution can be compared through a topic model (such as LDA). The closer the topic, the higher the content relevance. In addition, context relevance is also an important consideration. The strength of content relevance can be determined by evaluating the degree of association between the contextual information in the information fragment and the conference content, such as analyzing the semantic relationship between phrases, sentence structures and conference keyword groups.
[0099] (2) Quantification of timeliness-related parameters:
[0100] The timeliness parameter is used to measure the freshness of an information fragment. When quantifying, the release date of the information fragment can be considered. The closer the date is, the newer the information is and the more time-effective it is. At the same time, the update frequency of the information fragment is also an important indicator. Information fragments that are updated frequently have higher timeliness. In order to comprehensively evaluate timeliness, a timeliness score can be assigned to the information fragment, which is determined based on the release date, update frequency, and comparison with other information fragments.
[0101] (3) Quantification of credibility-related parameters:
[0102] Credibility parameters are used to evaluate the reliability of the source and author of the information fragment. When quantifying, first analyze the credibility of the source of the information fragment, such as news websites, academic journals, professional blogs, etc. Different sources have different credibility. Secondly, evaluate the authority of the author of the information fragment, such as the author's academic background, professional experience, etc. The higher the authority, the stronger the credibility. Finally, according to the credibility of the source and the authority of the author, assign a credibility score to the source of the information fragment to comprehensively reflect its credibility.
[0103] (4) Quantification of citation-related parameters:
[0104] The citation count parameter is used to measure the citation status of information fragments on the Internet. When quantifying, first count the citation count of the information fragment. The more citation counts, the higher the value of the information fragment. Secondly, analyze the distribution of citations of the information fragment, such as the distribution of citations from different sources, the distribution of citations in different fields, etc. The wider the citation distribution, the greater the influence of the information fragment. Finally, according to the citation count and citation distribution, assign a citation score to the citation status of the information fragment to comprehensively reflect its citation value and influence.
[0105] (5) Quantification of information relevance related parameters:
[0106] The information relevance parameter is used to evaluate the degree of relevance between information fragments and other information fragments. When quantifying, first analyze the relevance properties between information fragments and other information fragments, such as subject relevance, content relevance, etc. The closer the relevance properties, the higher the information relevance. Secondly, information fragments with high relevance can be grouped into information clusters, and information fragments within information clusters have higher value. In order to more comprehensively evaluate the information relevance, it is also possible to construct a relevance network between information fragments and analyze the position and role of information fragments in the network. Finally, according to the information relevance and the formation of information clusters, a relevance score is assigned to the relevance of the information fragments to comprehensively reflect its importance and value in the information network.
[0107] Step S400, real-time voice in the video conference is collected in real time, and the real-time voice is converted into real-time quotation information using voice recognition technology, and semantic analysis is performed on the real-time quotation information to extract a number of real-time keywords, and based on the occurrence frequency of the real-time keywords and the order correlation characteristics of their mutual occurrence, the information fragments in the information fragment set are screened, and the screened information fragments are displayed as explanatory recommendation materials.
[0108] In this step, the system starts to collect voice data from the video conference in real time; these voice data will be converted into real-time quotation information in text form through voice recognition technology; voice recognition technology can identify different speakers and languages, and accurately convert voice into text; then, the system will perform semantic analysis on these real-time quotation information and extract real-time keywords, which reflect the instant discussion points in the meeting; the system will filter the information fragments in the previously constructed information fragment set according to the frequency of occurrence of real-time keywords and the order correlation characteristics between them, and find the fragments most relevant to the current discussion. Finally, these filtered information fragments will be displayed to the meeting participants as explanatory recommendations to help them better understand the discussion content and provide decision support or supplementary information.
[0109] In some embodiments disclosed in the present invention, based on the occurrence frequency of real-time keywords and the order correlation characteristics of their occurrence, the method for screening information fragments in the information fragment set includes:
[0110] Step S401 , semantic analysis is performed on the information segments in the information segment set to determine the comparison keywords in the information segments, and the occurrence frequency and the order of occurrence correlation features of the comparison keywords are determined.
[0111] Step S402, compare the occurrence frequency and occurrence order association features of the comparison keywords with the real-time keywords, and determine the degree of consistency between the information fragment and the real-time quotation information based on the comparison result, and give priority to displaying the information fragment based on the size relationship of the consistency degree.
[0112] In some embodiments disclosed in the present invention, the method of comparing the occurrence frequency and occurrence order correlation features of the comparison keywords with the real-time keywords includes:
[0113] Step S4011, constructing a plane text grid template, wherein the plane text grid is a plurality of text grids arranged in a uniform array.
[0114] Step S4012, fill the real-time quotation information into the plane text grid template, mark the real-time keywords therein, and based on the semantic connection relationship between different real-time keywords, connect the text grid groups corresponding to the real-time keywords to obtain the real-time quotation text grid expression diagram.
[0115] Step S4013, fill the information fragment into the plane text grid template, mark the comparison keywords therein, and based on the semantic connection relationship between different comparison keywords, connect the text grid groups corresponding to the comparison keywords to obtain the information fragment text grid expression diagram.
[0116] Step S4014, sequentially shift and adjust all the text grids in the real-time quotation text grid expression diagram and the information fragment text grid expression diagram, and compare the real-time quotation text grid expression diagram and the information fragment text grid expression diagram, and determine the degree of consistency between the information fragment and the real-time quotation information based on the comparison result.
[0117] In some embodiments disclosed in the present invention, the expression for calculating the degree of conformity is:
[0118] .
[0119] in, For the degree of relevance, The conformity adjustment factor, The number of matching keyword pairs in the real-time quotation text expression graph and the information fragment text expression graph. is the number of real-time keywords in the real-time quotation text expression graph, The number of keywords compared in the information fragment text grid expression diagram, is the connection quantity adjustment factor, is the number of matching lines between text grid groups, is the number of lines in the real-time quotation text expression graph, The information fragment text grid expresses the number of connections in the graph, and d is a constant for adjusting the number of connections.
[0120] Conformance adjustment coefficient: This coefficient is used to adjust the calculation result of the conformance. It may be an empirical value or a value set according to the actual application scenario to ensure the accuracy and rationality of the calculation result. By adjusting this coefficient, the conformance can be fine-tuned in different situations to meet different application requirements.
[0121] Connection Quantity Adjustment Factor: This factor is used to adjust the impact of the connection quantity when calculating the match degree. It can help balance the relationship between the number of keyword matches and the number of connections, ensuring that the number of connections does not overly affect the results when calculating the total match degree.
[0122] Wire Quantity Adjustment Constant: This constant is used to further adjust the influence of the wire quantity when calculating the degree of compliance. It can be a fixed value used to fine-tune the wire quantity in specific situations to ensure the stability and consistency of the calculation results.
[0123] Among them, the method for judging the matching keyword pair is to set range circles for both the real-time keyword and the comparison keyword. If the range circles of the two keywords overlap, the two keywords are judged to be the matching keyword pair.
[0124] In some embodiments disclosed in the present invention, there is also disclosed a system for generating interpretation of element information in a video conference, including:
[0125] The first module is used to collect the meeting draft information of different meeting participants, and perform semantic analysis on each meeting draft information to determine a number of meeting keywords, and freely combine a number of meeting keywords belonging to the same meeting draft information to obtain a number of meeting keyword groups;
[0126] The second module is used to perform a network search for the conference keyword group, and based on the search results, determine the relevance of the conference keyword group, and record a number of conference keyword groups with a relevance greater than or equal to a preset value as reference conference keyword groups;
[0127] The third module is used to search the reference conference keyword phrases on the Internet and in a self-used database, and based on the information value analysis, to intercept the information fragments to obtain a number of information fragments, and record the collection of information fragments as an information fragment set;
[0128] The fourth module is used to collect real-time voice in video conferences in real time, and use voice recognition technology to convert real-time voice into real-time quotation information, and perform semantic analysis on the real-time quotation information to extract several real-time keywords. Based on the occurrence frequency of real-time keywords and the order correlation characteristics of their mutual occurrence, the information fragments in the information fragment set are screened, and the screened information fragments are displayed as explanatory recommendation materials.
[0129] The present invention discloses a method and system for generating interpretation of element information in an applied video conference, and relates to the field of video conference. Specifically, the present invention discloses freely combining a number of conference keywords belonging to the same conference draft information to obtain a number of conference keyword groups, performing a network search on the conference keyword groups, and recording a number of conference keyword groups with a correlation degree greater than or equal to a preset value as reference conference keyword groups, performing a network and a self-use database search on the reference conference keyword groups to obtain a number of information fragments, and recording the collection of information fragments as an information fragment set, screening the information fragments in the information fragment set based on the occurrence frequency of real-time keywords and the order correlation characteristics of their occurrence, and displaying the screened information fragments as explanatory recommendation materials. The present invention realizes timely interpretation of professional elements in professional video conferences through the above technical scheme, and improves the participants' cognition and understanding of the conference.
[0130] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.
Claims
1. A method for generating and interpreting element information in a video conference, characterized in that: include: Collecting the meeting draft information of different meeting participants, and performing semantic analysis on each meeting draft information, determining a number of meeting keywords, and freely combining a number of meeting keywords belonging to the same meeting draft information to obtain a number of meeting keyword groups; Conducting a network search for the conference keyword group, and determining the relevance of the conference keyword group based on the search results, and recording a number of conference keyword groups whose relevance is greater than or equal to a preset value as reference conference keyword groups; Search the reference conference keyword phrases on the Internet and in-house databases, and based on information value analysis, extract information fragments to obtain several information fragments, and record the collection of information fragments as an information fragment set; Real-time voice in the video conference is collected in real time, and the real-time voice is converted into real-time quotation information by using voice recognition technology, and the real-time quotation information is semantically analyzed to extract a number of real-time keywords, and based on the occurrence frequency of the real-time keywords and the order of their occurrence, the information fragments in the information fragment set are screened, and the screened information fragments are displayed as explanatory recommendation materials; Based on information value analysis, the methods for intercepting information fragments include: The retrieved information content is analyzed in terms of content relevance, timeliness, source credibility, number of citations, and information relevance, and weight coefficients that need attention are configured for content relevance, timeliness, source credibility, number of citations, and information relevance, and content-related value conversion operators, timeliness value conversion operators, source credibility value conversion operators, number of citations value conversion operators, and information relevance value conversion operators are constructed respectively, and the information value is obtained through comprehensive calculation; Real-time evaluation of the information value of the fragments in the information content is performed. If the information value is greater than or equal to a preset value, the complete part of the fragment of the information content is intercepted to obtain the information fragment; The expression for calculating information value is: ; in, is the information value, is the first weight coefficient that needs attention, is the content relevance parameter of the information content, is the second weight coefficient that needs attention, is a parameter related to the timeliness of information content, is the third weight coefficient that needs attention, is a parameter related to source credibility, is the fourth weight coefficient that needs attention, is the parameter related to the number of citations, is the fifth weight coefficient that needs attention, It is information-related parameter.
2. According to claim 1, a method for generating and interpreting element information in a video conference application, characterized in that: Methods for determining several conference keywords include: Acquire the conference theme, and set a number of reference keywords for the conference theme, and perform synonym analysis on the determined conference keywords based on the reference keywords, and filter the keywords in the conference draft information based on the result of the synonym analysis, and divide the keywords into focus keywords and general keywords based on the filtering result; Based on the semantic relationship between the focus keywords and the regular keywords in the sentence, the regular keywords are screened, and the screened regular keywords and focus keywords are simultaneously identified as conference keywords.
3. The method for generating and interpreting element information in a video conference according to claim 1, characterized in that: Methods for determining the relevance of a conference keyword group include: Construct query statements based on conference keyword groups, submit the query statements to search engines and academic databases, and select the retrieved document information according to the preset demand volume; Performing semantic analysis on the selected document information, including determining expression features of conference keywords in the conference keyword group relative to the document information, and determining expression features of synonyms of the conference keywords relative to the document information; Based on the first expression feature and the second expression feature, the relevance of the conference keyword group is determined.
4. A method for generating and interpreting element information in a video conference according to claim 3, characterized in that: The expression features include the proportion of keywords in each document information and the complete mapping ratio of conference keyword groups to document information; The keyword occupancy ratio is the proportion of conference keywords and synonyms of conference keywords in the conference keyword group relative to the keywords in the document information; Among them, the method for determining the complete mapping ratio is to mark the conference keywords and synonyms of the conference keywords that can be found mapped in the same document information, and determine the ratio of the number of marked keywords in the conference keyword group to the total number, and recognize the ratio as the complete mapping ratio. Among them, if the synonyms of the conference keywords are marked, the corresponding conference keywords in the conference keyword group are also marked.
5. The method for generating and interpreting element information in a video conference according to claim 3, characterized in that: The expression for calculating the relevance of the conference keyword group is: ; in, For the degree of relevance, is the fractional expansion adjustment coefficient, n is the number of document information, is the expression feature analysis function of the i-th document information, if is greater than or equal to the preset value, then Output 1, otherwise output 0. is the proportion of keywords in the i-th document information, is the occupancy ratio adjustment constant, is the complete mapping ratio of the i-th document information, and b is the relevance adjustment constant; Among them, the expression of keyword occupancy ratio is: ; in, is the total number of conference keywords and their synonyms in the i-th document information, is the number of all keywords in the i-th document information; Among them, the expression of the complete mapping ratio is: ; in, is the number of marked conference keywords in the conference keyword group, is the number of all conference keywords in the conference keyword group, and the marked conference keywords are conference keywords that can find mapping keywords or synonyms in the document information.
6. The method for generating and interpreting element information in a video conference according to claim 1, characterized in that: Based on the occurrence frequency of real-time keywords and the order correlation characteristics of their occurrence, the method for screening the information fragments in the information fragment set includes: Performing semantic analysis on the information fragments in the information fragment set, determining the comparison keywords in the information fragments, and determining the occurrence frequency and the order of occurrence correlation features of the comparison keywords; The occurrence frequency and order-related features of the comparison keywords are compared with the real-time keywords, and based on the comparison results, the degree of consistency between the information fragment and the real-time quotation information is determined, and based on the size relationship of the consistency, the information fragment is given priority display.
7. A method for generating and interpreting element information in a video conference according to claim 6, characterized in that: The method of comparing the occurrence frequency and occurrence order correlation features of the comparison keywords with the real-time keywords includes: A plane text grid template is constructed, wherein the plane text grid is a plurality of text grids arranged in a uniform array; Fill the real-time quotation information into the plane text grid template, mark the real-time keywords therein, and connect the text grid groups corresponding to the real-time keywords to obtain the real-time quotation text grid expression diagram based on the semantic connection relationship between different real-time keywords; Fill the information fragments into the flat text grid template, mark the comparison keywords therein, and connect the text grid groups corresponding to the comparison keywords based on the semantic connection relationship between different comparison keywords to obtain the information fragment text grid expression diagram; All the text grids in the real-time quotation text grid expression diagram and the information fragment text grid expression diagram are sequentially translated and adjusted, and the real-time quotation text grid expression diagram and the information fragment text grid expression diagram are compared, and based on the comparison result, the degree of consistency between the information fragment and the real-time quotation information is determined.
8. A method for generating and interpreting element information in a video conference according to claim 7, characterized in that: The expression for calculating the degree of match is: ; in, For the degree of relevance, The conformity adjustment factor, The number of matching keyword pairs in the real-time quotation text expression graph and the information fragment text expression graph. is the number of real-time keywords in the real-time quotation text expression graph, The number of keywords compared in the information fragment text grid expression diagram, is the connection quantity adjustment factor, is the number of matching lines between groups of text cells, is the number of lines in the real-time quotation text expression graph, The information fragment text grid expresses the number of connections in the graph, and d is the constant for adjusting the number of connections; Among them, the method for judging the matching keyword pair is to set range circles for both the real-time keyword and the comparison keyword. If the range circles of the two keywords overlap, the two keywords are judged to be the matching keyword pair.
9. A system for generating and explaining element information in a video conference, characterized in that: include: The first module is used to collect the meeting draft information of different meeting participants, and perform semantic analysis on each meeting draft information to determine a number of meeting keywords, and freely combine a number of meeting keywords belonging to the same meeting draft information to obtain a number of meeting keyword groups; The second module is used to perform a network search for the conference keyword group, and based on the search results, determine the relevance of the conference keyword group, and record a number of conference keyword groups with a relevance greater than or equal to a preset value as reference conference keyword groups; The third module is used to search the reference conference keyword phrases on the Internet and in a self-used database, and based on the information value analysis, to intercept the information fragments to obtain a number of information fragments, and record the collection of information fragments as an information fragment set; The fourth module is used to collect real-time voice in the video conference in real time, and use voice recognition technology to convert the real-time voice into real-time quotation information, and perform semantic analysis on the real-time quotation information to extract a number of real-time keywords, and based on the occurrence frequency of the real-time keywords and the order of occurrence correlation characteristics between each other, screen the information fragments in the information fragment set, and display the screened information fragments as explanatory recommendation materials; Based on information value analysis, the methods for intercepting information fragments include: The retrieved information content is analyzed in terms of content relevance, timeliness, source credibility, number of citations, and information relevance, and weight coefficients that need attention are configured for content relevance, timeliness, source credibility, number of citations, and information relevance, and content-related value conversion operators, timeliness value conversion operators, source credibility value conversion operators, number of citations value conversion operators, and information relevance value conversion operators are constructed respectively, and the information value is obtained through comprehensive calculation; Real-time evaluation of the information value of the fragments in the information content is performed. If the information value is greater than or equal to a preset value, the complete part of the fragment of the information content is intercepted to obtain the information fragment; The expression for calculating information value is: ; in, is the information value, is the first weight coefficient that needs attention, is the content relevance parameter of the information content, is the second weight coefficient that needs attention, is a parameter related to the timeliness of information content, is the third weight coefficient that needs attention, is a parameter related to source credibility, is the fourth weight coefficient that needs attention, is the parameter related to the number of citations, is the fifth weight coefficient that needs attention, It is information-related parameter.
