A high-efficiency video script creative generation method and system

Abstract

The application provides a high-efficiency video script creative generation method and system, and the application can make the finally generated script more original and personalized by identifying and retaining innovative features.The integration of multidimensional features, including content, structure, style, sound effects and narration, improves the coherence and diversity of the video script.Combined with natural language processing technology, the application can effectively extract and integrate innovative elements in the reference materials to generate high-quality creative scripts, solving the technical problem of the increasing demand for high-quality and innovative scripts in the prior art.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data recognition, in particular to a high-efficiency video script creative generation method and system. BACKGROUND

[0002] In the field of video script creation, with the vigorous development of the video content market, the demand for high-quality and innovative scripts is increasing. Traditional video script creation methods often rely on the personal experience and inspiration of the creator, which is low in efficiency and difficult to guarantee the novelty and diversity of the creativity. In the existing technology, there is a lack of systematic method to screen valuable reference content from a large number of materials, it is difficult to fully utilize the rich video resource library; in the aspect of shot processing, the classification, analysis and reorganization of shots are not fine enough, resulting in incoherent and optimized script logic; in the script similarity analysis, there is no mature and effective model to accurately screen unique script features; in the aspect of innovative element mining and integration, there is a lack of scientific means to identify and integrate novel elements, thereby limiting the innovation and appeal of the script. Therefore, an efficient video script creative generation method and system are needed to solve these problems. SUMMARY

[0003] In view of the above problems, the present application provides a high-efficiency video script creative generation method and system to solve the technical problems existing in the background art, and the technical solutions adopted are as follows:

[0004] A high-efficiency video script creative generation method, the method comprising:

[0005] S1: clearly defining the theme of the video to be created, searching in the video library according to the theme and emotional tags, screening a plurality of videos similar in theme and emotion, and obtaining the scripts of the plurality of videos similar in theme and emotion as an initial script set;

[0006] S2: classifying and labeling each shot in the initial script set according to its nature, length and angle, analyzing the logical relationship of each shot in the plurality of scripts, and reorganizing each shot in the plurality of scripts according to the analysis results to form a new video;

[0007] S3: obtaining the script of the new video, calculating the similarity of the new script and the initial script set based on the structural, style and sound effect features of the new video script through a script similarity calculation model, and screening scripts with low similarity, taking the scripts with low similarity as reference materials for the script features of the video to be created.

[0008] S4: performing natural language processing on the reference materials, screening innovative elements in the reference materials, integrating the innovative elements, and generating a script for the video to be created.

[0009] Preferably, the S1 comprises:

[0010] S11: explicitly determine the theme of the video to be created, and explicitly determine the description content of the video to be created, according to the description content, obtain the emotional main tone of the video to be created, and label the video to be created with an emotional label, the emotional label includes but is not limited to joy, sadness, anger, surprise and fear;

[0011] S12: according to the theme and emotional label of the video to be created, search the video library, filter out several videos with similar theme and emotion, and obtain the script from the filtered several videos, and unify into an initial script set.

[0012] Preferably, the S2 comprises:

[0013] S21: label each shot according to the nature of the shot, the nature includes but is not limited to action, dialogue and close-up;

[0014] S22: measure the length of each shot and label the length of each shot;

[0015] S23: record the shooting angle of each shot, the shooting angle includes front, side, back, top view and overhead view, and label the angle of each shot;

[0016] S24: obtain the logical relationship between each shot, the logical relationship includes but is not limited to causal relationship, time sequence and space conversion, analyze the logical relationship of multiple shots, construct the logical chain of each shot, and ensure the integrity of the logical chain;

[0017] S25: according to the analysis result of the logical relationship of multiple shots, recombine the shots to form a new video.

[0018] Preferably, the S3 comprises,

[0019] S31: for the script of the newly generated video, analyze the number sequence of the shots in the structural feature, analyze the style type of the video in the style feature, which is realistic, exaggerated and humorous, identify the volume and rhythm in the sound effect feature, and quantify all the features into data for analysis;

[0020] S32: input the new script feature data into the pre-constructed script similarity calculation model, calculate the similarity between the new script and each script in the initial script set, and obtain the calculation result, the higher the value corresponding to the calculation result, the higher the similarity, and vice versa.

[0021] S33: Compare the obtained numerical value corresponding to the calculation result with the preset similarity threshold value, filter and retain the script features lower than the similarity threshold value, and take the retained script features with low similarity as reference materials for the script features of the video to be created.

[0022] Preferably, the S4 comprises:

[0023] S41: Convert the reference materials into semantic vector representations through a natural language model, and identify the sentiment orientation in the text by applying a sentiment analysis algorithm, the sentiment emotion including positive, neutral and negative;

[0024] S42: Extract each word in the reference materials, and identify innovative elements different from conventional content through a novelty scoring mechanism;

[0025] S43: Aggregate the identified innovative elements, analyze the relevance and potential combination possibilities between each other according to the sentiment orientation, and fuse the innovative elements according to the analysis results to generate innovative element combinations;

[0026] S44: Generate the script of the video to be created according to the innovative element combinations, and proofread and optimize the script of the video to be created.

[0027] An efficient video script creative generation system, the system comprising:

[0028] The emotion matching and screening system: the theme of the video to be created is determined, the theme and the emotion label are searched in the video library, a plurality of videos with similar theme and emotion are screened out, and the scripts of the plurality of videos with similar theme and emotion are obtained and determined as an initial script set;

[0029] The shot multidimensional analysis and reorganization system: each shot in the initial script set is classified and labeled according to the nature, length and angle, the logical relationship of each shot in the plurality of scripts is analyzed, and each shot in the plurality of scripts is reorganized according to the analysis result to form a new video;

[0030] The multi-feature script similarity evaluation system: the script of the new video is obtained, the similarity between the new script and the initial script set is calculated based on the structural, style and sound effect features of the new video script through a script similarity calculation model, and the script with low similarity is screened out and taken as reference materials for the script features of the video to be created.

[0031] The innovative element fusion script generation system: the reference materials are subjected to natural language processing, the innovative elements in the reference materials are screened out, the innovative elements are fused, and the script of the video to be created is generated.

[0032] Preferably, the emotional matching screening system comprises:

[0033] The content emotional tone determination system: determines the theme of the video to be created, and determines the description content of the video to be created, obtains the emotional main tone of the video to be created according to the description content, and labels the video to be created with an emotion label, which includes but is not limited to joy, sadness, anger, surprise and fear;

[0034] The emotional label video retrieval system: retrieves the video library according to the theme and emotional label of the video to be created, screens out several videos with similar theme and emotion, and obtains scripts from the screened several videos to form an initial script set.

[0035] Preferably, the multi-dimensional analysis and reorganization system of the shot comprises:

[0036] The shot property recognition and labeling system: labels each shot according to its property, which includes but is not limited to action, dialogue and close-up;

[0037] The shot length measurement and labeling system: measures the length of each shot and labels the length of each shot;

[0038] The shot angle recording and labeling system: records the shooting angle of each shot, which includes front view, side view, back view, top view and overhead view, and labels the angle of each shot;

[0039] The shot logical relationship analysis system: obtains the logical relationship between each shot, which includes but is not limited to cause and effect relationship, time sequence and space conversion, analyzes the logical relationship of multiple shots, constructs the logical chain of each shot, and ensures the integrity of the logical chain;

[0040] The shot reorganization scheme development and execution system: reorganizes the shots according to the analysis results of the logical relationship of multiple shots to form a new video.

[0041] Preferably, the multi-feature script similarity evaluation system comprises:

[0042] The video script feature quantitative analysis system: analyzes the number and order of shots in structural features, analyzes the style type of the video in style features, which is realistic, exaggerated and humorous, identifies the volume and rhythm in sound effect features, and quantifies all the features into data for analysis;

[0043] The script similarity calculation execution system: input the new script feature data into the pre-constructed script similarity calculation model, calculate the similarity of the new script and each script in the initial script set, and obtain the calculation result. The higher the value corresponding to the calculation result is, the higher the similarity is, and vice versa.

[0044] The low-similarity script feature screening system: compare the value corresponding to the calculation result with the preset similarity threshold, screen and retain the script features lower than the similarity threshold, and retain the script features with low similarity as reference materials for the script features of the video to be created.

[0045] Preferably, the innovative element fusion script generation system comprises:

[0046] The sentiment analysis extraction system: converts the reference materials into semantic vector representation through a natural language model, and identifies the sentiment tendency in the text by applying a sentiment analysis algorithm. The sentiment emotion includes positive, neutral and negative.

[0047] The vocabulary novelty innovation identification system: extracts each vocabulary in the reference materials, and identifies innovative elements different from conventional content through a novelty scoring mechanism.

[0048] The sentiment-related innovation fusion system: aggregates the identified innovative elements, analyzes the relevance and potential combination possibilities between them according to the sentiment tendency, and fuses the innovative elements according to the analysis result to generate an innovative element combination.

[0049] The script generation and polishing system: generates the script of the video to be created according to the innovative element combination, and polishes and optimizes the script of the video to be created. The present application has the following beneficial effects: the script similarity calculation model accurately identifies and retains innovative features, making the finally generated script more original and personalized. The integration of multi-dimensional features enhances the coherence and richness of the video script by covering content, structure, style, sound effects and narration. At the same time, through natural language processing, the scheme can effectively obtain and fuse innovative elements in the reference materials, thereby generating high-quality creative scripts. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 The present application relates to a high-efficiency video script creative generation method. DETAILED DESCRIPTION

[0051] The preferred embodiments of the present application are described below in conjunction with the accompanying drawings, and it should be understood that the preferred embodiments described herein are only used to illustrate and explain the present application, and are not used to limit the present application.

[0052] One embodiment of the application is a high-efficiency video script creative generation method, characterized in that the method comprises:

[0053] A high-efficiency video script creative generation method, the method comprising:

[0054] S1: The theme of the video to be created is determined, and a search is performed in a video library according to the theme and emotional tags, a plurality of videos similar in theme and emotion are screened out, the scripts of the plurality of videos similar in theme and emotion are obtained, and the scripts are used as an initial script set;

[0055] S2: Each shot in the initial script set is classified and labeled according to properties, length, and angle, the logical relationship of each shot in the plurality of scripts is analyzed, and each shot in the plurality of scripts is reorganized according to the analysis result to form a new video;

[0056] S3: The script of the new video is obtained, the similarity of the new script to the initial script set is calculated through a script similarity calculation model based on the structural, style, and sound effect characteristics of the script of the new video, scripts with low similarity are screened out, and the scripts with low similarity are used as reference materials for the script characteristics of the video to be created.

[0057] S4: The reference materials are subjected to natural language processing, innovative elements in the reference materials are screened out, the innovative elements are fused, and the script of the video to be created is generated.

[0058] The working principle and effect of the above technical solution are: first, the theme of the video to be created is determined, combined with the emotional label, and the natural language processing technology and emotional analysis tool are used to identify the materials matching the target theme and emotion from a large number of videos. By obtaining the scripts of these videos, an initial script set is formed. Each shot in the initial script set is analyzed in detail, and classified and labeled based on factors such as the nature, length, and shooting angle of the shot. Computer vision technology and film language rules are used to understand the logical relationship between shots. By analyzing these logical relationships, the shots are recombined to create new video sequences, aiming to generate a video based on the original materials but with a completely new presentation. New video scripts are obtained from the recombined videos, and their characteristics are analyzed, including content, structure, style, sound effects, and narration. A script similarity calculation model is used to compare the new script with the initial script set to identify the differences between them. Special attention is paid to the features with low similarity, as these features often represent innovation and unique content, which are used as reference materials for the video script to be created. The remaining reference materials are subjected to natural language processing to screen out innovative elements. This process includes key phrase acquisition, emotional and semantic analysis, etc. After identifying elements with creativity and potential, they are integrated to create a comprehensive, coherent, and innovative new video script. The script similarity calculation model accurately identifies and retains innovative features, making the final generated script more original and personalized. The integration of multi-dimensional features, including content, structure, style, sound effects, and narration, enhances the coherence and richness of the video script. At the same time, through natural language processing technology, the scheme can effectively obtain and integrate innovative elements from reference materials, generating high-quality creative scripts.

[0059] One embodiment of the present application includes:

[0060] S11: The theme of the video to be created is determined, and the description content of the video to be created is determined. According to the description content, the main emotional tone of the video to be created is obtained, and the video to be created is labeled with an emotional label. The emotional label includes but is not limited to joy, sadness, anger, surprise, and fear.

[0061] S12: According to the theme and emotional label of the video to be created, the video library is searched, and a number of videos similar in theme and emotion are selected. Scripts are obtained from the selected videos, and an initial script set is formed.

[0062] The working principle and effects of the above technical solution are as follows: S1 mainly involves preliminary planning and preparation for video creation, ensuring that the video's theme, content, and emotion align with expectations. By clearly defining the video's theme and descriptive content, it's easier to grasp the overall direction and message to be conveyed. After clarifying the video's theme and emotional tags, this information can be used as keywords to search the video library. This allows finding videos with similar themes and emotional expressions to the video to be created as references. From the retrieved videos, several videos most similar to the theme and emotion of the video to be created are selected. Then, scripts are obtained from these videos. The scripts or storylines obtained from multiple videos are integrated and unified to form an initial script set. This script set serves as the framework or foundation for the video to be created. By searching and filtering similar videos in the video library and obtaining scripts, existing resources can be fully utilized, duplication of effort can be avoided, and work efficiency can be improved. Clearly defining the main emotional tone of the video and assigning emotional tags helps to better create atmosphere and evoke emotional resonance in the audience during the creation process, enhancing the video's appeal and impact. Extracting scripts and storylines from multiple similar videos can provide creators with diverse reference materials, helping them to create richer and more varied video works.

[0063] In one embodiment of the present invention, S2 includes:

[0064] S21: Label each shot according to its nature, including but not limited to action, dialogue, and close-up;

[0065] S22: Measure the length of each lens and mark the length of each lens;

[0066] S23: Record the shooting angle of each shot, including front, side, back, top, and bottom views, and label the angle of each shot;

[0067] S24: Obtain the logical relationship between each shot, including but not limited to causal relationship, temporal sequence and spatial transformation, and analyze the logical relationship between multiple shots, construct the logical chain of each shot, and ensure the integrity of the logical chain;

[0068] S25: Based on the analysis of the logical relationships between multiple shots, the shots are recombined to form a new video.

[0069] The working principle and effect of the above technical solution are as follows: first, each shot in the video is analyzed in detail to determine its nature, such as action, dialogue, close-up, etc. Then, according to the nature of the shot, it is labeled so that it can be quickly identified and utilized in the subsequent editing and synthesis process. By labeling the action, dialogue, and close-up characteristics of the shot, the content and function of the shot are accurately described. Using professional video editing software or tools, the length of each shot is accurately measured. The measurement results are labeled to accurately control the length of each shot during editing. This ensures the rhythm and smoothness of the video, avoiding overly long or short shots that affect the viewing experience of the audience. The shooting angle of each shot is analyzed to determine whether it is front, side, back, overhead, or overhead, etc. The shooting angle is labeled to allow selection of appropriate shots to present different visual effects as needed during editing. The logical relationship between each shot is analyzed in depth, such as cause and effect, time sequence, spatial transition, etc. According to the logical relationship, a logical chain of each shot is constructed to ensure the integrity and coherence of the logical chain. Based on the analysis results of the logical relationship of multiple shots, a detailed shot reorganization plan is developed. Using professional video editing software or tools, the shots are edited and synthesized according to the reorganization plan to form a new video work.

[0070] In one embodiment of the present application, the S3 comprises,

[0071] S31: For the script of the newly generated video, analyze the number of shots in terms of structural features, analyze the style of the video in terms of style features, and identify the volume and rhythm in terms of sound effects features, and quantify all the features into data for analysis;

[0072] S32: Input the new script feature data into the pre-built script similarity calculation model, calculate the similarity of the new script with each script in the initial script set, and obtain the calculation result. The higher the value corresponding to the calculation result, the higher the similarity, and vice versa.

[0073] S33: Compare the value corresponding to the obtained calculation result with the preset similarity threshold, filter and retain the script features below the similarity threshold, and use the retained script features with low similarity as reference materials for the script features of the video to be created.

[0074] Moreover, the similarity formula calculation method is as follows:

[0075] Define the new script as N and the initial script set as I. The similarity calculation function of the structural features is obtained by the following formula:

[0076]

[0077] wherein S(N, I) represents the similarity value of the structural feature of the structure feature, S N represents the structural feature quantization value of the new script N, S I represents the structural feature quantization value of the initial script set I, S max represents the maximum value in all script structural feature quantization values;

[0078] Similarly, the similarity value F(N, I) of the style feature and the similarity value A(N, I) of the sound effect feature are obtained, finally, the similarity calculation formula is obtained, and the similarity calculation formula is as follows:

[0079]

[0080] wherein Q(N, I) represents the similarity value of the new script and the initial script set;

[0081] And the closer the similarity value is to 1, the higher the similarity is, and vice versa.

[0082] The working principle and effects of the above technical solution are as follows: in the similarity value calculation formula of the structural feature, the style feature and the sound effect feature, the exponential function is used because it has a smooth transition characteristic and can well simulate the change trend of the similarity under different difference levels. The value range of the exponential function is naturally limited to (0, 1], which exactly meets the value range requirement of the similarity, that is, the similarity is a value between (0, 1], without the need for additional complex normalization operation. The change rate of the trigonometric function in a specific interval can be used to emphasize the sensitivity of the structural feature difference. It can highlight the significant influence of the similarity when the feature difference is large, making the similarity calculation more sensitive to large differences, which is a characteristic that other functions do not have. In the overall similarity value calculation formula, the geometric mean is used to integrate the three local similarities. The geometric mean has the characteristic of not overemphasizing or underemphasizing the influence of a certain dimension, so that the contribution of each dimension to the overall similarity is relatively balanced. For example, if a script has a high structural similarity but low style and sound effect similarities, the geometric mean will appropriately reduce the overall similarity, avoiding overestimating the overall similarity due to the advantage of a certain dimension. The nature of the geometric mean determines that only when the similarity of all dimensions is relatively high, the overall similarity will tend to 1. If the similarity of a certain dimension is very low, even if the similarity of other dimensions is very high, the overall similarity will be significantly affected and reduced, which is consistent with the actual situation of the script. Because a script needs to be highly similar to another script in many aspects such as structure, style and sound effect. The method for obtaining the structural feature quantitative value includes: obtaining the number of shots of the new script and the initial script set; the method for obtaining the style feature quantitative value includes: defining a specific vocabulary list related to each language style. For example, for the elegant and beautiful style, the words such as "splendid", "beautiful", "quiet" are included; for the humorous and witty style, the words such as "funny", "amusing", "humorous" are included. The frequency E of these specific words in the script dialogue and descriptive language is counted, the total number of characters in the script is calculated, and then divided by the total number of sentences to obtain the average sentence length L, and the new script N style feature value F N =a x E + b x L, a and b represent the word frequency weight and the sentence length weight respectively, and a + b = 1, and in the initial script set, F I =1; the method for obtaining the sound effect feature quantitative value includes: traversing the volume values of all sampling points in the new script, adding them up and dividing by the total number of sampling points to obtain the average volume. The maximum and minimum values of the volume are calculated, and the difference between the two is the volume variation amplitude; the average volume and the volume variation amplitude are normalized, According to the detected beats, the number of beats per minute is calculated, for example, D beats are detected in the time T (in minutes), then The sound effect characteristic value = c x the volume characteristic value + d x the rhythm characteristic value, c and d respectively represent the weights of the volume characteristic value and the rhythm characteristic value, and c + d = 1.

[0083] In one embodiment of the present application, the S4 comprises:

[0084] S41: converting the reference material into a semantic vector representation through a natural language model, and identifying the sentiment tendency in the text by applying a sentiment analysis algorithm, the sentiment emotion including positive, neutral and negative;

[0085] S42: extracting each word in the reference material, and identifying innovative elements different from the conventional content through a novelty scoring mechanism;

[0086] S43: aggregating the identified innovative elements, analyzing the correlation and potential combination possibility between each other according to the sentiment tendency analysis, and fusing the innovative elements according to the analysis result to generate an innovative element combination;

[0087] S44: generating a script of the to-be-created video according to the innovative element combination, and proofreading and optimizing the script of the to-be-created video.

[0088] The working principle and effect of the above technical solution are: the reference material is converted into semantic vector representation by a natural language model. These vectors capture the deep semantic information of the text, providing detailed content background. Using sentiment analysis algorithms, the sentiment orientation in the reference material is identified and quantified. Identifying whether the text expresses positive, negative, or neutral sentiment helps understand its impact and communication purpose. The words in the material are segmented and analyzed, and a novelty scoring mechanism is used to identify innovative and creative elements in them. The identified innovative elements are aggregated, and the relevance between these elements is evaluated through sentiment orientation analysis. Based on the relevance analysis results, these innovative elements are strategically integrated to generate an organic combination of innovative elements. Based on the integrated combination of innovative elements, a video script to be created is generated. The generation process uses an automated script arrangement system to improve efficiency and creative consistency. The generated script is proofread and optimized, and multiple iterations of evaluation are performed through a language model to ensure the coherence, sentiment expression accuracy, and style consistency of the text. At the same time, semantic vector representation can capture the deep meaning of the text, making the model more accurately understand the content of the reference material. The sentiment analysis algorithm can identify the sentiment orientation in the text, providing an emotional tone for script creation, making the content more in line with the emotional needs of the audience. The novelty scoring mechanism can filter out unique words and expressions, injecting freshness and creativity into the video script. By identifying innovative elements, the video script can be distinguished from other similar works in terms of content, enhancing its appeal. Through sentiment orientation analysis, the consistency of the innovative elements in the combination can be ensured, enhancing the coherence and appeal of the content. The integration of innovative elements can generate new creative points, making the video script more rich and interesting in terms of content. The script generated based on the combination of innovative elements has unique content that can attract the attention of the audience. The proofreading and optimization process can ensure that the grammar, logic, and expression of the script are in the best state, improving the overall quality of the video.

[0089] One embodiment of the present application is a high-efficiency video script creation system, characterized in that the system comprises:

[0090] Emotion matching screening system: determine the theme of the video to be created, retrieve the video library according to the theme of the video to be created and the emotion label, screen out several videos with similar themes and emotions, and obtain the scripts of the several videos with similar themes and emotions as the initial script set;

[0091] Multi-dimensional analysis and reorganization system of shots: classify and label each shot in the initial script set according to its nature, length, and angle, analyze the logical relationship of each shot in several scripts, and reorganize each shot in several scripts according to the analysis results to form a new video;

[0092] The script similarity calculation system: obtains the script of the new video, and based on the script of the new video, the script similarity calculation model is used to calculate the similarity between the new script and the initial script set based on the characteristics of the new script, including but not limited to content, structure, style, sound effects and narration, and the script features with low similarity are retained, and the retained script features with low similarity are used as reference materials for the script features of the video to be created;

[0093] The script generation system: the reference materials are subjected to natural language processing, the innovative elements in the reference materials are screened out, and the innovative elements are fused to generate the script of the video to be created.

[0094] The working principle and effect of the above technical solution are: first, the theme of the video to be created is determined, and the emotion label is combined to identify the materials matching the target theme and emotion from a large number of videos through natural language processing technology and emotion analysis tools. The script of these videos is obtained to form an initial script set. Each shot in the initial script set is analyzed in detail, and is classified and labeled based on the properties, length and shooting angle of the shot. Computer vision technology and film language rules are used to understand the logical relationship between the shots. By analyzing these logical relationships, the shots are recombined to create new video sequences, aiming to generate a video based on the original materials but with a new presentation. The new video script is obtained from the recombined video, and the characteristics of the new video script are analyzed, including content, structure, style, sound effects and narration. The script similarity calculation model is used to compare the new script with the initial script set to identify the differences therebetween. Special attention is paid to the features with low similarity, because these features often represent innovation and uniqueness, and they are used as reference materials for the script of the video to be created. The retained reference materials are subjected to natural language processing to screen out innovative elements. This process includes key phrase acquisition, emotion and semantic analysis, etc. After identifying the elements with creativity and potential, they are fused to create a comprehensive, coherent and innovative new video script. The script similarity calculation model is used to accurately identify and retain innovative features, so that the finally generated script is more original and personalized. The integration of multi-dimensional features enhances the coherence and richness of the video script, including content, structure, style, sound effects and narration. At the same time, through the natural language processing technology, the scheme can effectively acquire and fuse the innovative elements in the reference materials to generate high-quality creative scripts.

[0095] In an embodiment of the present application, the emotion matching screening system comprises:

[0096] The content emotional tone determination system: the theme of the video to be created is determined, and the description content of the video to be created is determined, the mood main tone of the video to be created is obtained according to the description content, and the mood label is given to the video to be created, and the mood label includes but is not limited to joy, sadness, anger, surprise and fear;

[0097] The emotional label video retrieval system: according to the theme and emotional label of the video to be created, the video library is retrieved, a plurality of videos similar in theme and emotion are screened out, and the scripts are obtained from the plurality of videos, and an initial script set is formed.

[0098] The working principle and effect of the above technical solution are as follows: S1 is mainly for the planning and preparation of video creation, to ensure that the theme, content and emotion of the video are consistent with the expectation. By determining the theme and description content of the video, the overall direction and information to be conveyed of the video can be easily grasped, and after determining the theme and emotional label of the video, these information can be used as keywords for retrieval in the video library. In this way, videos similar in theme and emotional expression to the video to be created can be found as references. A plurality of videos most similar in theme and emotion to the video to be created are screened out from the retrieved videos. Then, the scripts are obtained from these videos. The scripts or storylines obtained from multiple videos are integrated and unified to form an initial script set. This script set is the framework or basis of the video to be created. By retrieving and screening similar videos in the video library and obtaining scripts, existing resources can be fully utilized, and repetitive work can be avoided, improving work efficiency. Determining the emotional main tone of the video and giving the emotional label can help better create an atmosphere and evoke the emotional resonance of the audience during the creation process, improving the attractiveness and infectivity of the video. Obtaining scripts and storylines from multiple similar videos can provide creators with diverse reference materials, which helps to create more rich and diverse video works.

[0099] In an embodiment of the present application, the multi-dimensional analysis and reorganization system of the shot includes:

[0100] The shot property recognition and labeling system: according to the property of the shot, the property includes but is not limited to action, dialogue and close-up, each shot is labeled;

[0101] The shot length measurement and labeling system: the length of each shot is measured, and the length of each shot is labeled;

[0102] The shot angle recording and labeling system: records the shooting angle of each shot, the shooting angle includes front, side, back, top view and overhead view, and labels the angle of each shot;

[0103] A shot logical relationship analysis system: obtain the logical relationship between each shot, including but not limited to, causality, time sequence, and spatial conversion, analyze the logical relationship of multiple shots, construct the logical chain of each shot, and ensure the integrity of the logical chain;

[0104] A shot reorganization plan formulation and execution system: reorganize the shots according to the analysis results of the logical relationship of multiple shots to form a new video.

[0105] The working principle and effect of the above technical solution are as follows: first, each shot in the video is analyzed in detail to determine its nature, such as action, dialogue, close-up, etc. Then, according to the nature of the shot, it is labeled so that it can be quickly identified and utilized in the subsequent editing and synthesis process. By labeling the action, dialogue, and close-up characteristics of the shot, the content and function of the shot are accurately described. Using professional video editing software or tools, the length of each shot is accurately measured. The measurement results are labeled to accurately control the length of each shot during editing. This ensures the rhythm and smoothness of the video, avoiding overly long or short shots that affect the viewing experience of the audience. The shooting angle of each shot is analyzed to determine whether it is front, side, back, overhead, or overhead, etc. The shooting angle is labeled so that appropriate shots can be selected to present different visual effects during editing. The logical relationship between each shot is analyzed in depth, such as causality, time sequence, spatial conversion, etc. According to the logical relationship, the logical chain of each shot is constructed to ensure the integrity and coherence of the logical chain. According to the analysis results of the logical relationship of multiple shots, a detailed shot reorganization plan is formulated. Using professional video editing software or tools, the shots are edited and synthesized according to the reorganization plan to form a new video work.

[0106] One embodiment of the present application, the innovative element fusion script generation system includes:

[0107] An emotion analysis extraction system: generates a semantic vector representation of the reference material through a natural language model, and identifies the emotional tendency in the text by applying an emotion analysis algorithm;

[0108] An innovative element recognition system: obtains each word in the reference material, and identifies innovative elements different from conventional content through a novelty scoring mechanism;

[0109] An emotion-related innovation fusion system: aggregates the identified innovative elements, analyzes their relevance and potential combination possibilities based on emotional tendency analysis, and fuses the innovative elements based on the analysis results to generate an innovative element combination;

[0110] The script generation and polishing system generates a script for the video to be created based on the combination of innovative elements and polishes the script for the video to be created.

[0111] The working principle and effects of the above technical solution are as follows: the reference material is converted into semantic vector representation by a natural language model. These vectors capture the deep semantic information of the text, providing detailed content background. A sentiment analysis algorithm is used to identify and quantify the sentiment orientation in the reference material. Identifying whether the text expresses positive, negative, or neutral sentiment helps understand its influence and communication purpose. Each word in the material is segmented and analyzed, and a novelty scoring mechanism is used to identify innovative and creative elements therein. The identified innovative elements are aggregated, and the correlation between these elements is evaluated based on sentiment orientation analysis. Based on the correlation analysis results, these innovative elements are strategically integrated to generate an organic combination of innovative elements. Based on the integrated combination of innovative elements, a script for the video to be created is generated. The generation process uses an automated script arrangement system to improve efficiency and creative consistency. The generated script is proofread and optimized, and the language model is iteratively evaluated to ensure the coherence of the text, the accuracy of the emotional expression, and the consistency of the style. At the same time, the semantic vector representation can capture the deep meaning of the text, making the model more accurately understand the content of the reference material. The sentiment analysis algorithm can identify the sentiment orientation in the text, providing an emotional tone for script creation, making the content more in line with the emotional needs of the audience. The novelty scoring mechanism can filter out unique words and expressions, injecting freshness and creativity into the video script. By identifying innovative elements, the video script can be distinguished from other similar works in terms of content, enhancing its appeal. Through sentiment orientation analysis, the innovative elements can maintain emotional consistency when combined, enhancing the coherence and appeal of the content. The integration of innovative elements can generate new creative points, making the video script more rich and interesting in terms of content. The script generated based on the combination of innovative elements has unique content that can attract the attention of the audience. The proofreading and optimization process can ensure that the script is grammatically, logically, and expressively optimal, improving the overall quality of the video.

[0112] Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application also intends to include these modifications and variations.

Claims

1. A method for generating creative video scripts, characterized in that, The method includes: S1: Define the theme of the video to be created, search the video library based on the theme and emotional tags, select several videos with similar themes and emotions, and obtain the scripts of the several videos with similar themes and emotions as the initial script set; S2: Classify and label each shot in the initial script set according to its nature, length, and angle. At the same time, analyze the logical relationship between each shot in several scripts. Based on the analysis results, reorganize each shot in several scripts to form a new video. S3: Obtain the script of the new video, and based on the structural, stylistic and sound effect features of the new video script, calculate the similarity between the new script and the initial script set through the script similarity calculation model, filter out the scripts with low similarity, and use the scripts with low similarity as reference materials for the features of the video script to be created. S4: Perform natural language processing on the reference materials, filter out the innovative elements in the reference materials, integrate the innovative elements, and generate the script for the video to be created.

2. The efficient video script creative generation method according to claim 1, characterized in that, S1 includes: S11: Define the theme of the video to be created, and define the description of the video to be created. Based on the description, obtain the main emotional tone of the video to be created, and assign an emotional tag to the video to be created. The emotional tags include joy, sadness, anger, surprise, and fear. S12: Search the video library based on the theme and emotion tags of the video to be created, select several videos with similar themes and emotions, and obtain scripts from the selected videos to form an initial script set.

3. The efficient video script creative generation method according to claim 1, characterized in that, S2 includes: S21: Label each shot according to its nature, which includes action, dialogue, and close-up; S22: Measure the length of each lens and mark the length of each lens; S23: Record the shooting angle of each shot, including front, side, back, top, and bottom views, and label the angle of each shot; S24: Obtain the logical relationship between each shot, including causal relationship, temporal sequence and spatial transformation, and analyze the logical relationship between multiple shots to construct the logical chain of each shot and ensure the integrity of the logical chain; S25: Based on the analysis of the logical relationships between multiple shots, the shots are recombined to form a new video.

4. The efficient video script creative generation method according to claim 1, characterized in that, S3 includes, Get S31: For the script of the newly generated video, in terms of structural features, analyze the number and order of shots; in terms of style features, analyze whether the video style is realistic, exaggerated, or humorous; in terms of sound effects features, identify volume and rhythm; and quantify all features into data for analysis. S32: Input the new script feature data into the pre-built script similarity calculation model, calculate the similarity between the new script and each script in the initial script set, and obtain the calculation result. The higher the value of the calculation result, the higher the similarity, and vice versa. S33: Compare the numerical values ​​corresponding to the obtained calculation results with the preset similarity threshold, filter and retain script features that are below the similarity threshold, and use the retained script features with low similarity as reference material for the script features of the video to be created.

5. The efficient video script creative generation method according to claim 1, characterized in that, S4 includes: S41: Using a natural language model, the reference material is transformed into a semantic vector representation, and the sentiment tendency in the text is identified by applying a sentiment analysis algorithm. The sentiment tendency includes positive, neutral, and negative. S42: Extract each word from the reference material and identify innovative elements that differ from conventional content through a novelty scoring mechanism; S43: Aggregate the identified innovative elements, analyze their correlation and potential combinations based on sentiment, and fuse the innovative elements based on the analysis results to generate a combination of innovative elements; S44: Generate the script for the video to be created based on the combination of innovative elements, and proofread and optimize the script for the video to be created.

6. A highly efficient video script creative generation system, characterized in that, The system includes: Emotional Matching and Filtering System: Define the theme of the video to be created, search the video library based on the theme and emotional tags, filter out several videos with similar themes and emotions, and obtain the scripts of the several videos with similar themes and emotions as the initial script set; The multi-dimensional shot analysis and reassembly system classifies and labels each shot in the initial script set according to its nature, length, and angle. It also analyzes the logical relationship between each shot in several scripts and reassembles each shot in several scripts based on the analysis results to form a new video. Multi-feature script similarity evaluation system: Obtain the script of a new video, and based on the structural, stylistic and sound effect features of the new video script, calculate the similarity between the new script and the initial script set through a script similarity calculation model, filter out scripts with low similarity, and use the scripts with low similarity as reference materials for the features of the video script to be created; Innovative Element Fusion Script Generation System: Performs natural language processing on the reference materials, filters out innovative elements from the reference materials, fuses the innovative elements, and generates a script for the video to be created.

7. The efficient video script creative generation system according to claim 6, characterized in that, The sentiment matching and screening system includes: Content Emotional Tone Determination System: Clearly define the theme of the video to be created, and at the same time, clearly define the description content of the video to be created. Based on the description content, obtain the main emotional tone of the video to be created, and assign an emotional tag to the video to be created. The emotional tags include joy, sadness, anger, surprise, and fear. Emotion Tag Video Retrieval System: Based on the theme and emotion tags of the video to be created, the system searches the video library, selects several videos with similar themes and emotions, and extracts scripts from the selected videos to form an initial script set.

8. The efficient video script creative generation system according to claim 6, characterized in that, The lens multi-dimensional analysis and reconstruction system includes: Shot type identification and labeling system: Each shot is labeled according to its type, which includes action, dialogue and close-up; Shot duration measurement and labeling system: Measure the length of each shot and label the length of each shot; Lens angle recording and annotation system: records the shooting angle of each shot, including front, side, back, top, and bottom views, and annotates the angle of each shot; Shot Logic Relationship Analysis System: Acquires the logical relationships between each shot, including causal relationships, temporal sequence, and spatial transformation, analyzes the logical relationships of multiple shots, constructs the logical chain of each shot, and ensures the integrity of the logical chain; Shot Reassembly Solution Development and Execution System: Based on the analysis of the logical relationships between multiple shots, the shots are reassembled to form a new video.

9. The efficient video script creative generation system according to claim 6, characterized in that, The multi-feature script similarity evaluation system includes: Video script feature quantification analysis system: For newly generated video scripts, in terms of structural features, it analyzes the number and order of shots; in terms of style features, it analyzes whether the video style is realistic, exaggerated, or humorous; in terms of sound effects features, it identifies volume and rhythm; and quantifies all features into data for analysis. The script similarity calculation execution system inputs the feature data of the new script into the pre-built script similarity calculation model, calculates the similarity between the new script and each script in the initial script set, and obtains the calculation results. The higher the value of the calculation result, the higher the similarity, and vice versa. Low-similarity script feature screening system: The system compares the calculated values ​​with a preset similarity threshold, filters and retains script features that are below the similarity threshold, and uses the retained low-similarity script features as reference material for the script features of the video to be created.

10. The efficient video script creative generation system according to claim 6, characterized in that, The innovative element fusion script generation system includes: Sentiment Analysis Extraction System: Through a natural language model, reference materials are transformed into semantic vector representations, and sentiment analysis algorithms are applied to identify the sentiment tendencies in the text, which include positive, neutral, and negative sentiment tendencies. Lexical Novelty Innovation Recognition System: Extracts each word from the reference material and identifies innovative elements that differ from conventional content through a novelty scoring mechanism; Emotional Association Innovation Fusion System: This system aggregates identified innovative elements, analyzes their correlation and potential combinations based on emotional tendencies, and fuses the innovative elements based on the analysis results to generate innovative element combinations. Script generation and refinement system: Based on the combination of innovative elements, it generates the script for the video to be created, and then proofreads and optimizes the script for the video to be created.

Images