Scenario quality evaluation method and device, electronic equipment and storage medium

Abstract

The invention provides a script quality evaluation method and device, electronic equipment and a storage medium. The method comprises the steps of obtaining text data of a to-be-evaluated script; based on the text data, calculating a contribution degree index of each role to plot development; based on the text data, calculating a support degree index of a relation network of each role for plot development; and performing matching degree analysis on the contribution degree index of each role and the corresponding support degree index to generate a script quality evaluation result. Therefore, the actual supporting effect of the role relation network on role shaping can be accurately reflected, and the accuracy of script analysis is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of script analysis, and in particular to a script quality evaluation method and device, an electronic device, and a storage medium. BACKGROUND

[0002] In modern film and television drama and stage play creation, character shaping and character relationship building are the core elements that determine the quality of a script. With the rapid development of the cultural industry, script evaluation has gradually shifted from purely subjective artistic judgment to a standardized process supported by quantitative analysis. Existing technical solutions mainly include three categories: the first category is a rule-based text analysis method that identifies and divides scenes through pre-set keywords to count character screen time; the second category uses machine learning algorithms to identify character interaction patterns in scripts through trained models; and the third category combines complex network theory to build character relationship graphs and calculate node centrality indicators.

[0003] However, the existing technical solutions have overly simple character relationship quantitative indicators that cannot accurately reflect the actual supporting role of relationships in character shaping, resulting in inaccurate script analysis results. SUMMARY

[0004] The purpose of the embodiments of the present application is to provide a script quality evaluation method and device, an electronic device, and a storage medium to solve the problem of overly simple character relationship quantitative indicators in existing technical solutions that cannot accurately reflect the actual supporting role of relationships in character shaping, resulting in inaccurate script analysis results. The specific technical solutions are as follows:

[0005] In a first aspect, the present application provides a script quality evaluation method, comprising:

[0006] obtaining text data of a script to be evaluated;

[0007] calculating a contribution degree indicator of each character to plot development based on the text data;

[0008] calculating a support degree indicator of a relationship network of each character to plot development based on the text data;

[0009] performing matching degree analysis on the contribution degree indicator of each character and the corresponding support degree indicator to generate a script quality evaluation result.

[0010] In one possible implementation, the calculation of the contribution degree indicator of each character to plot development comprises:

[0011] dividing the text data by episodes to obtain episode script data;

[0012] calculating a plot importance score of each character in the episode script data to obtain a single-episode contribution degree score corresponding to each character;

[0013] For each role, a contribution degree index of the role to the plot development is generated according to all single-episode contribution degree scores corresponding to the role.

[0014] In a possible implementation, the calculation of the plot importance scores of the roles in the script data to obtain single-episode contribution degree scores corresponding to the roles includes:

[0015] For each role, a set of play behaviors of the role is extracted from the script data, and the set of play behaviors includes a feature of appearance, a feature of dialogue, and a feature of behavior;

[0016] Based on the set of play behaviors, an influence degree of the role on a plot development of a current episode is analyzed, where the feature of appearance is used to calculate a role participation basis value, the feature of dialogue is used to analyze a plot driving force contribution value, and the feature of behavior is used to evaluate a conflict influence value.

[0017] The role participation basis value, the plot driving force contribution value, and the conflict influence value are weighted and comprehensively calculated to generate a single-episode contribution degree score of the role in the current episode.

[0018] In a possible implementation, the calculation of a support degree index of a relationship network of the roles to the plot development includes:

[0019] The text data is divided according to episodes to obtain script data of each episode;

[0020] For each set of script data, an interactive relationship between roles in the script data is identified, a contribution degree of the interactive relationship to a plot of a current episode is calculated, and a single-episode support degree score of each role in the current episode is calculated;

[0021] For each role, a support degree index of the role to the plot development is generated according to all single-episode support degree scores corresponding to the role.

[0022] In a possible implementation, the calculation of a single-episode support degree score of each role in a current episode based on a contribution degree of an interactive relationship to a plot of the current episode includes:

[0023] For each set of interactive relationships in the current episode, a contribution degree score of the interactive relationship to the plot of the current episode is calculated;

[0024] For each role, a target interactive relationship containing the role is identified;

[0025] Contribution degree scores corresponding to all target interactive relationships are summed to obtain a single-episode support degree score of the role in the current episode.

[0026] In a possible implementation, the calculating the contribution score of the interactive relationship to the current episode plot comprises:

[0027] extracting an interactive behavior feature set of the interactive relationship, the interactive behavior feature set comprising a dialogue interaction feature, a plot interaction feature, and a conflict interaction feature;

[0028] calculating a multi-dimensional contribution value corresponding to the interactive relationship based on the interactive behavior feature set, wherein the dialogue interaction feature is used to calculate a plot driving contribution value; the plot interaction feature is used to calculate a character shaping contribution value; and the conflict interaction feature is used to calculate a dramatic tension contribution value;

[0029] performing weighted comprehensive calculation on the plot driving contribution value, the character shaping contribution value, and the dramatic tension contribution value to generate the contribution score of the interactive relationship to the current episode plot.

[0030] In a possible implementation, the matching degree analysis of the contribution degree indicators of the roles and the corresponding support degree indicators comprises:

[0031] for each role, performing normalization processing on the contribution degree indicators corresponding to the role to obtain contribution degree data, and performing normalization processing on the support degree indicators corresponding to the role to obtain support degree data;

[0032] calculating a difference degree value between the contribution degree data and the support degree data, the difference degree value being used to represent a matching degree of role importance and relationship network support degree;

[0033] generating a script quality evaluation result based on the difference degree values of all the roles, the script quality evaluation result comprising a comprehensive score reflecting the matching situation of all the roles in the script, and / or an unbalanced role whose difference degree value exceeds a preset threshold and corresponding matching deviation features, and / or improvement suggestions generated according to the matching deviation features of the unbalanced roles.

[0034] In a second aspect, the present application provides a script quality evaluation device, comprising:

[0035] an acquisition module configured to acquire text data of a script to be evaluated;

[0036] a first calculation module configured to calculate, based on the text data, contribution degree indicators of roles to plot development;

[0037] a second calculation module configured to calculate, based on the text data, support degree indicators of relationship networks of the roles to plot development;

[0038] an analysis module configured to perform matching degree analysis of the contribution degree indicators of the roles and the corresponding support degree indicators to generate a script quality evaluation result.

[0039] In a possible implementation, the first calculation module is specifically configured to:

[0040] divide the text data according to episodes to obtain script data of each episode;

[0041] calculate, for the script data of each episode, a plot importance score of each character in the script data to obtain a single-episode contribution score corresponding to each character;

[0042] generate, for each character, a contribution index of the character to plot development according to all single-episode contribution scores corresponding to the character.

[0043] In a possible implementation, the first calculation module is further configured to:

[0044] extract, for each character, a set of acting behavior features of the character from the script data, the set of acting behavior features including a feature of appearance, a feature of dialogue, and a feature of behavior;

[0045] analyze, based on the set of acting behavior features, an influence degree of the character on a plot development of a current episode, where the feature of appearance is used to calculate a character participation basis value, the feature of dialogue is used to analyze a plot driving force contribution value, and the feature of behavior is used to evaluate a conflict impact value;

[0046] perform weighted comprehensive calculation on the character participation basis value, the plot driving force contribution value, and the conflict impact value to generate a single-episode contribution score of the character in the current episode.

[0047] In a possible implementation, the second calculation module is specifically configured to:

[0048] divide the text data according to episodes to obtain script data of each episode;

[0049] identify, for the script data of each episode, an interaction relationship between characters in the script data, calculate a single-episode support score of each character in a current episode based on a contribution degree of the interaction relationship to a plot of the current episode;

[0050] generate, for each character, a support index of the character to plot development according to all single-episode support scores corresponding to the character.

[0051] In a possible implementation, the second calculation module is further configured to:

[0052] calculate, for each group of interaction relationships in a current episode, a contribution score of the interaction relationship to a plot of the current episode;

[0053] identifying, for each role, all target interaction relationships containing the role;

[0054] summing up the contribution degree scores corresponding to all target interaction relationships to obtain a single-episode support degree score of the role in the current episode.

[0055] In one possible implementation, the second calculation module is further configured to:

[0056] extracting an interaction behavior feature set of the interaction relationship, the interaction behavior feature set including dialogue interaction features, plot interaction features, and conflict interaction features;

[0057] calculating, based on the interaction behavior feature set, a multi-dimensional contribution value corresponding to the interaction relationship, wherein the dialogue interaction features are used to calculate a plot driving contribution value; the plot interaction features are used to calculate a role shaping contribution value; and the conflict interaction features are used to calculate a dramatic tension contribution value;

[0058] performing weighted comprehensive calculation on the plot driving contribution value, the role shaping contribution value, and the dramatic tension contribution value to generate a contribution degree score of the interaction relationship to the plot of the current episode.

[0059] In one possible implementation, the analysis module is specifically configured to:

[0060] for each role, performing normalization processing on a contribution degree index corresponding to the role to obtain contribution degree data, and performing normalization processing on a support degree index corresponding to the role to obtain support degree data;

[0061] calculating a difference degree value between the contribution degree data and the support degree data, the difference degree value being used to represent a matching degree of role importance and relationship network support degree;

[0062] generating a script quality evaluation result based on the difference degree values of all roles, the script quality evaluation result including a comprehensive score reflecting a matching situation of all roles in the script, and / or, an unbalanced role whose difference degree value exceeds a preset threshold and a corresponding matching deviation feature, and / or, improvement suggestions generated according to the matching deviation features of the unbalanced roles.

[0063] In a third aspect, an electronic device is provided, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory complete communication with each other through the communication bus;

[0064] The memory is configured to store a computer program.

[0065] The processor is configured to execute the program stored on the memory to implement the method steps of any of the first aspect.

[0066] In a fourth aspect, a computer readable storage medium is provided, and the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the method steps in any of the first aspect.

[0067] In a fifth aspect, a computer program product containing instructions which, when executed on a computer, cause the computer to perform the script quality evaluation method in any of the above aspects.

[0068] The embodiments of the present application have the following beneficial effects:

[0069] The embodiments of the present application provide a script quality evaluation method and device, electronic equipment and a storage medium. In the present application, after obtaining script text data, the contribution degree index and the relationship network support degree index of each role are calculated according to the script text data, and then the correlation evaluation mechanism between the role and its relationship network is established through matching degree analysis, so that the actual support effect of the role relationship network on the role shaping can be accurately reflected, and the accuracy of script analysis is improved.

[0070] Of course, implementing any product or method of the present application does not necessarily require all the advantages described above. BRIEF DESCRIPTION OF DRAWINGS

[0071] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate embodiments consistent with the present application and serve to explain the principles of the present application together with the specification.

[0072] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the accompanying drawings needed to be used in the embodiments or prior art description will be briefly introduced as follows. Obviously, for those skilled in the art, other drawings can also be obtained based on these drawings without any creative effort.

[0073] One or more embodiments are exemplarily illustrated by the pictures in the drawings corresponding thereto, and these exemplary illustrations do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings represent similar elements, unless otherwise specified. The drawings in the drawings do not constitute a proportional limitation.

[0074] Figure 1 A flowchart of a script quality evaluation method provided by the embodiments of the present application is provided.

[0075] Figure 2 A flowchart of another script quality evaluation method provided by the embodiments of the present application is provided.

[0076] Figure 3 A flowchart of another script quality evaluation method provided by the embodiments of the present application is provided.

[0077] Figure 4 A structural schematic diagram of a script quality evaluation device provided by an embodiment of the present application is shown in the figure.

[0078] Figure 5 A structural schematic diagram of an electronic device provided by an embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0079] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described below in connection with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0080] The following disclosure provides many different embodiments, or examples, for implementing different structures of the present application. For the purpose of simplicity and clarity, the description in the following text describes the components and settings of specific examples. Of course, they are only examples, and the purpose is not to limit the present application. In addition, the present application can repeatedly refer to numbers and / or letters in different examples. Such repetition is for the purpose of simplification and clarity, and does not indicate the relationship between the various embodiments and / or settings discussed.

[0081] Figure 1 A flowchart of a script quality evaluation method provided by an embodiment of the present application is shown in the figure. The method can be applied to one or more electronic devices such as a smart phone, a notebook computer, a desktop computer, a portable computer, and a server. In addition, the execution subject of the method can be hardware or software. When the execution subject is hardware, the execution subject can be one or more of the above-mentioned electronic devices. For example, a single electronic device can execute the method, or multiple electronic devices can cooperate with each other to execute the method. When the execution subject is software, the method can be implemented as multiple software or software modules, or as a single software or software module. No specific limitation is made herein.

[0082] As shown in the figure, the method specifically includes: Figure 1

[0083] S101, obtaining text data of a script to be evaluated.

[0084] The script to be evaluated refers to a complete script text that needs to be analyzed for quality, including elements such as character dialogue and scene description.

[0085] The text data refers to converting the content of the script into a structured or semi-structured data form that can be processed by a computer.

[0086] ​In the embodiments of the present application, the original script file is acquired through a data acquisition interface, and after preprocessing such as denoising and format conversion, standardized text data containing complete script content and meeting the analysis requirements are generated. This step provides a standardized data basis for subsequent analysis, ensures that script data from different sources can be uniformly processed, while retaining key dramatic elements in the original script, and provides reliable data support for subsequent quantitative analysis.

[0087] In S102, a contribution degree index of each role to plot development is calculated based on the text data.

[0088] The contribution degree index is used to quantify the comprehensive influence of a role on plot development, including contribution to plot advancement, conflict and theme expression.

[0089] In the embodiments of the present application, by analyzing the performance characteristics (such as appearance characteristics, dialogue characteristics and behavior characteristics) of each role in the text data, the participation degree, plot driving force contribution and conflict influence of each role are determined, and then the contribution degree index of each role to plot development is calculated.

[0090] In this way, the traditional subjective role importance evaluation can be converted into a quantifiable index, providing an objective basis for subsequent matching analysis, and solving the problem of poor consistency of manual evaluation.

[0091] In S103, a support degree index of the relationship network of each role to plot development is calculated based on the text data.

[0092] The support degree index is used to measure the overall support of all interactive relationships of a role to plot development, and reflects the effectiveness of the relationship network in role shaping.

[0093] In the embodiments of the present application, first, the interactive relationships such as dialogue and conflict between roles are identified, then the contribution weight of each relationship to plot development and role character shaping is evaluated, and finally the support degree index is accumulated. In this way, the limitations of traditional relationship calculation can be broken through, the dramatic value of interactive relationships can be quantified, the actual support effect of the relationship network on role shaping can be accurately reflected, and data support for identifying roles with "inconsistent roles and relationships" can be provided.

[0094] In S104, the contribution degree index of each role is matched with the corresponding support degree index for matching degree analysis, and a script quality evaluation result is generated.

[0095] Matching degree analysis refers to the process of quantitatively comparing the difference between the contribution degree index and the support degree index of a role to evaluate their coordination.

[0096] In the embodiments of the present application, by matching the contribution degree index of each role with the corresponding support degree index, a script quality evaluation result is generated, the correlation evaluation of the importance of the role and the support degree of the relationship network is realized, the structural defects of "role and relationship mismatch" can be automatically identified, accurate quantitative basis is provided for script modification, and the technical blank of traditional methods that cannot evaluate the coordination of role-relationship is solved.

[0097] Specifically, S104 can include the following steps: for each role, performing normalization processing on the contribution degree index corresponding to the role to obtain contribution degree data, and performing normalization processing on the support degree index corresponding to the role to obtain support degree data; calculating a difference degree value between the contribution degree data and the support degree data, the difference degree value being used to represent the matching degree of the importance of the role and the support degree of the relationship network; based on the difference degree values of all roles, generating a script quality evaluation result, the script quality evaluation result including a comprehensive score reflecting the matching situation of all roles in the script, and / or imbalance roles and corresponding matching deviation features whose difference degree values exceed a preset threshold, and / or improvement suggestions generated according to the matching deviation features of each imbalance role.

[0098] Normalization processing refers to a data standardization method of unifying different dimension indexes to the same numerical range (such as the interval [0, 1]) through linear transformation. The difference degree value is a scalar index quantifying the deviation degree of the contribution degree data and the support degree data, the numerical value of which is negatively correlated with the matching degree of the role, and can be obtained by calculating the absolute difference value or the square difference value. The matching deviation feature is used to describe the specific type of mismatch between the contribution degree and the support degree of the role (such as "high contribution degree-low support degree").

[0099] The technical solution firstly eliminates the dimension difference between indexes through normalization processing to ensure comparability, then calculates the difference degree values of two dimensions of each role to accurately quantify the matching degree, and finally generates three types of evaluation results based on the difference degree distribution: 1) the comprehensive score of the whole script reflects the overall coordination; 2) specific imbalance roles and their deviation types are identified; 3) differentiated improvement suggestions are generated for different deviation types (such as adding key interactive relationships for "high contribution degree-low support degree" roles).

[0100] As a specific example, assuming that a script contains three roles, after normalization processing, the data is as follows: normalized index data: role A: contribution degree = 1.0, support degree = 0.67; role B: contribution degree = 0.4, support degree = 1.0; role C: contribution degree = 0.0, support degree = 0.0. Then, the difference degree value is calculated (two implementation methods):

[0101] Method one: absolute difference value method

[0102] Role A: |1.0-0.67|=0.33; Role B: |0.4-1.0|=0.6; Role C: |0.0-0.0|=0.0;

[0103] Then, the average of all role differences is 0.31, and finally, the matching degree score is 69 points by the formula "score=(1-average difference) x 100".

[0104] Method two: square difference formula

[0105] Role A: (1.0-0.67)2=0.1089; Role B: (0.4-1.0)2=0.36; Role C: (0.0-0.0)2=0.0.

[0106] Then, the average of all role differences is 0.156, and finally, the matching degree score is 84.4 points by the formula "score=(1-average difference) x 100".

[0107] Among them, the square difference calculation amplifies the influence of larger deviations, such as the difference of role B from 0.6 to 0.36, which increases the overall score.

[0108] At the same time, identify the unbalanced role B (0.6) whose difference exceeds the threshold value 0.5, and its "high support-low contribution" feature indicates that the role relationship network is rich but the role is insufficient, and it is suggested to increase the participation of key plot.

[0109] This scheme realizes the overall evaluation of the script quality through standardized comparison and typization analysis, and can not only locate the specific structural problems and provide the operable modification direction, but also significantly improve the accuracy and practicality of the script evaluation.

[0110] In the embodiment of the application, after obtaining the script text data, the contribution degree index and the relationship network support degree index of each role are calculated according to the script text data, and then the correlation evaluation mechanism between the role and its relationship network is established through matching degree analysis, so that the actual support effect of the role relationship network on the role shaping can be accurately reflected, and the accuracy of the script analysis is improved.

[0111] Referring to Figure 2 , another embodiment flowchart of the script quality evaluation method provided in the embodiment of the application. The Figure 2 flowchart shown in the above Figure 1 , describes how to calculate the contribution degree index of each role to the plot development. As Figure 2 shown, the flowchart can include the following steps:

[0112] S201, divide the text data according to the episodes to obtain the script data of each episode.

[0113] episode division, refers to the process of splitting a complete script into several independent units according to the natural division structure of the script (such as the number of episodes of a TV series or the number of scenes of a play).

[0114] In the embodiments of the present application, the original text data is converted into a structured data set organized by episodes by analyzing the episode markers (such as "Xth episode" format identifiers) of the script or manually preset division rules, ensuring that each episode script data contains complete scene, dialogue and character interaction information. This scheme not only preserves the complete dramatic structure of the script, but also provides standardized data input for subsequent episode-by-episode analysis, solving the problem of coarse granularity in overall analysis of long scripts.

[0115] S202, for each episode script data, calculate the plot importance score of each character in the script data to obtain the single-episode contribution score corresponding to each character.

[0116] Single-episode contribution score is a numerical indicator that quantifies the importance of a character in a single-episode plot, reflecting the character's overall impact on the current episode plot development.

[0117] In the embodiments of the present application, each episode script data (usually the original script) is input into the fine-tuned large model in turn, and the built-in attention mechanism of the model is used to analyze the script data and extract the importance score of each character in the plot.

[0118] Specifically, the calculation of the plot importance score of each character in the script data to obtain the single-episode contribution score corresponding to each character can include the following steps: for each character, extract the character's performance feature set from the script data, the performance feature set including appearance features, dialogue features and behavior features; based on the performance feature set, analyze the influence of the character on the current episode plot development, wherein: the appearance features are used to calculate the character's participation basis value; the dialogue features are used to analyze the plot driving force contribution value; the behavior features are used to evaluate the conflict impact value; the character's participation basis value, the plot driving force contribution value and the conflict impact value are weighted and combined to generate the single-episode contribution score of the character in the current episode.

[0119] Performance feature set refers to a multi-dimensional feature set extracted from the script text that represents the dramatic influence of a character, including: 1) appearance features (such as appearance frequency, scene coverage rate, etc. basic data); 2) dialogue features (such as dialogue emotional intensity, topic dominance, etc. language features); 3) behavior features (such as decision criticality, conflict initiation intensity, etc. action features).

[0120] The scheme includes the following three stages:

[0121] 1. Feature Extraction Phase: The large model extracts a set of dramatic action features through multi-level text analysis. For appearance features, the model uses named entity recognition to locate characters and combines it with a scene segmentation algorithm to calculate appearance frequency, duration, and participation in key scenes. For dialogue features, the model uses a sentiment analysis module to quantify the emotional value of lines (e.g., a continuous interval from -1 to 1), identifies the degree of thematic control of the dialogue through topic modeling, and uses causal reasoning to analyze the intensity of the lines' influence on the plot. For behavioral features, the model constructs an action-influence relationship graph based on event extraction technology and uses a graph neural network to assess the importance of decision nodes and the scope of conflict. This phase automatically focuses on key dramatic elements in the script through an attention mechanism, avoiding the limitations of manually set features.

[0122] Second, during the feature quantification phase, the model converts raw features into standardized dramatic value metrics. A baseline value for character engagement is generated using a weighted algorithm for key scenes (e.g., a weight coefficient of 2.0 is set for mainline scenes). A bidirectional LSTM (Long Short-Term Memory) model analyzes the causal impact of dialogue chains, outputting a normalized value between 0 and 1. Conflict impact is assigned a graded value based on the number and intensity of subsequent conflict scenarios triggered by an action (e.g., actions that trigger more than three subsequent conflicts are assigned a value between 0.8 and 1.0). Each quantification process utilizes thresholds guided by dramatic theory; for example, dialogue with an emotional intensity exceeding 0.7 is automatically labeled as high-value.

[0123] During the rating generation phase, the system dynamically assigns weighting parameters based on the script type: For plot-driven dramas (such as mysteries), the weightings for plot driving force are set at 45%, conflict at 40%, and engagement at 15%. For character-driven dramas (such as ethical dramas), the weightings are adjusted to 35%, 30%, and 35%, respectively. After weighting, a sigmoid function is used to map the raw scores to a reasonable range of 60-95 (avoiding extreme values), ultimately generating a rating that provides dramatic interpretation (e.g., 82 for "this character plays a key role in the episode, but the character arc is incomplete"). The median value for each dimension is retained during the rating process to generate subsequent improvement suggestions.

[0124] Through the fusion of multi-dimensional features, this solution not only retains the objectivity of traditional statistical methods (such as appearance counts), but also combines deep semantic analysis (such as dialogue influence assessment). It can comprehensively capture the substantive contributions of characters in many aspects such as dramatic conflict and plot advancement, and solve the problem of one-sided evaluation of a single indicator.

[0125] S203. For each character, generate an index of the character's contribution to the plot development based on all the single-episode contribution scores corresponding to the character.

[0126] The whole drama contribution index refers to a comprehensive index generated by aggregating the single-episode contribution scores of a character in all episodes, reflecting the overall importance of the character in the entire script.

[0127] In the embodiments of the present application, first, time series analysis is performed on the single-episode contribution scores of the character in each episode to identify the importance change trend of the character in different stages of the plot; then, a weighted aggregation algorithm (such as key episode number weight addition, continuity correction factor, etc.) is used to calculate the overall contribution, wherein the score weight of the key episode set can be 1.5-2 times that of the ordinary set; finally, a two-dimensional index (such as character A: total score 85 + rising trend) is generated, which has both "overall performance" and "development trajectory". This scheme, through dynamic weight adjustment across episodes, not only retains the fineness of single-episode analysis, but also accurately reflects the comprehensive value of the character in the complete plot, solving the problem of ignoring the characteristics of the development stage in traditional methods.

[0128] Figure 2 As shown in the flow, first, the drama structure integrity of the script is preserved through episode-by-episode processing, making the analysis granularity accurate to the single-episode level; second, the single-episode score based on the large model can capture the substantial contribution of the character in the dimensions of plot advancement and conflict development, overcoming the superficial defects of traditional statistical methods; finally, through time-weighted whole-drama index aggregation, both the importance peak of the character in key episodes and its development trajectory in the whole drama are considered, and finally the role evaluation result with timeliness and overall performance is generated, providing complete data support for script optimization from macrostructure to micro-adjustment.

[0129] Referring to Figure 3 An embodiment flowchart of another script quality evaluation method provided in the embodiments of the present application. The Figure 3 The flowchart shown in the above Figure 1 Based on the flowchart shown in the above Figure 3 As shown, the flowchart can include the following steps:

[0130] S301, divide the text data by episodes to obtain episode script data.

[0131] Episode division refers to the process of splitting the complete script into several independent units according to the natural episode structure of the script (such as the number of episodes of a TV series or the number of scenes of a play).

[0132] In the embodiments of the present application, the original text data is converted into a structured data set organized by episodes through parsing the episode markers (such as "Xth episode" format identifier) of the script or manually preset division rules, ensuring that each episode script data contains complete scene, dialogue and character interaction information. This scheme not only preserves the complete dramatic structure of the script, but also provides standardized data input for subsequent episode-by-episode analysis, solving the problem of coarse granularity in long script overall analysis.

[0133] S302, for each episode script data, identify the interaction relationship between the characters in the script data, calculate the single-episode support score of each character in the current episode based on the contribution degree of the interaction relationship to the plot of the current episode.

[0134] Interaction relationship refers to the dramatic connection between characters through dialogue, conflict or cooperation. Single-episode support score is used to quantify the overall support of the character relationship network to the current plot.

[0135] In the embodiments of the present application, each episode script data (usually the original script) is input into the fine-tuned large model in turn, and the script data is analyzed through the attention mechanism built-in the model to identify the interaction relationship between characters (including two characters A and B), and score the importance of each pair of interaction relationship to the plot, and then calculate the single-episode support score of each character in each episode.

[0136] Specifically, based on the contribution degree of the interaction relationship to the plot of the current episode, the single-episode support score of each character in the current episode can include the following steps: for each group of interaction relationships in the current episode, calculate the contribution score of the interaction relationship to the plot of the current episode; for each character, identify all target interaction relationships containing the character; sum all target interaction relationship contribution scores to obtain the single-episode support score of the character in the current episode.

[0137] Contribution score, refers to the value evaluation of a single interaction relationship to the plot quantified by a large language model. In this step, the large model first evaluates each set of interaction relationships (A, B) based on its multi-modal understanding capabilities (such as semantic analysis, emotion recognition) to obtain an original score, then the model corrects the original score by weighting the pre-trained drama knowledge module, and outputs a standardized contribution score of 0-10 points (such as Score_AB=7.2). For each character X, the system automatically aggregates the scores of all associated interaction relationships (X, Y) of the character, and generates a single set support score of 0-100 points through weighted summation (such as primary relationship weight 1.2, secondary relationship weight 0.8). This technical solution calculates the contribution score of the interaction relationship of the character and aggregates to generate the support score, realizes the objective quantification of the value of the relationship network of the character, and can accurately reflect the overall support of the interaction relationship of each character to the plot, providing data support for identifying key characters with weak relationship networks.

[0138] Among them, calculating the contribution score of the interaction relationship to the plot of the current episode includes: extracting the interaction behavior feature set of the interaction relationship, the interaction behavior feature set includes dialogue interaction features, plot interaction features and conflict interaction features; based on the interaction behavior feature set, a multi-dimensional contribution value corresponding to the interaction relationship is calculated, wherein the dialogue interaction features are used to calculate the plot driving contribution value; the plot interaction features are used to calculate the character shaping contribution value; the conflict interaction features are used to calculate the dramatic tension contribution value; the plot driving contribution value, the character shaping contribution value and the dramatic tension contribution value are weighted and integrated to generate the contribution score of the interaction relationship to the plot of the current episode.

[0139] Interaction behavior feature set, refers to a multi-dimensional feature set extracted from the dialogue and plot of the script, representing the dramatic value of the interaction relationship, including: 1) dialogue interaction features (such as topic turning intensity, emotional interaction depth, etc. Language features); 2) plot interaction features (such as the number and importance of key events participated in together); 3) conflict interaction features (such as the degree of conflict escalation, the duration of confrontation, etc.). In this scheme, first, the large model extracts the three types of features, then calculates the plot driving contribution value (quantifies the influence of interaction on the main line development), the character shaping contribution value (evaluates the character personality dimensions revealed by interaction) and the dramatic tension contribution value (measures the conflict intensity generated by interaction); finally, the pre-set drama weight (such as plot driving 40%, character shaping 30%, dramatic tension 30%) is used for weighted fusion to generate a standardized contribution score of 0-10 points. This scheme evaluates through multi-dimensional feature fusion, not only retains the relationship structure data of traditional social network analysis, but also increases the depth quantification of interaction dramatic value, and can accurately identify key interaction relationships that have substantial support for plot development.

[0140] S303, for each role, generating a support degree index of the role to the plot development according to all single-episode support degree scores corresponding to the role.

[0141] The whole-episode support degree index is a comprehensive index generated by aggregating the single-episode support degree scores of the role in all episodes, reflecting the overall support strength of the relationship network of the role in the whole script.

[0142] In the embodiments of the present application, first, time series analysis is performed on the single-episode support degree scores of the role in each episode to identify the trend of the relationship network strength changing with the plot development; then, data aggregation is performed using a key scene weighting algorithm (such as a climax episode score weight increase of 50%) and continuity correction (such as an additional score for a continuous interaction relationship); finally, a two-dimensional index (such as role A: total score 78 + fluctuation index 0.2) containing "overall support strength" and "development stability" is generated. Through dynamic evaluation across episodes, this scheme not only reflects the peak support of the relationship network of the role in key scenes, but also embodies its continuous effect in the whole script, solving the problem that traditional methods ignore the dynamic evolution of the relationship network.

[0143] Figure 3 The flowchart shown ensures a fine grasp of the script structure by identifying the interaction relationship of the role in each episode; the single-episode support degree score is calculated based on multi-dimensional features to achieve deep quantification of the dramatic value of the relationship network of the role; and finally, the comprehensive evaluation system reflecting the dynamic evolution process of the relationship network is established through the whole-episode support degree index generation. This technical solution systematically solves the technical defects of traditional methods that cannot evaluate the continuous support of the relationship network to the role, and provides a complete data chain support from single-episode interaction analysis to whole-episode relationship evolution for script optimization.

[0144] Based on the same technical concept, the embodiments of the present application also provide a script quality evaluation device, as shown in Figure 4 The device comprises:

[0145] The acquisition module 41 is configured to acquire text data of a script to be evaluated.

[0146] The first calculation module 42 is configured to calculate a contribution degree index of each role to the plot development based on the text data.

[0147] The second calculation module 43 is configured to calculate a support degree index of the relationship network of each role to the plot development based on the text data.

[0148] The analysis module 44 is configured to perform matching degree analysis on the contribution degree index and the corresponding support degree index of each role to generate a script quality evaluation result.

[0149] In one possible implementation, the first calculation module is specifically configured to:

[0150] dividing the text data according to episodes to obtain script data of each episode;

[0151] For each script data, a plot importance score of each character in the script data is calculated to obtain a single-episode contribution score corresponding to each character;

[0152] For each character, a contribution index of the character to plot development is generated according to all single-episode contribution scores corresponding to the character.

[0153] In a possible implementation, the first calculation module is further configured to:

[0154] For each character, a set of acting behavior features of the character is extracted from the script data, the set of acting behavior features including a feature of appearance, a feature of dialogue and a feature of behavior;

[0155] Based on the set of acting behavior features, an influence degree of the character on a plot development of a current episode is analyzed, wherein the feature of appearance is used to calculate a basic value of character participation, the feature of dialogue is used to analyze a contribution value of plot driving force, and the feature of behavior is used to evaluate an influence value of conflict;

[0156] The basic value of character participation, the contribution value of plot driving force and the influence value of conflict are weighted and comprehensively calculated to generate a single-episode contribution score of the character in the current episode.

[0157] In a possible implementation, the second calculation module is specifically configured to:

[0158] dividing the text data according to episodes to obtain script data of each episode;

[0159] For each script data, an interactive relationship between characters in the script data is identified, and a single-episode support score of each character in a current episode is calculated based on a contribution degree of the interactive relationship to a plot of the current episode;

[0160] For each character, a support index of the character to plot development is generated according to all single-episode support scores corresponding to the character.

[0161] In a possible implementation, the second calculation module is further configured to:

[0162] For each interactive relationship in a current episode, a contribution score of the interactive relationship to a plot of the current episode is calculated;

[0163] For each character, all target interactive relationships containing the character are identified;

[0164] The contribution scores corresponding to all target interactive relationships are summed up to obtain the single episode support score of the character in the current series.

[0165] In a possible implementation manner, the second calculation module is further configured to:

[0166] Extracting an interactive behavior feature set of the interactive relationship, wherein the interactive behavior feature set includes dialogue interaction features, plot interaction features, and conflict interaction features;

[0167] Based on the interactive behavior feature set, the multi-dimensional contribution value corresponding to the interactive relationship is calculated, wherein the dialogue interaction feature is used to calculate the plot advancement contribution value; the plot interaction feature is used to calculate the character creation contribution value; and the conflict interaction feature is used to calculate the dramatic tension contribution value;

[0168] The contribution value of plot promotion, the contribution value of character creation and the contribution value of dramatic tension are weighted and comprehensively calculated to generate a contribution score of the interactive relationship to the plot of the current episode.

[0169] In one possible implementation, the analysis module is specifically configured to:

[0170] For each role, normalizing the contribution index corresponding to the role to obtain contribution data, and normalizing the support index corresponding to the role to obtain support data;

[0171] Calculating a difference value between the contribution data and the support data, wherein the difference value is used to represent the degree of matching between the role importance and the relationship network support;

[0172] Based on the difference values ​​of all characters, a script quality assessment result is generated, which includes a comprehensive score reflecting the matching situation of all characters in the play, and / or unbalanced characters whose difference values ​​exceed a preset threshold and the corresponding matching deviation characteristics, and / or improvement suggestions generated based on the matching deviation characteristics of each unbalanced character.

[0173] Based on the same technical concept, the embodiment of the present application also provides an electronic device, such as Figure 5 As shown, it includes a processor 111, a communication interface 112, a memory 113 and a communication bus 114, wherein the processor 111, the communication interface 112, and the memory 113 communicate with each other through the communication bus 114.

[0174] Memory 113, for storing computer programs;

[0175] The processor 111 is configured to execute the program stored in the memory 113 by performing the following steps:

[0176] Obtain text data of a script to be evaluated;

[0177] Based on the text data, calculate a contribution degree index of each role to the plot development;

[0178] Based on the text data, calculate a support degree index of the relationship network of each role to the plot development;

[0179] Conduct matching degree analysis on the contribution degree index of each role and the corresponding support degree index, and generate a script quality evaluation result.

[0180] The communication bus mentioned in the above electronic device can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0181] The communication interface is used for communication between the above electronic device and other devices.

[0182] The memory can include a Random Access Memory (RAM) and can also include a Non-Volatile Memory (NVM), such as at least one disk memory. Optionally, the memory can also be at least one storage device located away from the aforementioned processor.

[0183] The processor mentioned above can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processing (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.

[0184] In another embodiment provided in the present application, a computer readable storage medium is also provided, and the computer readable storage medium stores a computer program. When the computer program is executed by a processor, the steps of any of the above script quality evaluation methods are implemented.

[0185] In yet another embodiment provided in the present application, a computer program product containing instructions, which, when executed on a computer, cause the computer to perform the script quality evaluation method of any of the above embodiments, is also provided.

[0186] The device embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the embodiment according to actual needs.

[0187] From the above description of the embodiments, those skilled in the art can clearly understand that the embodiments can be implemented by means of software plus a general hardware platform, and of course can also be implemented by hardware. Based on such understanding, the above technical solutions, essentially or in terms of related art, can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in the various embodiments or some parts of the embodiments.

[0188] It should be understood that the terms used herein are for the purpose of describing particular example embodiments only and are not intended to be limiting. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. The terms "comprises", "comprising", "includes", "including" and "has" are inclusive and therefore specify the presence of stated features, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, steps, operations, elements, components, and / or groups thereof. The method steps, processes, and operations described herein are not to be construed as necessarily requiring their performance in the particular order in which they are described, unless specifically identified as an order dependent step. It is also to be understood that additional or alternative steps can be employed.

[0189] The above description is merely illustrative of the application and does not limit the application, as the application can be modified in various ways. Accordingly, the application is not limited to the embodiments disclosed herein, but is intended to cover all modifications falling within the spirit and scope of the application. Therefore, the application will not be limited to the embodiments shown herein, but is intended to cover the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A script quality assessment method, characterized in that: The method comprises: Obtain the text data of the script to be evaluated; Based on the text data, calculating the contribution index of each character to the development of the plot; Based on the text data, calculating the support index of each character's relationship network for the plot development; Perform a matching analysis on each character's contribution index and the corresponding support index to generate the script quality assessment results.

2. The method according to claim 1, characterized in that The calculation of each character's contribution to the plot development index includes: Divide the text data by episode to obtain script data for each episode; For each episode of the script data, calculate the plot importance score of each character in the script data to obtain the corresponding single episode contribution score of each character; For each character, an index of the character's contribution to the plot development is generated based on the contribution scores of all episodes corresponding to the character.

3. The method according to claim 2, characterized in that The step of calculating the plot importance score of each character in the script data to obtain the single episode contribution score corresponding to each character includes: For each character, extracting a dramatic behavior feature set of the character from the script data, the dramatic behavior feature set including appearance features, dialogue features, and behavior features; Based on the set of dramatic behavior features, the degree of influence of the character on the development of the current episode is analyzed, wherein: the appearance features are used to calculate the base value of the character's participation; the dialogue features are used to analyze the contribution value of the plot driving force; and the behavior features are used to evaluate the impact value of contradictions and conflicts; A weighted comprehensive calculation is performed on the basic value of the role participation, the contribution value of the plot driving force and the contradiction and conflict impact value to generate a single episode contribution score of the role in the current series.

4. The method according to claim 1, wherein The calculation of the support index of each character's relationship network for the plot development includes: Divide the text data by episode to obtain script data for each episode; For each episode of the script data, identify the interactive relationships between the characters in the script data, and calculate the single episode support score of each character in the current episode based on the contribution of the interactive relationships to the plot of the current episode; For each character, an index of the character's support for the plot development is generated based on the support scores of all single episodes corresponding to the character.

5. The method according to claim 4, characterized in that The calculation of the single episode support score of each character in the current episode based on the contribution of the interactive relationship to the plot of the current episode includes: For each set of interactive relationships in the current episode, calculate the contribution score of the interactive relationship to the plot of the current episode; For each role, identify all target interactions involving the role; The contribution scores corresponding to all target interactive relationships are summed up to obtain the single episode support score of the character in the current series.

6. The method according to claim 5, characterized in that Calculating the contribution score of the interactive relationship to the current episode plot includes: Extracting an interactive behavior feature set of the interactive relationship, wherein the interactive behavior feature set includes dialogue interaction features, plot interaction features, and conflict interaction features; Based on the interactive behavior feature set, the multi-dimensional contribution value corresponding to the interactive relationship is calculated, wherein the dialogue interaction feature is used to calculate the plot advancement contribution value; the plot interaction feature is used to calculate the character creation contribution value; and the conflict interaction feature is used to calculate the dramatic tension contribution value; The contribution value of plot promotion, the contribution value of character creation and the contribution value of dramatic tension are weighted and comprehensively calculated to generate a contribution score of the interactive relationship to the plot of the current episode.

7. The method according to claim 1, characterized in that The matching analysis of each character's contribution index and the corresponding support index is performed to generate a script quality assessment result, including: For each role, normalizing the contribution index corresponding to the role to obtain contribution data, and normalizing the support index corresponding to the role to obtain support data; Calculating a difference value between the contribution data and the support data, wherein the difference value is used to represent the degree of matching between the role importance and the relationship network support; Based on the difference values ​​of all characters, a script quality assessment result is generated, which includes a comprehensive score reflecting the matching situation of all characters in the play, and / or unbalanced characters whose difference values ​​exceed a preset threshold and the corresponding matching deviation characteristics, and / or improvement suggestions generated based on the matching deviation characteristics of each unbalanced character.

8. A script quality assessment device, characterized in that: The device comprises: An acquisition module is used to obtain text data of the script to be evaluated; A first calculation module is used to calculate the contribution index of each character to the development of the plot based on the text data; A second calculation module is used to calculate the support index of the relationship network of each character for the development of the plot based on the text data; The analysis module is used to perform matching analysis on the contribution index of each role and the corresponding support index to generate the script quality assessment results.

9. An electronic device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus; Memory for storing computer programs; The processor is used to implement the script quality assessment method described in any one of claims 1 to 7 when executing the program stored in the memory.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the script quality assessment method according to any one of claims 1 to 7.

Images