A speech generation model building system based on digital visualization

Abstract

The present invention relates to the technical field of speech generation model building system, and provides a speech generation model building system based on digital visualization, including a data processing terminal, a model building terminal, a visualization interface terminal and a speech synthesis terminal; the data processing terminal is used for data processing and feature extraction of input data; the visualization interface terminal is used for providing a user operation interface and reading the model parameters input by the user; the model building terminal is used for building a speech generation model according to the model parameters input by the user; the speech synthesis terminal is used for controlling the speech generation module to generate corresponding speech information according to the data processed and the speech features. The present invention has the effect of improving the working efficiency of the speech generation model.

Description

Technical Field

[0001] The present invention relates to the technical field of speech generation model building systems, and in particular to a speech generation model building system based on digital visualization. Background Art

[0002] Speech synthesis technology has a wide range of applications in human-computer interaction, intelligent assistants and other fields. Traditional speech synthesis methods are usually based on rules, splicing or statistics, but these methods often have difficulty in generating natural and fluent speech. In recent years, the development of deep learning technology has made significant progress in speech synthesis models based on neural networks. The speech generation model building system based on digital visualization is a system that uses digital graphical interfaces and visualization tools to build, debug and optimize speech generation models. This system combines speech synthesis technology with a graphical user interface, allowing users to intuitively operate and manage the various components of the speech generation model, thereby simplifying the model design and adjustment process.

[0003] Now many speech generation model systems have been developed. After a lot of searching and reference, we found that the speech generation model systems in the prior art are such as the speech generation model systems disclosed in the publication numbers CN114495898A, CN116884389A, CN111627418A, EP1528538A1, US20240038217A1, and JP2019124940A. These speech generation model systems generally include: a data acquisition terminal, a speech generation model loading terminal, and a speech information output terminal; the data acquisition terminal is used to acquire speech data from the user; the speech generation model loading terminal is used to load a preset speech generation model; the speech information output terminal is used to generate corresponding speech information using the speech generation model according to the user's speech data. Since the speech generation model of the above-mentioned speech generation model system is relatively single, the processing method of speech data and the generation process of speech information are relatively single, which is not conducive to meeting the different needs of multiple users, resulting in the defect of reduced working efficiency of the speech generation model of the system. Summary of the invention

[0004] The purpose of the present invention is to propose a speech generation model building system based on digital visualization in view of the shortcomings of the above-mentioned speech generation model building system.

[0005] The present invention adopts the following technical solution:

[0006] A speech generation model building system based on digital visualization includes a data processing terminal, a model building terminal, a visualization interface terminal and a speech synthesis terminal; the data processing terminal is used to process and extract features of input data; the visualization interface terminal is used to provide a user operation interface and read the model parameters input by the user; the model building terminal is used to build a speech generation model according to the model parameters input by the user; the speech synthesis terminal is used to control the speech generation module to generate corresponding speech information according to the data and speech features after data processing;

[0007] The data processing terminal includes a voice data processing module, a text data processing module and a voice feature extraction module; the voice data processing module is used to preprocess the input voice data; the preprocessing includes loading, format conversion and noise reduction; the text data processing module is used to read and recognize the input text data; the voice feature extraction module is used to extract voice features from the preprocessed voice data.

[0008] Optionally, the visualization interface terminal includes an interface management module, a model parameter reading module and an automatic interaction module; the interface management module is used to provide a user operation interface; the user operation interface includes various model parameter adjustment items, a model effect preview window and a data display window for stored data; the model parameter reading module is used to read the specific data that has been adjusted by the user in various model parameter adjustment items; the automatic interaction module is used to perform automatic interaction according to the user operation situation to complete the regulation of the user operation interface.

[0009] Optionally, the automatic interaction module includes a user operation process monitoring information acquisition submodule, a monitoring information analysis submodule, an interaction demand index calculation submodule and an automatic interaction control submodule; the user operation process monitoring information acquisition submodule is used to acquire user video information from the monitoring camera and user operation record information from the system; the monitoring information analysis submodule is used to analyze the user video information and user operation record information, and generate corresponding user operation interface still time data, mouse movement data and user interface viewing time data; the interaction demand index calculation submodule is used to calculate the corresponding interaction demand index based on the user operation interface still time data, mouse movement data and user interface viewing time data; the automatic interaction control submodule is used to automatically interactively control the user operation interface according to the interaction demand index, and automatically retrieve and display the speech generation model parameters that are used most frequently by the corresponding user.

[0010] Optionally, the interaction demand index calculation submodule includes a still time score calculation unit, a mouse movement score calculation unit, a user viewing time score calculation unit and an interaction demand index integration unit; the still time score calculation unit is used to calculate the still time score of the corresponding user based on the still time data of the user operation interface; the mouse movement score calculation unit is used to calculate the corresponding mouse movement score based on the mouse movement data of the user when performing the operation; the user viewing time score calculation unit is used to calculate the corresponding user viewing time score based on the user viewing interface time data; the interaction demand index integration unit is used to integrate the still time score, mouse movement score and user viewing time score into a corresponding interaction demand index.

[0011] Optionally, the visualization interface terminal further includes a model parameter confirmation prompt module; the model parameter confirmation prompt module is used to generate corresponding model parameter confirmation prompt information according to the user's operation process after the user completes the model parameter adjustment.

[0012] Optionally, the model parameter confirmation prompt module includes a user operation process line of sight information acquisition submodule, a mouse operation record information acquisition submodule, a parameter adjustment time information acquisition submodule, a prompt index calculation submodule and a model parameter confirmation prompt information generation submodule; the user operation process line of sight information acquisition submodule is used to obtain the user operation process line of sight information from the eye tracker; the mouse operation record information acquisition submodule is used to obtain the mouse operation record information when the user adjusts various model parameter adjustment items; the parameter adjustment time information acquisition submodule is used to obtain the parameter adjustment time when the user adjusts various model parameter adjustment items; the prompt index calculation submodule is used to calculate the corresponding prompt index according to the user operation process line of sight information, mouse operation record information and parameter adjustment time; the model parameter confirmation prompt information generation submodule is used to generate corresponding model parameter confirmation prompt information according to the prompt index.

[0013] A method for building a speech generation model based on digital visualization, such as the above-mentioned system for building a speech generation model based on digital visualization, the method for building a speech generation model includes:

[0014] S1, data processing and feature extraction of input data;

[0015] S2, providing a user operation interface and reading the model parameters entered by the user;

[0016] S3, constructing a speech generation model according to the model parameters entered by the user;

[0017] S4, generating corresponding voice information according to the processed data and voice features.

[0018] The beneficial effects achieved by the present invention are:

[0019] 1. The data processing terminal, model building terminal, visual interface terminal and speech synthesis terminal are set up. Data processing and feature extraction are first realized through an independent data processing terminal, and then a visual interface is provided for direct user interaction to complete the adjustment of different model parameters, so as to quickly build an accurate speech generation model. Finally, the customized speech generation model is used to complete the speech information output, which is conducive to improving the adaptability and accuracy of the speech generation model and improving the working efficiency of the system's speech generation model;

[0020] 2. The setting of the voice data processing module, the text data processing module and the voice feature extraction module improves the efficiency of data processing by processing voice data and text data simultaneously, which is conducive to improving the working efficiency of the system, improving the accuracy of data processing, and then improving the accuracy and construction efficiency of the voice generation model, thereby helping to improve the working efficiency of the system's voice generation model;

[0021] 3. The settings of the interface management module, model parameter reading module and automatic interaction module improve the comprehensiveness and convenience of the user operation interface by managing various model parameter adjustment items, model effect preview windows and stored data display windows, which is conducive to improving the accuracy and efficiency of the speech generation model construction, thereby improving the working efficiency of the system's speech generation model;

[0022] 4. The setting of the user operation process monitoring information acquisition submodule, monitoring information analysis submodule, interaction demand index calculation submodule and automatic interaction control submodule calculates the corresponding interaction demand index through the user operation interface static time data, mouse movement data and user viewing interface time data, and then automatically interacts and controls the user operation interface according to the interaction demand index, which is conducive to further improving the intelligence of the system, improving the accuracy and efficiency of the speech generation model construction, and thus improving the working efficiency of the system's speech generation model;

[0023] 5. The still time score calculation unit, the mouse movement score calculation unit, the user viewing time score calculation unit and the interaction demand index integration unit are set up in coordination with the still time score algorithm, the mouse movement score algorithm and the user viewing time score algorithm. The still time score of the corresponding user is calculated through the still time data of the user operation interface, the corresponding mouse movement score is calculated through the mouse movement data of the user when performing the operation, and the corresponding user viewing time score is calculated through the user viewing interface time data. Then, the still time score, the mouse movement score and the user viewing time score are integrated into the corresponding interaction demand index, which is conducive to improving the accuracy of the interaction demand index, thereby improving the accuracy and efficiency of the automatic interaction control, and thus helping to improve the working efficiency of the system's speech generation model;

[0024] 6. The model parameter confirmation prompt module includes a user operation process sight line information acquisition submodule, a mouse operation record information acquisition submodule, a parameter adjustment time information acquisition submodule, a prompt index calculation submodule and a model parameter confirmation prompt information generation submodule. The setting cooperates with the prompt index algorithm, and calculates the corresponding prompt index through the user operation process sight line information, mouse operation record information and parameter adjustment time, which is conducive to improving the accuracy of the prompt index, and then is conducive to further improving the accuracy of the speech generation model, thereby helping to improve the working efficiency of the system's speech generation model;

[0025] 7. The voice data automatic deletion module includes a user monitoring information acquisition submodule, a user location information acquisition submodule, an automatic deletion evaluation index calculation submodule and a voice data automatic deletion instruction generation submodule. The settings cooperate with the automatic deletion evaluation index algorithm to determine whether the automatic deletion operation is performed through an accurate automatic deletion evaluation index, which is conducive to resetting the system for other users in advance, while improving the privacy and security of the current user's data.

[0026] To further understand the features and technical contents of the present invention, please refer to the following detailed description and drawings of the present invention. However, the drawings provided are only for reference and description and are not intended to limit the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a schematic diagram of the overall structure of the present invention;

[0028] Figure 2 It is a structural schematic diagram of the automatic interaction module in the present invention;

[0029] Figure 3 This is a structural diagram of the model parameter confirmation prompt module in the present invention;

[0030] Figure 4 A schematic diagram of a method flow of a method for building a speech generation model based on digital visualization in the present invention;

[0031] Figure 5 It is a structural diagram of the automatic voice data deletion module in the present invention. DETAILED DESCRIPTION

[0032] The following is an explanation of the embodiments of the present invention through specific embodiments. Those skilled in the art can understand the advantages and effects of the present invention from the contents disclosed in this specification. The present invention can be implemented or applied through other different specific embodiments, and the details in this specification can also be modified and changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. In addition, the drawings of the present invention are only simple schematic illustrations and are not depicted according to actual dimensions. It is stated in advance. The following embodiments will further explain the relevant technical contents of the present invention in detail, but the disclosed contents are not intended to limit the scope of protection of the present invention.

[0033] Embodiment 1: This embodiment provides a system for building a speech generation model based on digital visualization. Figure 1 As shown, a speech generation model building system based on digital visualization includes a data processing terminal, a model building terminal, a visualization interface terminal and a speech synthesis terminal; the data processing terminal is used to process and extract features of input data; the visualization interface terminal is used to provide a user operation interface and read the model parameters input by the user; the model building terminal is used to build a speech generation model according to the model parameters input by the user; the speech synthesis terminal is used to control the speech generation module to generate corresponding speech information according to the data and speech features after data processing;

[0034] The data processing terminal includes a voice data processing module, a text data processing module and a voice feature extraction module; the voice data processing module is used to preprocess the input voice data; the preprocessing includes loading, format conversion and noise reduction; the text data processing module is used to read and recognize the input text data; the voice feature extraction module is used to extract voice features from the preprocessed voice data.

[0035] Optionally, the visualization interface terminal includes an interface management module, a model parameter reading module and an automatic interaction module; the interface management module is used to provide a user operation interface; the user operation interface includes various model parameter adjustment items, a model effect preview window and a data display window for stored data; the model parameter reading module is used to read the specific data that has been adjusted by the user in various model parameter adjustment items; the automatic interaction module is used to perform automatic interaction according to the user operation situation to complete the regulation of the user operation interface.

[0036] Optional, combined Figure 2As shown, the automatic interaction module includes a user operation process monitoring information acquisition submodule, a monitoring information analysis submodule, an interaction demand index calculation submodule and an automatic interaction control submodule; the user operation process monitoring information acquisition submodule is used to obtain user video information from the monitoring camera and user operation record information from the system; the monitoring information analysis submodule is used to analyze the user video information and user operation record information, and generate corresponding user operation interface still time data, mouse movement data and user interface viewing time data; the interaction demand index calculation submodule is used to calculate the corresponding interaction demand index according to the user operation interface still time data, mouse movement data and user interface viewing time data; the automatic interaction control submodule is used to automatically interact with the user operation interface according to the interaction demand index, and automatically retrieve and display the voice generation model parameters that are used most frequently by the corresponding user.

[0037] Optionally, the interaction demand index calculation submodule includes a still time score calculation unit, a mouse movement score calculation unit, a user viewing time score calculation unit and an interaction demand index integration unit; the still time score calculation unit is used to calculate the still time score of the corresponding user based on the still time data of the user operation interface; the mouse movement score calculation unit is used to calculate the corresponding mouse movement score based on the mouse movement data of the user when performing the operation; the user viewing time score calculation unit is used to calculate the corresponding user viewing time score based on the user viewing interface time data; the interaction demand index integration unit is used to integrate the still time score, mouse movement score and user viewing time score into a corresponding interaction demand index.

[0038] When the user sees the user operation interface, the automatic interaction module starts working for a preset time. If the interaction demand index calculated after the preset time reaches the standard, the automatic interaction module automatically controls the user operation interface, that is, automatically retrieves and displays the voice generation model parameters that are used most by the corresponding user. The preset time is set by the administrator based on experience. Generally, the preset time is 60 seconds. When the static duration score calculation unit is working, the following formula is satisfied:

[0039] ;

[0040] Wherein, Time1 represents the score of the static duration; t1 represents the static duration, i.e., the duration of the user's mouse being static from the time the user operation interface is displayed to the preset time; the static duration of the user's mouse means that the user has no mouse operation within 3 seconds; n t1 It indicates the duration from the time the user operation interface is displayed to the first time the user's mouse stops moving. The above durations are all in seconds.

[0041] The mouse movement score calculation unit performs calculation based on the mouse movement conditions except for the static time within the preset time, and satisfies the following formula:

[0042] ;

[0043] Among them, Mouse1 represents the mouse movement score; m ref Indicates the mouse movement benchmark score, which is set by the administrator based on experience; L all It indicates the total length of the track of the mouse pointer moved by the user in the user operation interface within the preset time; t1 indicates the static time;

[0044] The user viewing time score calculation unit satisfies the following formula when calculating within the preset time:

[0045] ;

[0046] ;

[0047] Among them, V1 represents the user's viewing time score; v a represents the duration of the user viewing the user operation interface for the ath time within the preset time; A represents the total number of times the user views the user operation interface within the preset time; f(v a ) represents the selection function based on the multiple of duration;

[0048] The interactive demand index integration unit satisfies the following formula when working:

[0049] ;

[0050] Where R represents the interaction demand index; α1, α2 and α3 represent different index integration coefficients, which are set by the administrator according to the actual situation. α1>α2>α3>1. The older the user is, the larger α1, α2 and α3 are. For example, if the user is 20 years old, α1=2, α2=1.5, α3=1.2; if the user is 40 years old, α1=2.5, α2=2, α3=1.5. When R≥r ref When the interaction demand index reaches the standard, the automatic interaction control submodule is used to automatically interact with the user operation interface, automatically retrieve and display the speech generation model parameters that are used most frequently by the corresponding user; ref Indicates the automatic interaction threshold, which is set by the administrator based on experience.

[0051] Optionally, the visualization interface terminal further includes a model parameter confirmation prompt module; the model parameter confirmation prompt module is used to generate corresponding model parameter confirmation prompt information according to the user's operation process after the user completes the model parameter adjustment.

[0052] Optional, combined Figure 3 As shown, the model parameter confirmation prompt module includes a user operation process line of sight information acquisition submodule, a mouse operation record information acquisition submodule, a parameter adjustment time information acquisition submodule, a prompt index calculation submodule and a model parameter confirmation prompt information generation submodule; the user operation process line of sight information acquisition submodule is used to obtain the user operation process line of sight information from the eye tracker; the mouse operation record information acquisition submodule is used to obtain the mouse operation record information when the user adjusts each model parameter adjustment item; the parameter adjustment time information acquisition submodule is used to obtain the parameter adjustment time when the user adjusts each model parameter adjustment item; the prompt index calculation submodule is used to calculate the corresponding prompt index according to the user operation process line of sight information, mouse operation record information and parameter adjustment time; the model parameter confirmation prompt information generation submodule is used to generate the corresponding model parameter confirmation prompt information according to the prompt index.

[0053] When the prompt index calculation submodule is working, the corresponding prompt index is calculated for each model parameter adjustment item in the user operation interface, satisfying the following formula:

[0054] ;

[0055] ;

[0056] ;

[0057] Wherein, C represents the prompt index; Time2 represents the adjustment time of a corresponding model parameter adjustment item, that is, the time spent by the user in adjusting the parameters of the above-mentioned model parameter adjustment item in the user operation interface; λ represents the index value conversion coefficient, which is set by the administrator based on experience; Mouse2 represents the mouse operation score in the mouse operation record information; t2 represents the total time the mouse is in the corresponding model parameter adjustment item area when the user adjusts the parameters of the above-mentioned model parameter adjustment item in the user operation interface in the mouse operation record information; N represents the total number of times the mouse enters the corresponding model parameter adjustment item area when the user adjusts the parameters of the above-mentioned model parameter adjustment item in the user operation interface in the mouse operation record information; Gaze represents the sight score in the sight information of the user operation process; t3 represents the total time the user's sight is in the corresponding model parameter adjustment item area when the user adjusts the parameters of the above-mentioned model parameter adjustment item in the user operation interface in the sight information of the user operation process; L section It represents the moving distance of the sight line in the corresponding model parameter adjustment item area when the user adjusts one of the above model parameter adjustment items in the user operation interface in the sight line information during the user operation process; β1 and β2 represent the mouse operation score conversion coefficient and the sight line score conversion coefficient respectively, both of which are set by the administrator based on experience.

[0058] When C ≥ c ref When the model parameter confirmation prompt information generating submodule generates model parameter confirmation prompt information for indicating that the corresponding model parameter adjustment item needs to be confirmed by the user; c ref Indicates the warning threshold, which is set by the administrator based on experience.

[0059] A method for building a speech generation model based on digital visualization, such as the above-mentioned system for building a speech generation model based on digital visualization, combined with Figure 4 As shown, the speech generation model building method includes:

[0060] S1, data processing and feature extraction of input data;

[0061] S2, providing a user operation interface and reading the model parameters entered by the user;

[0062] S3, constructing a speech generation model according to the model parameters entered by the user;

[0063] S4, generating corresponding voice information according to the processed data and voice features.

[0064] In summary, the above system first realizes data processing and feature extraction through an independent data processing terminal, and then provides a visual interface for users to directly interact, complete the adjustment of different model parameters, and then quickly build an accurate speech generation model. Finally, the customized speech generation model is used to complete the voice information output, which is conducive to improving the adaptability and accuracy of the speech generation model; the efficiency of data processing is improved by the simultaneous processing of voice data processing and text data processing, which is conducive to improving the work efficiency of the system, improving the accuracy of data processing, and then improving the accuracy and construction efficiency of the speech generation model; the comprehensiveness and convenience of the user operation interface are improved by managing various model parameter adjustment items, model effect preview windows and stored data display windows, which is conducive to improving the accuracy and efficiency of speech generation model construction; the corresponding interaction demand index is calculated through the user operation interface static time data, mouse movement data and user viewing interface time data, and then the user operation interface is adjusted according to the interaction demand index. Automatic interactive control is conducive to further improving the intelligence of the system and improving the accuracy and efficiency of building a speech generation model; in conjunction with the stillness time score algorithm, the mouse movement score algorithm and the user viewing time score algorithm, the stillness time score of the corresponding user is calculated through the stillness time data of the user operation interface, the corresponding mouse movement score is calculated through the mouse movement data when the user is operating, and the corresponding user viewing time score is calculated through the user viewing interface time data, and then the stillness time score, the mouse movement score and the user viewing time score are integrated into the corresponding interaction demand index, which is conducive to improving the accuracy of the interaction demand index, thereby improving the accuracy and efficiency of automatic interactive control; in conjunction with the prompt index algorithm, the corresponding prompt index is calculated through the user's operating process line of sight information, mouse operation record information and parameter adjustment time, which is conducive to improving the accuracy of the prompt index, thereby further improving the accuracy of the speech generation model, thereby helping to improve the working efficiency of the system's speech generation model.

[0065] Embodiment 2: This embodiment includes all the contents of embodiment 1, and provides a system for building a voice generation model based on digital visualization, wherein the visualization interface terminal further includes a voice data automatic deletion module; the voice data automatic deletion module is used to generate a corresponding voice data automatic deletion instruction according to the user's location information and user monitoring information after the corresponding voice generation model is built and used. The voice data automatic deletion instruction is used to drive the system to automatically delete the voice data that has been used by the corresponding user.

[0066] Combination Figure 5As shown, the voice data automatic deletion module includes a user monitoring information acquisition submodule, a user location information acquisition submodule, an automatic deletion evaluation index calculation submodule and a voice data automatic deletion instruction generation submodule; the user monitoring information acquisition submodule is used to acquire user monitoring information from a surveillance camera; the user location information acquisition submodule is used to identify and generate user location information of each user based on the user monitoring information; the automatic deletion evaluation index calculation submodule is used to calculate the automatic deletion evaluation index based on the user monitoring information and the user location information; the voice data automatic deletion instruction generation submodule is used to generate a corresponding voice data automatic deletion instruction based on the automatic deletion evaluation index.

[0067] In an application scenario where a user is sitting and using the system and there are other users in the place where the system is located, after the current user completes a speech generation using the system, when the current user stands up from the seat, the automatic deletion evaluation index calculation submodule starts to work and calculates the automatic deletion evaluation index to satisfy the following formula:

[0068] ;

[0069] ;

[0070] Among them, Delete represents the automatic deletion evaluation index; h(NUM) represents the selection function based on the total number of other users in the system location; NUM represents the total number of other users in the system location; L i represents the distance between the i-th user and the system in the place where the system is located when the current user stands up from his seat; L3 represents the relative moving distance between the current user and the system within 3 seconds after he stands up from his seat; when Delete≥d ref When the voice data automatic deletion instruction generation submodule generates a voice data automatic deletion instruction for driving the system to automatically delete the voice data that has been used by the current user; d ref Indicates the assessment threshold, which is set by the administrator based on experience.

[0071] To sum up, the voice data automatic deletion module includes the user monitoring information acquisition submodule, the user location information acquisition submodule, the automatic deletion assessment index calculation submodule and the voice data automatic deletion instruction generation submodule, which are set up in conjunction with the automatic deletion assessment index algorithm. It determines whether the automatic deletion operation is performed through an accurate automatic deletion assessment index, which is conducive to resetting the system for other users in advance, while improving the privacy and security of the current user's data.

[0072] The contents disclosed above are only preferred feasible embodiments of the present invention, and do not limit the protection scope of the present invention. Therefore, all equivalent technical changes made using the contents of the present invention description and drawings are included in the protection scope of the present invention. In addition, the elements therein can be updated as technology develops.

Claims

1. A speech generation model building system based on digital visualization, characterized in that: It includes a data processing terminal, a model building terminal, a visual interface terminal and a speech synthesis terminal; the data processing terminal is used to process and extract features of the input data; the visual interface terminal is used to provide a user operation interface and read the model parameters input by the user; the model building terminal is used to build a speech generation model according to the model parameters input by the user; the speech synthesis terminal is used to control the speech generation module to generate corresponding speech information according to the data and speech features after data processing; The data processing terminal includes a voice data processing module, a text data processing module and a voice feature extraction module; The voice data processing module is used to pre-process the input voice data; the pre-processing includes loading, format conversion and noise reduction; the text data processing module is used to read and recognize the input text data; the voice feature extraction module is used to extract voice features from the pre-processed voice data; The visual interface terminal includes an interface management module, a model parameter reading module and an automatic interaction module; the interface management module is used to provide a user operation interface; the user operation interface includes various model parameter adjustment items, a model effect preview window and a stored data display window; the model parameter reading module is used to read the specific data adjusted by the user in various model parameter adjustment items; the automatic interaction module is used to automatically interact according to the user operation situation to complete the regulation of the user operation interface; The automatic interaction module includes a user operation process monitoring information acquisition submodule, a monitoring information analysis submodule, an interaction demand index calculation submodule and an automatic interaction control submodule; The user operation process monitoring information acquisition submodule is used to acquire user video information from the monitoring camera and user operation record information from the system; The monitoring information analysis submodule is used to analyze user video information and user operation record information to generate corresponding user operation interface stillness time data, mouse movement data and user interface viewing time data; the interaction demand index calculation submodule is used to calculate the corresponding interaction demand index based on the user operation interface stillness time data, mouse movement data and user interface viewing time data; the automatic interaction control submodule is used to automatically interact with the user operation interface according to the interaction demand index, and automatically retrieve and display the voice generation model parameters that are used most frequently by the corresponding user.

2. A speech generation model building system based on digital visualization as claimed in claim 1, characterized in that: The interaction demand index calculation submodule includes a static time score calculation unit, a mouse movement score calculation unit, a user viewing time score calculation unit and an interaction demand index integration unit; The static time length score calculation unit is used to calculate the static time length score of the corresponding user according to the static time data of the user operation interface; The mouse movement score calculation unit is used to calculate the corresponding mouse movement score based on the mouse movement data of the user when performing an operation; the user viewing time score calculation unit is used to calculate the corresponding user viewing time score based on the user viewing interface time data; the interaction demand index integration unit is used to integrate the static time score, the mouse movement score and the user viewing time score into a corresponding interaction demand index.

3. A speech generation model building system based on digital visualization as described in claim 2, characterized in that: The visualization interface terminal also includes a model parameter confirmation prompt module; the model parameter confirmation prompt module is used to generate corresponding model parameter confirmation prompt information according to the user's operation process after the user completes the model parameter adjustment.

4. A speech generation model building system based on digital visualization as described in claim 3, characterized in that: The model parameter confirmation prompt module includes a user operation process sight line information acquisition submodule, a mouse operation record information acquisition submodule, a parameter adjustment time information acquisition submodule, a prompt index calculation submodule and a model parameter confirmation prompt information generation submodule; The user operation process sight line information acquisition submodule is used to acquire the user operation process sight line information from the eye tracker; the mouse operation record information acquisition submodule is used to acquire the mouse operation record information when the user adjusts each model parameter adjustment item; the parameter adjustment time length information acquisition submodule is used to acquire the parameter adjustment time length when the user adjusts each model parameter adjustment item; The prompt index calculation submodule is used to calculate the corresponding prompt index according to the user's line of sight information, mouse operation record information and parameter adjustment time during the operation process; the model parameter confirmation prompt information generation submodule is used to generate the corresponding model parameter confirmation prompt information according to the prompt index.

5. A method for building a speech generation model based on digital visualization, applied to a system for building a speech generation model based on digital visualization as claimed in claim 4, characterized in that: The speech generation model building method comprises: S1, data processing and feature extraction of input data; S2, providing a user operation interface and reading the model parameters entered by the user; S3, constructing a speech generation model according to the model parameters entered by the user; S4, generating corresponding voice information according to the processed data and voice features.

Images