Speech synthesis method, system, terminal and storage medium

Abstract

The application provides a speech synthesis method, system, terminal and storage medium, the method comprising: inputting a text sample into a speech synthesis model to extract phonemes, obtaining text phonemes; phoneme coding is carried out on the text phonemes, obtaining phoneme features, inputting style sample audio into the speech synthesis model to extract features, obtaining emotion features and global features; the phoneme features, emotion features and global features are combined to obtain sample combined features, and the sample combined features are decoded to obtain decoded audio; determine the model loss according to the decoded audio and the standard audio of the text sample, update the parameters of the speech synthesis model according to the model loss; input the text to be synthesized and the target style label into the converged speech synthesis model to synthesize speech, and obtain the target style audio. The speech synthesis model after convergence can effectively synthesize different style speeches, and the diversity of speech synthesis style is improved.

Description

Technical Field

[0001] This invention relates to the field of speech synthesis technology, and in particular to a speech synthesis method, system, terminal, and storage medium. Background Technology

[0002] Speech synthesis is a technology that transforms text information into fluent, understandable spoken output. The speech synthesis process involves first converting text information into linguistic features or phonemes, and then converting those features or phonemes into audio waveforms.

[0003] In existing speech synthesis processes, speech models can only synthesize speech of a single style, resulting in poor diversity of speech synthesis styles. Summary of the Invention

[0004] The purpose of this invention is to provide a speech synthesis method, system, terminal, and storage medium to solve the problem of poor diversity in speech synthesis styles in the prior art.

[0005] The present invention is implemented as follows: a speech synthesis method, the method comprising:

[0006] Obtain audio samples of text and style, and input the text samples into a speech synthesis model to extract phonemes, thereby obtaining text phonemes;

[0007] The text phonemes are phoneme encoded to obtain phoneme features, and the style sample audio is input into the speech synthesis model for feature extraction to obtain emotional features and global features.

[0008] The phoneme features, the emotion features, and the global features are combined to obtain sample combination features, and the sample combination features are decoded to obtain decoded audio.

[0009] The model loss is determined based on the decoded audio and the standard audio of the text sample, and the parameters of the speech synthesis model are updated based on the model loss until the speech synthesis model converges.

[0010] The text to be synthesized and the target style label are input into the converged speech synthesis model to synthesize speech and obtain the target style audio.

[0011] Preferably, the style sample audio is input into the speech synthesis model for feature extraction to obtain emotional features and global features, including:

[0012] The emotional features are obtained by performing feature downsampling on the style sample audio based on the speech synthesis model.

[0013] The global features are obtained by performing attention processing on the style sample audio based on the speech synthesis model.

[0014] Preferably, the style sample audio is downsampled according to the speech synthesis model to obtain the emotional features, including:

[0015] The frequency features and audio features of the style sample audio are obtained according to the speech synthesis model, and the frequency features and audio features are downsampled to obtain frequency sampling features and audio sampling features.

[0016] Attention processing is performed on the frequency sampling features and the audio sampling features to obtain the emotion features.

[0017] Preferably, attention processing is performed on the style sample audio based on the speech synthesis model to obtain the global features, including:

[0018] The style sample audio is convolved according to the speech synthesis model to obtain the first convolution feature, and the first convolution feature is vectorized to obtain the style vector.

[0019] Attention processing is applied to the style vector and the phoneme features to obtain the global features.

[0020] Preferably, the global features are obtained by performing attention processing on the phoneme features, the style vector, and the phoneme features, including:

[0021] The style vector is used as the key matrix and value matrix, and the phoneme features are used as the query matrix for attention processing to obtain the first attention feature;

[0022] The first attention feature is subjected to multiple convolutional processing to obtain the second convolutional feature and the third convolutional feature, and the second convolutional feature and the third convolutional feature are weighted to obtain the global feature.

[0023] Preferably, the text sample is input into a speech synthesis model for phoneme extraction to obtain text phonemes, including:

[0024] The characters in the text sample are converted into phonemes according to the speech synthesis model to obtain character phonemes, and phoneme embedding is performed on the character phonemes to obtain the text phonemes.

[0025] Another objective of this invention is to provide a speech synthesis system, the system comprising:

[0026] The phoneme extraction module is used to acquire audio samples of text and style samples, and input the text samples into the speech synthesis model to extract phonemes and obtain text phonemes.

[0027] The feature extraction module is used to encode the phonemes of the text to obtain phoneme features, and input the style sample audio into the speech synthesis model to extract features to obtain emotion features and global features;

[0028] The feature decoding module is used to combine the phoneme features, the emotion features, and the global features to obtain sample combination features, and to perform feature decoding on the sample combination features to obtain decoded audio;

[0029] The parameter update module is used to determine the model loss based on the decoded audio and the standard audio of the text sample, and update the parameters of the speech synthesis model based on the model loss until the speech synthesis model converges.

[0030] The speech synthesis module is used to input the text to be synthesized and the target style label into the converged speech synthesis model to synthesize speech and obtain the target style audio.

[0031] Preferably, the feature extraction module is further used for:

[0032] The emotional features are obtained by performing feature downsampling on the style sample audio based on the speech synthesis model.

[0033] The global features are obtained by performing attention processing on the style sample audio based on the speech synthesis model.

[0034] In this embodiment of the invention, by inputting text samples into a speech synthesis model for phoneme extraction, text phonemes can be effectively extracted from the text samples. By phoneme encoding of the text phonemes, feature extraction of the text phonemes can be effectively performed to obtain phoneme features. By inputting style sample audio into the speech synthesis model for feature extraction, emotional features and global features in the style sample audio can be effectively extracted. By combining the phoneme features, emotional features, and global features, sample combination features can be effectively obtained. By decoding the sample combination features, the sample combination features can be effectively decoded into audio information to obtain decoded audio. By using the decoded audio and the standard audio of the text samples, the model loss can be effectively determined. The parameters of the speech synthesis model are updated using the model loss, so that the converged speech synthesis model can effectively synthesize speech of different styles, thereby improving the diversity of speech synthesis styles. Attached Figure Description

[0035] Figure 1 This is a flowchart of the speech synthesis method provided in the first embodiment of the present invention;

[0036] Figure 2 This is a schematic diagram of the speech synthesis system provided in the second embodiment of the present invention;

[0037] Figure 3 This is a schematic diagram of the structure of the terminal device provided in the third embodiment of the present invention. Detailed Implementation

[0038] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0039] To illustrate the technical solution described in this invention, specific embodiments are described below.

[0040] Example 1

[0041] Please see Figure 1 This is a flowchart of a speech synthesis method provided in the first embodiment of the present invention. This speech synthesis method can be applied to any device or system, and includes the following steps:

[0042] Step S10: Obtain audio samples of text and style, and input the text samples into the speech synthesis model for phoneme extraction to obtain text phonemes;

[0043] Both text samples and style sample audio can have their text content and audio style set according to requirements. By inputting text samples into the speech synthesis model for phoneme extraction, the phoneme features (text phonemes) of the text samples can be effectively extracted.

[0044] Optionally, the text sample is input into a speech synthesis model for phoneme extraction to obtain text phonemes, including:

[0045] The characters in the text sample are converted into phonemes according to the speech synthesis model to obtain character phonemes, and phoneme embedding processing is performed on the character phonemes to obtain the text phonemes;

[0046] In this process, by converting characters in the text sample into phonemes, the characters can be effectively converted into phonemes, resulting in character phonemes. The character phonemes are then segmented, and the segmented character phonemes are further processed by phoneme embedding to obtain text phonemes.

[0047] Step S20: Phoneme encoding is performed on the text phonemes to obtain phoneme features, and the style sample audio is input into the speech synthesis model for feature extraction to obtain emotional features and global features;

[0048] In this process, text phonemes are input into a phoneme encoder for phoneme encoding to obtain the features of the text phonemes. The phoneme features are then obtained by inputting style sample audio into a speech synthesis model for feature extraction. This effectively extracts emotional features and global features from the style sample audio. Emotional features are used to characterize the emotional information of the style sample audio, while global features are used to characterize the overall audio attributes of the style sample audio.

[0049] Optionally, the style sample audio is input into the speech synthesis model for feature extraction to obtain emotional features and global features, including:

[0050] The style sample audio is downsampled using the speech synthesis model to obtain the emotional features; attention processing is then performed on the style sample audio using the speech synthesis model to obtain the global features.

[0051] Further, the style sample audio is downsampled according to the speech synthesis model to obtain the emotional features, including:

[0052] The frequency features and audio features of the style sample audio are obtained according to the speech synthesis model, and the frequency features and audio features are downsampled to obtain frequency sampling features and audio sampling features.

[0053] Attention processing is performed on the frequency sampling features and the audio sampling features to obtain the emotion features;

[0054] Specifically, by acquiring the frequency features and audio features of the style sample audio separately, feature downsampling is effectively guaranteed. By downsampling the frequency features and audio features separately, the data complexity and computational load of the frequency features and audio features are effectively reduced. By performing attention processing on the frequency sampling features and audio sampling features, the frequency sampling features and audio sampling features can be effectively fused into emotional features, thereby improving the robustness of emotional features.

[0055] Furthermore, attention processing is performed on the style sample audio based on the speech synthesis model to obtain the global features, including:

[0056] The style sample audio is convolved according to the speech synthesis model to obtain the first convolution feature, and the first convolution feature is vectorized to obtain the style vector.

[0057] Attention processing is applied to the style vector and the phoneme features to obtain the global features;

[0058] In this process, by performing vector transformation on the first convolutional feature, the first convolutional feature can be effectively converted into a style vector, which facilitates the subsequent attention processing between the convolutional feature and the phoneme feature. The convolutional kernel of this convolutional processing can be set according to the requirements. By performing attention processing on the style vector and the phoneme feature, the style vector and the phoneme feature can be effectively fused into a global feature, so that the global feature can effectively carry feature information such as style and phoneme.

[0059] Preferably, the global features are obtained by performing attention processing on the phoneme features, the style vector, and the phoneme features, including:

[0060] The style vector is used as the key matrix and value matrix, and the phoneme features are used as the query matrix for attention processing to obtain the first attention feature;

[0061] The first attention feature is subjected to multiple convolutional processing to obtain the second convolutional feature and the third convolutional feature, and the second convolutional feature and the third convolutional feature are weighted to obtain the global feature.

[0062] Step S30: Combine the phoneme features, the emotion features, and the global features to obtain sample combination features, and perform feature decoding on the sample combination features to obtain decoded audio;

[0063] In this method, by combining phoneme features, emotion features, and global features, sample combination features can be effectively obtained. By decoding the sample combination features, the sample combination features can be effectively decoded into audio information, thus obtaining decoded audio and achieving the effect of speech synthesis of text samples.

[0064] Step S40: Determine the model loss based on the decoded audio and the standard audio of the text sample, and update the parameters of the speech synthesis model based on the model loss until the speech synthesis model converges;

[0065] Specifically, the model loss is obtained by calculating the difference between the standard audio and the decoded audio of the text sample. The model loss is then used to update the parameters of the speech synthesis model, so that the converged speech synthesis model can effectively synthesize speech of different styles.

[0066] Step S50: Input the text to be synthesized and the target style label into the converged speech synthesis model to perform speech synthesis and obtain the target style audio.

[0067] The target style label is used to represent the audio style required by the user. By inputting the text to be synthesized and the target style label into the converged speech synthesis model, the text to be synthesized can be effectively converted into audio with the desired audio style.

[0068] Optionally, this embodiment also includes:

[0069] Model training phase:

[0070] The style sample audio is discretized using a trained encoder to obtain a speech representation token. The text sample and the speech representation token are then used as input to train the GPT synthesis model (speech synthesis model).

[0071] Prepare speech prompt libraries with different styles or emotions as needed, and use these speech prompt libraries to fine-tune the trained GPT synthesis model to obtain the fine-tuned GPT synthesis model.

[0072] Speech synthesis stage:

[0073] Obtain the text to be synthesized, along with the required style or emotion type description. Select the corresponding speech from the speech prompt library as the target style label. Input the target style label and the text to be synthesized into the GPT fine-tuning synthesis model for speech synthesis to obtain the target style audio.

[0074] In this embodiment, by inputting text samples into the speech synthesis model for phoneme extraction, text phonemes can be effectively extracted from the text samples. By phoneme encoding of the text phonemes, feature extraction of the text phonemes can be effectively performed to obtain phoneme features. By inputting style sample audio into the speech synthesis model for feature extraction, emotional features and global features in the style sample audio can be effectively extracted. By combining the phoneme features, emotional features, and global features, sample combination features can be effectively obtained. By decoding the sample combination features, the sample combination features can be effectively decoded into audio information to obtain decoded audio. By using the decoded audio and the standard audio of the text samples, the model loss can be effectively determined. The parameters of the speech synthesis model are updated using the model loss, so that the converged speech synthesis model can effectively synthesize speech of different styles, improving the diversity of speech synthesis styles.

[0075] Example 2

[0076] Please see Figure 2 This is a schematic diagram of the structure of the speech synthesis system 100 provided in the second embodiment of the present invention, including:

[0077] The phoneme extraction module 10 is used to acquire text samples and style sample audio, and input the text samples into the speech synthesis model for phoneme extraction to obtain text phonemes. Both the text samples and style sample audio can be customized with different text content and audio styles as needed. By inputting the text samples into the speech synthesis model for phoneme extraction, the phoneme features (text phonemes) of the text samples can be effectively extracted.

[0078] Optionally, the phoneme extraction module 10 is further configured to: convert the characters in the text sample into phonemes according to the speech synthesis model to obtain character phonemes, and perform phoneme embedding processing on the character phonemes to obtain the text phonemes. Specifically, by converting the characters in the text sample into phonemes, the characters can be effectively converted into phonemes, the character phonemes are segmented, and the segmented character phonemes are embedded to obtain the text phonemes.

[0079] The feature extraction module 11 is used to encode the text phonemes to obtain phoneme features, and input the style sample audio into the speech synthesis model for feature extraction to obtain emotional features and global features. Specifically, inputting the text phonemes into the phoneme encoder for phoneme encoding to obtain text phoneme features, and then inputting the style sample audio into the speech synthesis model for feature extraction, effectively extracts emotional features and global features from the style sample audio. Emotional features are used to characterize the emotional information of the style sample audio, and global features are used to characterize the overall audio attributes of the style sample audio.

[0080] Optionally, the feature extraction module 11 is further configured to: perform feature downsampling processing on the style sample audio according to the speech synthesis model to obtain the emotional features;

[0081] The global features are obtained by performing attention processing on the style sample audio based on the speech synthesis model.

[0082] Furthermore, the feature extraction module 11 is also used to: obtain the frequency features and audio features of the style sample audio according to the speech synthesis model, and downsample the frequency features and audio features respectively to obtain frequency sampling features and audio sampling features;

[0083] Attention processing is applied to the frequency sampling features and the audio sampling features to obtain the emotional features. Specifically, by separately acquiring the frequency and audio features of the style sample audio, feature downsampling is effectively ensured. Downsampling the frequency and audio features separately effectively reduces their data complexity and computational load. Attention processing of the frequency and audio sampling features effectively fuses them into emotional features, improving the robustness of the emotional features.

[0084] Furthermore, the feature extraction module 11 is also used to: perform convolution processing on the style sample audio according to the speech synthesis model to obtain a first convolution feature, and perform vector transformation on the first convolution feature to obtain a style vector;

[0085] Attention processing is applied to the style vector and the phoneme features to obtain the global features. Specifically, vector transformation is performed on the first convolutional features, effectively converting them into style vectors, which facilitates subsequent attention processing with the phoneme features. The convolution kernel for this processing can be set according to requirements. By applying attention processing to the style vector and phoneme features, the style vector and phoneme features are effectively fused into global features, ensuring that the global features effectively carry style and phoneme information.

[0086] Preferably, the feature extraction module 11 is further configured to: perform attention processing on the style vector as a key matrix and a value matrix, and the phoneme features as a query matrix, to obtain the first attention feature;

[0087] The first attention feature is subjected to multiple convolutional processing to obtain the second convolutional feature and the third convolutional feature, and the second convolutional feature and the third convolutional feature are weighted to obtain the global feature.

[0088] The feature decoding module 12 is used to combine the phoneme features, the emotion features, and the global features to obtain sample combination features, and to decode the sample combination features to obtain decoded audio. Specifically, by combining the phoneme features, emotion features, and global features, sample combination features can be effectively obtained. By decoding the sample combination features, audio information can be effectively decoded, thus achieving the effect of speech synthesis for text samples.

[0089] The parameter update module 13 is used to determine the model loss based on the decoded audio and the standard audio of the text sample, and to update the parameters of the speech synthesis model based on the model loss until the speech synthesis model converges. Specifically, the model loss is obtained by calculating the difference between the standard audio and the decoded audio of the text sample, and the parameters of the speech synthesis model are updated using the model loss, enabling the converged speech synthesis model to effectively synthesize speech of different styles.

[0090] The speech synthesis module 14 is used to input the text to be synthesized and the target style label into the converged speech synthesis model for speech synthesis to obtain audio in the target style. The target style label represents the audio style desired by the user. By inputting the text to be synthesized and the target style label into the converged speech synthesis model for speech synthesis, the text to be synthesized can be effectively converted into audio in the desired audio style.

[0091] In this embodiment, by inputting text samples into the speech synthesis model for phoneme extraction, text phonemes can be effectively extracted from the text samples. By phoneme encoding of the text phonemes, feature extraction of the text phonemes can be effectively performed to obtain phoneme features. By inputting style sample audio into the speech synthesis model for feature extraction, emotional features and global features in the style sample audio can be effectively extracted. By combining the phoneme features, emotional features, and global features, sample combination features can be effectively obtained. By decoding the sample combination features, the sample combination features can be effectively decoded into audio information to obtain decoded audio. By using the decoded audio and the standard audio of the text samples, the model loss can be effectively determined. The parameters of the speech synthesis model are updated using the model loss, so that the converged speech synthesis model can effectively synthesize speech of different styles, improving the diversity of speech synthesis styles.

[0092] Example 3

[0093] Figure 3 This is a structural block diagram of a terminal device 2 provided in the third embodiment of this application. For example... Figure 3 As shown, the terminal device 2 in this embodiment includes: a processor 20, a memory 21, and a computer program 22 stored in the memory 21 and executable on the processor 20, such as a speech synthesis method program. When the processor 20 executes the computer program 22, it implements the steps in the various embodiments of the speech synthesis methods described above.

[0094] For example, the computer program 22 may be divided into one or more modules, which are stored in the memory 21 and executed by the processor 20 to complete this application. The one or more modules may be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program 22 in the terminal device 2. The terminal device may include, but is not limited to, the processor 20 and the memory 21.

[0095] The processor 20 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0096] The memory 21 can be an internal storage unit of the terminal device 2, such as a hard drive or memory of the terminal device 2. The memory 21 can also be an external storage device of the terminal device 2, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card equipped on the terminal device 2. Furthermore, the memory 21 can include both internal and external storage units of the terminal device 2. The memory 21 is used to store the computer program and other programs and data required by the terminal device. The memory 21 can also be used to temporarily store data that has been output or will be output.

[0097] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0098] If an integrated module is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. This computer-readable storage medium can be non-volatile or volatile. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable storage medium can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the contents of a computer-readable storage medium may be appropriately added to or subtracted from the contents as required by the legislation and patent practice in a jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, a computer-readable storage medium may not include electrical carrier signals and telecommunication signals.

[0099] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A speech synthesis method, characterized in that, The method includes: Obtain audio samples of text and style, and input the text samples into a speech synthesis model to extract phonemes, thereby obtaining text phonemes; The text phonemes are phoneme encoded to obtain phoneme features, and the style sample audio is input into the speech synthesis model for feature extraction to obtain emotional features and global features. The phoneme features, the emotion features, and the global features are combined to obtain sample combination features, and the sample combination features are decoded to obtain decoded audio. The model loss is determined based on the decoded audio and the standard audio of the text sample, and the parameters of the speech synthesis model are updated based on the model loss until the speech synthesis model converges. The text to be synthesized and the target style label are input into the converged speech synthesis model to perform speech synthesis and obtain the target style audio. The style sample audio is input into the speech synthesis model for feature extraction to obtain emotional features and global features, including: The frequency features and audio features of the style sample audio are obtained according to the speech synthesis model, and the frequency features and audio features are downsampled to obtain frequency sampling features and audio sampling features. Attention processing is performed on the frequency sampling features and the audio sampling features to obtain the emotion features; The style sample audio is convolved according to the speech synthesis model to obtain the first convolution feature, and the first convolution feature is vectorized to obtain the style vector. Attention processing is applied to the style vector and the phoneme features to obtain the global features.

2. The speech synthesis method as described in claim 1, characterized in that, The global features are obtained by performing attention processing on the phoneme features, the style vector, and the phoneme features, including: The style vector is used as the key matrix and value matrix, and the phoneme features are used as the query matrix for attention processing to obtain the first attention feature; The first attention feature is subjected to multiple convolutional processing to obtain the second convolutional feature and the third convolutional feature, and the second convolutional feature and the third convolutional feature are weighted to obtain the global feature.

3. The speech synthesis method as described in claim 1, characterized in that, The text sample is input into a speech synthesis model for phoneme extraction to obtain text phonemes, including: The characters in the text sample are converted into phonemes according to the speech synthesis model to obtain character phonemes, and phoneme embedding is performed on the character phonemes to obtain the text phonemes.

4. A speech synthesis system, characterized in that, The system includes: The phoneme extraction module is used to acquire audio samples of text and style samples, and input the text samples into the speech synthesis model to extract phonemes and obtain text phonemes. The feature extraction module is used to encode the text phonemes to obtain phoneme features, and input the style sample audio into the speech synthesis model to extract features to obtain emotion features and global features; The feature decoding module is used to combine the phoneme features, the emotion features, and the global features to obtain sample combination features, and to perform feature decoding on the sample combination features to obtain decoded audio; The parameter update module is used to determine the model loss based on the decoded audio and the standard audio of the text sample, and update the parameters of the speech synthesis model based on the model loss until the speech synthesis model converges. The speech synthesis module is used to input the text to be synthesized and the target style label into the converged speech synthesis model to synthesize speech and obtain the target style audio. The feature extraction module is further configured to: obtain the frequency features and audio features of the style sample audio according to the speech synthesis model, and downsample the frequency features and audio features respectively to obtain frequency sampling features and audio sampling features; Attention processing is performed on the frequency sampling features and the audio sampling features to obtain the emotion features; The style sample audio is convolved according to the speech synthesis model to obtain the first convolution feature, and the first convolution feature is vectorized to obtain the style vector. Attention processing is applied to the style vector and the phoneme features to obtain the global features.

5. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 3.

6. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 3.

Images