Electronic book data processing method, computing device and computer storage medium

Abstract

The invention discloses an electronic book data processing method, computing equipment and a computer storage medium. The method comprises the following steps: acquiring an electronic book text, and inputting the electronic book text into a generative recognition model; generating a dialogue text recognition instruction and a role recognition instruction, and inputting the dialogue text recognition instruction and the role recognition instruction into the generative recognition model; and obtaining at least one dialogue text in the electronic book text output by the generative recognition model and a role corresponding to any dialogue text. According to the scheme, the dialogue text and the corresponding role in the electronic book text are obtained by inputting the dialogue text recognition instruction and the role recognition instruction into the generative recognition model, the recognition efficiency of the dialogue text and the role in the electronic book text is improved, and the labor cost is saved.

Description

Technical Field

[0001] The present invention relates to the technical field of electronic books, and in particular to an electronic book data processing method, a computing device and a computer storage medium. Background Art

[0002] With the continuous development of science and technology and society, e-books are favored by many readers for their energy-saving and environmental protection, large capacity, and rich forms. E-books usually contain a lot of key information, such as character dialogues.

[0003] In the prior art, a manual extraction method is usually adopted to extract character dialogues in an e-book. However, this method is inefficient and has high labor costs. Summary of the invention

[0004] In view of the above problems, the present invention is proposed to provide an electronic book data processing method, a computing device and a computer storage medium that overcome the above problems or at least partially solve the above problems.

[0005] According to one aspect of the present invention, there is provided an electronic book data processing method, comprising:

[0006] Obtaining an electronic book text, and inputting the electronic book text into a generative recognition model;

[0007] Generate a dialogue text recognition instruction and a role recognition instruction, and input the dialogue text recognition instruction and the role recognition instruction into the generative recognition model;

[0008] At least one dialogue text in the electronic book text output by the generative recognition model and a role corresponding to any dialogue text are obtained.

[0009] According to another aspect of the present invention, there is provided a computing device, comprising: a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus;

[0010] The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform the following operations:

[0011] Obtaining an electronic book text, and inputting the electronic book text into a generative recognition model;

[0012] Generate a dialogue text recognition instruction and a role recognition instruction, and input the dialogue text recognition instruction and the role recognition instruction into the generative recognition model;

[0013] At least one dialogue text in the electronic book text output by the generative recognition model and a role corresponding to any dialogue text are obtained.

[0014] According to another aspect of the present invention, a computer storage medium is provided, wherein the storage medium stores at least one executable instruction, and the executable instruction enables a processor to execute operations corresponding to the above-mentioned electronic book data processing method.

[0015] In the electronic book data processing method, computing device and computer storage medium disclosed in the present invention, the electronic book text, dialogue text recognition instructions and role recognition instructions are input into a generative recognition model, and the generative recognition model outputs the dialogue text in the electronic book text and the role corresponding to the dialogue text, thereby improving the recognition efficiency and recognition effect of the dialogue text and the role in the electronic book text and saving labor costs.

[0016] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0018] Figure 1 A schematic flow chart of an electronic book data processing method provided by the first embodiment of the present invention is shown;

[0019] Figure 2 A schematic diagram showing a flow chart of an electronic book data processing method provided by the second embodiment of the present invention;

[0020] Figure 3 A schematic flow chart of an electronic book data processing method provided in Embodiment 3 of the present invention is shown;

[0021] Figure 4 A schematic flow chart of an electronic book data processing method provided by a fourth embodiment of the present invention is shown;

[0022] Figure 5 A schematic flow chart of an electronic book data processing method provided in Embodiment 5 of the present invention is shown;

[0023] Figure 6 A schematic flow chart of an electronic book data processing method provided by Embodiment 6 of the present invention is shown;

[0024] Figure 7A schematic flow chart of an electronic book data processing method provided by Embodiment 7 of the present invention is shown;

[0025] Figure 8 A schematic diagram of the structure of a computing device provided in Embodiment 9 of the present invention is shown. DETAILED DESCRIPTION

[0026] The exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present invention and to enable the scope of the present invention to be fully communicated to those skilled in the art.

[0027] Embodiment 1

[0028] Figure 1 FIG. 1 is a flow chart of an electronic book data processing method provided by Embodiment 1 of the present invention. Figure 1 As shown, the method specifically comprises the following steps:

[0029] Step S110, obtaining the e-book text, and inputting the e-book text into the generative recognition model.

[0030] The acquired e-book text is the text to be processed later, and the e-book text is part or all of the text data in the e-book. In an optional implementation, the acquisition of the e-book text may be achieved by using one or more of the following acquisition methods:

[0031] Acquisition method 1: Acquisition of e-book texts in fixed units. For example, the fixed unit may be chapters, so that the text data of one chapter is acquired each time; or the fixed unit may be N pages of e-book content, so that the text data of N pages is acquired each time; or the fixed unit may be books, so that the e-book text of the entire book is acquired each time, and so on. This acquisition method is simple and easy to implement, and is suitable for large-scale application and implementation.

[0032] Acquisition method 2: dynamically adjust the data volume of the acquired e-book text according to the performance parameters of the generative recognition model. For example, the data volume of the e-book text can be determined according to the current throughput of the generative recognition model. The larger the current throughput, the more data volume of the e-book text is acquired, and the smaller the current throughput, the less data volume of the e-book text is acquired. Alternatively, the data volume of the acquired e-book text can be dynamically adjusted according to the current response time of the generative recognition model. The shorter the current response time, the more data volume of the e-book text is acquired, and the longer the current response time, the less data volume of the e-book text is acquired. This acquisition method can make the data volume of the acquired e-book text compatible with the performance of the current generative recognition model, thereby ensuring the overall efficiency of the e-book data processing.

[0033] Acquisition method three: In response to a text designation operation, obtain the e-book text that matches the text designation operation. Specifically, a text designation portal can be provided to the user, through which the user can input a text identifier that needs to be subsequently processed, and the text data that matches the user input data will be used as the e-book text to be acquired. For example, the user inputs "Chapters 1-3 of Book A" through the portal, then Chapters 1-3 of Book A are the e-book texts that need to be acquired. Using this acquisition method, the corresponding e-book text can be processed in a customized manner according to user needs, thereby improving the flexibility of user operations and meeting the personalized needs of users.

[0034] After obtaining the e-book text, the e-book text is input into the generative recognition model. Among them, the generative recognition model can be a generative large language model (Generative Large Language Mode), such as ChatGPT (Chat Generative Pre-trained Transformer) model, BERT (Bidirectional EncoderRepresentation from Transformers) model, or ERNIE (Enhanced Representation throughKnowledge Integration) model, etc. The generative recognition model is essentially a generative model, which is an AI (Artificial Intelligence) model obtained by training the neural network structure through a large amount of language data. The generative recognition model can understand natural language well and output relevant results that conform to natural language instructions.

[0035] Step S120, generating a dialogue text recognition instruction and a role recognition instruction, and inputting the dialogue text recognition instruction and the role recognition instruction into a generative recognition model.

[0036] The dialogue text recognition instruction is a related instruction for instructing the generative recognition model to recognize the dialogue text in the e-book text. Since the generative recognition model can effectively understand natural language, the dialogue text recognition instruction can specifically be a natural language instruction.

[0037] In an optional embodiment, the dialogue text recognition instruction includes: an instruction for recognizing dialogue text based on preset symbols. The instruction for recognizing dialogue text based on preset symbols is used to instruct the generative recognition model to recognize dialogue text based on preset symbols. The preset symbols can be double quotes, single quotes, and other symbols that have dialogue prompting functions. For example, the instruction for recognizing dialogue text based on preset symbols can be "recognize dialogue within text double quotes" or "treat the content within double quotes as dialogue", so that the generative recognition model treats the data within a pair of double quotes as a dialogue text when processing the e-book text.

[0038] In an optional implementation, the dialogue text recognition instruction includes: an instruction indicating the output range of the dialogue text. The instruction indicating the output range of the dialogue text is used to instruct the generative recognition model to determine the output range of the dialogue text. For example, the instruction indicating the output range of the dialogue text can be "keep the content in each double quote, and do not keep the content outside the double quotes", so that the generative recognition model only outputs the dialogue text in the double quotes; or, the instruction can be a protagonist dialogue text recognition instruction, that is, when there are many characters in the e-book text, in order to save processing resources and improve processing efficiency, only the dialogue text of the main character can be recognized, and the comparison text corresponding to the secondary character can be ignored, etc.

[0039] In an optional implementation, the dialogue text recognition instruction includes: an instruction indicating the output format of the dialogue text. The instruction indicating the output format of the dialogue text is used to instruct the generative recognition model to determine the output format of the dialogue text. For example, the instruction indicating the output format of the dialogue text may be "add &end& at the end of the dialogue", or "new line after each dialogue", or "add new lines before and after multiple double quote dialogues appear in the same line", etc. By constraining the output format of the dialogue text, it is possible to facilitate the subsequent rapid processing of the dialogue text and improve the overall execution efficiency of the method.

[0040] In an optional implementation, the dialogue text recognition instruction includes: an instruction of a processing rule in an abnormal dialogue text recognition state. The instruction of the processing rule in an abnormal dialogue text recognition state is used to instruct the generative recognition model what processing method to use when the dialogue text recognition state is abnormal. For example, the instruction of the processing rule in an abnormal dialogue text recognition state may be "when double quotation marks do not appear in pairs, automatically fill in the double quotation marks on the other side", etc.

[0041] The role identification instruction is an instruction for instructing the generative recognition model to identify the role corresponding to the dialogue file. Since the generative recognition model can effectively understand natural language, the role identification instruction can be a natural language instruction.

[0042] In an optional embodiment, the role identification instruction includes: an instruction for identifying a role based on a preset rule. The instruction for identifying a role based on a preset rule is used to instruct the generative recognition model to use a preset rule to identify the dialogue text. The preset rule may be a semantic rule, that is, the AI ​​model identifies the semantics of the e-book text to determine the role corresponding to the dialogue text; or it may be based on a preset symbol nearest rule, such as taking the named entity appearing before the double quotation marks as the role corresponding to the dialogue text in the double quotation marks, and so on.

[0043] In an optional implementation, the role identification instruction includes: an instruction indicating the role identification range. The instruction indicating the role identification range is used to indicate which dialogue texts the generative recognition model is to perform role identification on. For example, the instruction indicating the role identification range may be "narration does not need to mark the role", etc. Alternatively, the instruction may be a protagonist identification instruction, that is, when there are many roles in the e-book text, in order to save processing resources and improve processing efficiency, only the protagonist may be marked.

[0044] In an optional implementation, the role recognition instruction includes: an instruction indicating a role output format. The instruction indicating the role output format is used to indicate the format in which the generative recognition model outputs the role information. For example, the instruction indicating the role output format may be "marking the corresponding role with [] before each dialogue text".

[0045] In an optional embodiment, the role recognition instruction includes: an instruction for processing rules in the case of abnormal role recognition. The instruction for processing rules in the case of abnormal role recognition is used to instruct the generative model on how to process in the case of abnormal role recognition. For example, the instruction for processing rules in the case of abnormal role recognition may be "for dialogue texts that cannot identify the role, replace with [%%%]", so as to facilitate subsequent manual recognition and other processing for dialogue texts that cannot identify the role.

[0046] The dialogue text recognition instruction and the role recognition instruction are input into the generative recognition model, the generative recognition model executes the dialogue text recognition instruction to obtain at least one dialogue text in the e-book text, and the generative recognition model executes the role recognition instruction to obtain the role corresponding to any dialogue text. The dialogue text recognition instruction and the role recognition instruction can be presented as independent instructions or integrated into a composite instruction, which is not limited in the embodiment of the present invention.

[0047] Step S130, obtaining at least one dialogue text in the electronic book text output by the generative recognition model and a role corresponding to any dialogue text.

[0048] After the dialogue text recognition instructions and the role recognition instructions are input into the generative recognition model, the generative recognition model recognizes the dialogue text in the e-book text and marks the role corresponding to the dialogue text. For example, if a part of the e-book text is "Zhang Fei shouted "Who is coming?"", the output result of the generative recognition model can be "[Zhang Fei] Who is coming", that is, the dialogue text "Who is coming" and the role "Zhang Fei" corresponding to the dialogue text are obtained.

[0049] In an optional implementation, in order to enable the generative large language model to adapt to the e-book data processing scenario in the embodiment of the present invention, the implementation can configure the relevant parameters in the generative large language model. The relevant parameters include but are not limited to: randomness parameter, existence penalty parameter, repetition penalty parameter, maximum number of tokens, and / or sorting parameter, etc. The randomness parameter is a parameter that controls the randomness of the model output, for example, it can include a temperature parameter, a top-p sampling parameter, a top-k sampling parameter, etc. The higher the value of the randomness parameter, the more random the result output by the model. Therefore, in the embodiment of the present invention, the randomness parameter can be set to 0 so that the model can output the result in a stable manner to ensure the stability of the model output; the presence penalty parameter (Presence penalty) and the repetition penalty parameter (Frequency penalty) are parameters for determining whether the model prefers new words. After the presence penalty and repetition penalty functions are turned on, the model will prefer to output new words and omit repeated words. In the embodiment of the present invention, the presence penalty and / or repetition penalty functions are turned off. In this way, even if the same character has multiple dialogue texts, the character can be marked for each of the multiple dialogue texts; the maximum number of tokens (max_tokens) is the maximum number of words in the output result. In the embodiment of the present invention, the maximum number of tokens is set to the maximum value to ensure the integrity of the model output result; in addition, the ranking parameter can be set to 1, that is, the result with the largest probability value is retained, and so on.

[0050] It can be seen that in the e-book data processing method provided by the embodiment of the present invention, the e-book text, dialogue text recognition instructions and role recognition instructions are input into the generative recognition model, and the generative recognition model outputs the dialogue text in the e-book text and the role corresponding to the dialogue text, thereby improving the recognition efficiency of the dialogue text and the role in the e-book text and saving labor costs.

[0051] In the prior art, the e-book audio of the entire e-book uses the same timbre, so that users cannot accurately distinguish the dialogues of different characters in the e-book, which affects the user's e-book reading speed and reading accuracy, resulting in poor user experience, and reduces the user usage of the audiobook function, and affects the user retention rate of the e-book platform. In order to solve at least one of the above problems in the prior art, the present invention identifies the characters in the e-book and the dialogue text data corresponding to the characters through a generative recognition model, and assigns different timbres to different characters, so that users can accurately distinguish the dialogues of different characters in the process of listening to the book, thereby improving the user experience and improving the user's reading efficiency and reading accuracy.

[0052] The following describes in detail the implementation process of generating electronic book audio by identifying the characters in the electronic book and the dialogue texts corresponding to the characters according to the generative recognition model through Examples 2, 3, 4, 5 and 6.

[0053] Embodiment 2

[0054] Figure 2 FIG. 2 is a flow chart of an electronic book data processing method provided by Embodiment 2 of the present invention. Figure 2 As shown, the method specifically comprises the following steps:

[0055] Step S210, obtaining the e-book text, and inputting the e-book text into the generative recognition model.

[0056] Step S220, generating a dialogue text recognition instruction and a role recognition instruction, and inputting the dialogue text recognition instruction and the role recognition instruction into a generative recognition model.

[0057] Step S230, obtaining at least one dialogue text in the electronic book text output by the generative recognition model and a role corresponding to any dialogue text.

[0058] Step S240, determining the timbre that matches any character, and converting the dialogue text corresponding to the character into dialogue audio presented in the matching timbre.

[0059] For any character identified in the e-book text, a timbre matching the character is determined, that is, different timbres can be matched for different characters. For example, the character "Zhang Fei" can be matched with a "rough male voice" timbre, the character "Xiao Qiao" can be matched with a "gentle female voice" timbre, and so on.

[0060] After determining the timbre matching the role, the TTS (Text to Speech) module can be called to convert the dialogue text corresponding to the role into the corresponding dialogue audio, and the dialogue audio is presented with the timbre matching the role, so that the dialogues of different roles can be played with different timbres. In the specific conversion process, the phoneme library corresponding to the matching timbre can be determined, and the dialogue text is converted into dialogue audio based on the phonemes in the phoneme library.

[0061] Step S250, generating an e-book audio corresponding to the e-book text according to the dialogue audio corresponding to the character in the e-book text.

[0062] The dialogue audios corresponding to the characters in the e-book text are synthesized to obtain the e-book audio corresponding to the e-book text, that is, the e-book audio contains the dialogue audios corresponding to the characters in the e-book text.

[0063] Specifically, for other texts other than the recognized dialogue text (such as environment descriptions, dialogue texts of minor characters, etc.), background audio presented in the default timbre can be generated, and the background audio and each dialogue audio are merged to obtain the e-book audio.

[0064] It can be seen that in the e-book data processing method provided by the embodiment of the present invention, after the dialogue text and the characters are extracted based on the generative recognition model, different timbres are matched for different characters, and the dialogue text corresponding to each character is converted into dialogue audio presented with the matched timbre, and the e-book audio corresponding to the e-book text is generated according to the dialogue audio corresponding to the character in the e-book text, so that the dialogues of different characters are presented in the e-book audio with different timbres, which makes it easier for users to distinguish the dialogues of different characters, improves the user's reading experience and improves reading efficiency, and is conducive to improving the user retention rate of the e-book platform.

[0065] Embodiment 3

[0066] Figure 3 FIG. 1 is a flow chart of an electronic book data processing method provided by Embodiment 3 of the present invention. Figure 3 As shown, the method specifically comprises the following steps:

[0067] Step S310, obtaining the e-book text, and inputting the e-book text into the generative recognition model.

[0068] Step S320, generate dialogue text recognition instructions, role recognition instructions, and text position marking instructions, input the dialogue text recognition instructions, role recognition instructions, and text position marking instructions into the generative recognition model, and obtain at least one dialogue text in the e-book text output by the generative recognition model, the role corresponding to any dialogue text, and the position information of the dialogue text.

[0069] In this embodiment, in addition to generating the dialogue text recognition instruction and the role recognition instruction, a text position marking instruction is also generated. The text position marking instruction is used to instruct the generative recognition model to mark the position of the comparison text in the e-book text. For example, the text position marking instruction can be "mark the string index of this dialogue text in the original text after &end&", and this string index can indicate which character in the original text the first character in the dialogue text corresponds to, etc.

[0070] In addition to outputting at least one dialogue text in the e-book text and the role corresponding to any dialogue text, the generative recognition model further outputs the position information of the dialogue text. For example, the output data can be "[Zhang Fei] Who is coming &end& 283", so that it can be determined that the first character of the dialogue text "Who is coming" is the 283rd character in the e-book text, which is convenient for realizing the positioning of the dialogue text in the original text.

[0071] Step S330, determine the tone color matching any role, and convert the dialogue text corresponding to this role into dialogue audio presented in this matching tone color.

[0072] Step S340, generate e-book audio corresponding to the e-book text according to the dialogue audio corresponding to the role in the e-book text and the position information of the dialogue text.

[0073] Specifically, according to the position information of the dialogue text, the position of the dialogue text in the e-book text can be determined, and then the insertion position of the dialogue audio corresponding to this comparison file in the e-book audio can be determined, and the corresponding dialogue audio is inserted at this insertion position to achieve the precise splicing of different dialogue audios and / or background audios, and improve the accuracy of the generated e-book audio.

[0074] It can be seen that in the e-book data processing method provided by the embodiment of the present invention, a text position marking instruction is also generated, so that the position information of the dialogue text can be marked by the generative recognition model, thereby facilitating the accurate insertion of the dialogue audio corresponding to the comparison text into the corresponding position of the e-book audio, and improving the accuracy of the generated e-book audio.

[0075] Embodiment Four

[0076] Figure 4 shows a schematic flowchart of an e-book data processing method provided by the fourth embodiment of the present invention. As Figure 4 shown, this method specifically includes the following steps:

[0077] Step S410, obtain the e-book text, and input this e-book text into the generative recognition model.

[0078] Step S420, generate dialogue text recognition instructions, role recognition instructions, and role portrait generation instructions, input the dialogue text recognition instructions, role recognition instructions, and role portrait generation instructions into the generative recognition model, and obtain at least one dialogue text in the e-book text output by the generative recognition model, the role corresponding to any dialogue text, and the role portrait information of any role.

[0079] In addition to the dialogue text recognition instructions and the role recognition instructions, this embodiment also generates a role portrait generation instruction. The role portrait generation instruction is used to instruct the generative recognition model to generate role portrait information of the role, and the role portrait information may include the gender, age, occupation, personality type, etc. of the role. Among them, because the generative recognition model has a strong semantic analysis ability, the role portrait information output by the model has a high accuracy.

[0080] In addition to outputting at least one dialogue text in the e-book text and the character corresponding to any dialogue text, the generative recognition model further outputs the character portrait information of any character. For example, the character portrait information outputted for the character "Zhang Fei" may be "male-30-military general-rough", etc.

[0081] Step S430, based on the character portrait information of any character, determine the timbre that matches the character, and convert the dialogue text corresponding to the character into dialogue audio presented in the matching timbre.

[0082] Determining the timbre that matches the character based on the character portrait information can improve the compatibility of the timbre of the dialogue audio with the character and enhance the user experience.

[0083] In a specific implementation process, the timbre library pre-stores relevant information of at least one candidate timbre (such as phoneme information, etc.), and each candidate timbre has a timbre label matching it, which is used to describe the profile of the timbre, for example, the timbre label can be "rough male voice", "sweet female voice", "childish child voice", etc. When determining the timbre matching the role, the timbre label of the candidate timbre can be obtained.

[0084] Further, for any character, the similarity between the timbre label of each candidate timbre and the character portrait information of the character is calculated, and the timbre matching the character is selected from the candidate timbre according to the similarity. Specifically, the text similarity between the timbre label and the character portrait information can be calculated respectively, and the candidate timbre corresponding to the timbre label with the highest text similarity is used as the timbre matching the character. And the dialogue text corresponding to the character is converted into dialogue audio presented with the matching timbre.

[0085] Step S440, generating an e-book audio corresponding to the e-book text according to the dialogue audio corresponding to the character in the e-book text.

[0086] It can be seen that in the e-book data processing method provided by the embodiment of the present invention, in addition to generating dialogue text recognition instructions and character recognition instructions, character portrait generation instructions are also generated, so that the generative recognition model can output the character portrait information of the character, and based on the character portrait information, a timbre that matches the character highly can be obtained, thereby improving the matching degree between the dialogue audio and the character and improving the user experience.

[0087] Embodiment 5

[0088] Figure 5 FIG. 5 is a flow chart showing a method for processing electronic book data provided by Embodiment 5 of the present invention. Figure 5 As shown, the method specifically comprises the following steps:

[0089] Step S510, obtaining the e-book text, and inputting the e-book text into the generative recognition model.

[0090] Step S520, generate a dialogue text recognition instruction and a role recognition instruction, input the dialogue text recognition instruction and the role recognition instruction into the generative recognition model, and obtain at least one dialogue text in the e-book text output by the generative recognition model and a role corresponding to any dialogue text.

[0091] Step S530, in response to the timbre configuration operation for any character, determine the timbre that matches the character, and convert the dialogue text corresponding to the character into dialogue audio presented with the matching timbre.

[0092] In this embodiment, a timbre configuration portal may be provided for relevant personnel, through which a timbre may be specified for a role.

[0093] For example, in the TTS tool or server, each character identified in the e-book text is displayed, and in response to a trigger operation for any character, the timbre configuration page of the character is displayed. A timbre selection entry is provided in the timbre configuration page, for example, timbre label options such as "rough male voice" and "sweet female voice" can be presented. When a timbre label option is selected, the timbre corresponding to the option is used as the timbre matching the character.

[0094] After the timbre matching the character is determined, the timbre label is assigned to each dialogue text corresponding to the character, so that when generating dialogue audio, the timbre of the dialogue text to be converted can be quickly determined according to the timbre label, thereby improving the efficiency of generating dialogue audio. In addition, for some onomatopoeias that cannot be determined for a character, onomatopoeia audio matching the onomatopoeia can be generated, and so on.

[0095] Step S540, generating an e-book audio corresponding to the e-book text according to the dialogue audio corresponding to the character in the e-book text.

[0096] It can be seen that in the e-book data processing method provided by the embodiment of the present invention, a tone configuration entry can also be provided to determine the tone that matches the character according to the tone configuration operation, thereby improving the flexibility of the solution and the accuracy of determining the tone that matches the character, further improving the user experience.

[0097] Embodiment 6

[0098] Figure 6 FIG. 6 is a flow chart showing a method for processing electronic book data provided by Embodiment 6 of the present invention. Figure 6 As shown, the method specifically comprises the following steps:

[0099] Step S610, obtaining the e-book text, and inputting the e-book text into the generative recognition model.

[0100] Step S620, generate dialogue text recognition instructions, role recognition instructions, and scene recognition instructions, input the dialogue text recognition instructions, role recognition instructions, and scene recognition instructions into the generative recognition model, and obtain at least one dialogue text in the e-book text output by the generative recognition model, the role corresponding to any dialogue text, and the scene information corresponding to any dialogue text.

[0101] In addition to the generated dialogue text recognition instructions and role recognition instructions, this embodiment also generates a scene recognition instruction. The scene recognition instruction is used to instruct the generative recognition model to generate scene information corresponding to the dialogue text. The scene information includes: weather information, and / or location information, etc. Among them, since the generative recognition model has a strong semantic analysis ability, the scene information output by the model has a high accuracy. This embodiment does not limit the specific way in which the model outputs the scene information. For example, the scene information can be obtained according to the weather and location keyword recognition method.

[0102] In addition to outputting at least one dialogue text in the e-book text and the character corresponding to any dialogue text, the generative recognition model further outputs the scene information corresponding to any dialogue text. For example, if a part of the e-book text is "On a night of thunder and lightning, Zhang San screamed "Why" in the playground", then the scene information of the dialogue text "Why" of the character "Zhang San" is "Thunderstorm-Playground".

[0103] Step S630, determining the timbre that matches any character, and converting the dialogue text corresponding to the character into dialogue audio presented in the matching timbre.

[0104] Step S640, generating scene audio corresponding to the dialogue text according to the scene information corresponding to the dialogue text.

[0105] For any scene information, search for scene audio matching the scene information from the preset audio library. For example, if the scene information is "thunderstorm", obtain thunder audio from the audio library; if the scene information is "wind sound", obtain wind sound audio from the audio library, and so on. And according to the correspondence between the scene information and the dialogue text, determine the correspondence between the scene audio and the dialogue text.

[0106] Step S650, generating an e-book audio corresponding to the e-book text according to the dialogue audio corresponding to the character in the e-book text and the scene audio corresponding to the dialogue text.

[0107] Specifically, according to the correspondence between the dialogue text and the scene audio and the position information of the dialogue text, the insertion position of the scene audio in the e-book audio can be determined, and then the scene audio can be inserted into the corresponding position. Among them, the insertion position of the dialogue audio can be consistent with the insertion position of the corresponding scene audio, so that the dialogue audio can be superimposed with the corresponding scene audio to simulate the dialogue effect in the corresponding scene. For example, the scene information of the dialogue text "Why" of the character "Zhang San" is "Thunderstorm-Playground", then the scene audio of thunder sound can be superimposed in the dialogue audio of "Why", so as to simulate the effect of dialogue in a thunderstorm and enhance the user's reading experience.

[0108] It can be seen that in the e-book data processing method provided by the embodiment of the present invention, in addition to generating dialogue text recognition instructions and character recognition instructions, scene recognition instructions are also generated, so that the generative recognition model can output scene information corresponding to the dialogue text, generate corresponding scene audio based on the scene information, and insert the scene audio into the corresponding position of the e-book audio, so that the dialogue audio can be superimposed with the corresponding scene audio, so as to achieve the effect of dialogue in the corresponding scene, enhance the vividness of the e-book audio, and enhance the user experience.

[0109] Embodiment 7

[0110] Figure 7 FIG. 1 is a flow chart of an electronic book data processing method provided by Embodiment 7 of the present invention. Figure 7 As shown, the method specifically comprises the following steps:

[0111] Step S710, obtaining an electronic book text, and inputting the electronic book text into a generative recognition model.

[0112] Step S720, generating a dialogue text recognition instruction and a role recognition instruction, and inputting the dialogue text recognition instruction and the role recognition instruction into a generative recognition model.

[0113] Step S730, obtaining at least one dialogue text in the electronic book text output by the generative recognition model and a role corresponding to any dialogue text.

[0114] Step S740, generating a dialogue novel according to at least one dialogue text in the electronic book text and a character corresponding to any dialogue text.

[0115] Dialogue novels are a type of novel written with dialogue between characters as the basic structure and form of expression. They narrate events, unfold plots, explain the environment, and portray characters through dialogue between characters. They have a novel structure, a strong sense of reader involvement, and are suitable for fragmented time reading, which makes them popular among some users.

[0116] The existing method of generating dialogue novels is manual conversion, which involves manually analyzing the contents of other forms of e-books and converting them into corresponding dialogue novels. However, this method has high labor costs and low conversion efficiency, and is not suitable for large-scale application and implementation.

[0117] After obtaining at least one dialogue text in the electronic book text and a character corresponding to any dialogue text based on a generative recognition model, the embodiment of the present invention converts the character and the corresponding dialogue text into a corresponding dialogue novel according to the dialogue novel format.

[0118] It can be seen that in the e-book data processing method provided by the embodiment of the present invention, after extracting the dialogue text and characters based on the generative recognition model, a dialogue novel is generated according to the characters and the corresponding dialogue text, which is beneficial to improving the generation efficiency of the dialogue novel and saving labor costs.

[0119] Embodiment 8

[0120] Embodiment 8 of the present invention provides a non-volatile computer storage medium, wherein the computer storage medium stores at least one executable instruction, and the computer executable instruction can execute the electronic book data processing method in any of the above method embodiments.

[0121] The executable instructions may be specifically used to cause the processor to perform the following operations:

[0122] Obtaining an electronic book text, and inputting the electronic book text into a generative recognition model;

[0123] Generate a dialogue text recognition instruction and a role recognition instruction, and input the dialogue text recognition instruction and the role recognition instruction into the generative recognition model;

[0124] At least one dialogue text in the electronic book text output by the generative recognition model and a role corresponding to any dialogue text are obtained.

[0125] In an optional implementation, the executable instructions may be specifically used to cause the processor to perform the following operations:

[0126] Determine a timbre that matches any character, and convert the dialogue text corresponding to the character into dialogue audio presented in the matching timbre;

[0127] According to the dialogue audio corresponding to the character in the e-book text, an e-book audio corresponding to the e-book text is generated.

[0128] In an optional implementation, the executable instruction causes the processor to perform the following operations:

[0129] Generate a text position marking instruction, input the text position marking instruction into the generative recognition model, and obtain the position information of the dialogue text output by the generative recognition model;

[0130] The electronic book audio corresponding to the electronic book text is generated according to the dialogue audio corresponding to the character in the electronic book text and the position information of the dialogue text.

[0131] In an optional implementation, the executable instruction causes the processor to perform the following operations:

[0132] Generate a character portrait generation instruction, input the character portrait generation instruction into the generative recognition model, and obtain character portrait information of any character output by the generative recognition model;

[0133] According to the character portrait information of any character, determine the timbre that matches the character.

[0134] In an optional implementation, the executable instruction causes the processor to perform the following operations:

[0135] Get the timbre label of the candidate timbre;

[0136] For any character, the similarity between the timbre label of each candidate timbre and the character portrait information of the character is calculated, and the timbre matching the character is selected from the candidate timbre according to the similarity.

[0137] In an optional implementation, the executable instruction causes the processor to perform the following operations:

[0138] In response to a timbre configuration operation for any character, a timbre matching the character is determined.

[0139] In an optional implementation, the executable instruction causes the processor to perform the following operations:

[0140] Generate a scene recognition instruction, input the scene recognition instruction into the generative recognition model, and obtain scene information corresponding to the dialogue text output by the generative recognition model;

[0141] Generate scene audio corresponding to the dialogue text according to the scene information;

[0142] The electronic book audio corresponding to the electronic book text is generated according to the dialogue audio corresponding to the character in the electronic book text and the scene audio corresponding to the dialogue text.

[0143] In an optional implementation, the executable instruction causes the processor to perform the following operations:

[0144] A dialogue novel is generated according to at least one dialogue text in the electronic book text and a character corresponding to any dialogue text.

[0145] In an optional implementation, the dialogue text recognition instruction includes:

[0146] Instructions for recognizing conversation text based on preset symbols;

[0147] and / or, instructions indicating the output range of the dialogue text;

[0148] and / or, instructions indicating the output format of the dialogue text;

[0149] And / or, the dialogue text identifies instructions for processing rules under abnormal conditions.

[0150] In an optional implementation, the role identification instruction includes:

[0151] Instructions for identifying roles based on preset rules;

[0152] and / or, instructions indicating the scope of character identification;

[0153] and / or, instructions indicating the output format of the character;

[0154] And / or, instructions for processing rules under role recognition abnormal conditions.

[0155] It can be seen that in the storage medium provided by the embodiment of the present invention, the electronic book text, dialogue text recognition instructions and role recognition instructions are input into the generative recognition model, and the generative recognition model outputs the dialogue text in the electronic book text and the role corresponding to the dialogue text, thereby improving the recognition efficiency of the dialogue text and the role in the electronic book text and saving labor costs.

[0156] Embodiment 9

[0157] Figure 8The figure shows a schematic diagram of the structure of a computing device provided by the ninth embodiment of the present invention. The specific embodiment of the present invention does not limit the specific implementation of the computing device.

[0158] like Figure 8 As shown, the computing device may include: a processor (processor) 802 , a communication interface (Communications Interface) 804 , a memory (memory) 806 , and a communication bus 808 .

[0159] The processor 802, the communication interface 804, and the memory 806 communicate with each other via a communication bus 808. The communication interface 804 is used to communicate with other devices such as a client or other server network elements. The processor 802 is used to execute a program 810, which can specifically execute the relevant steps in the above-mentioned embodiment of the method for processing electronic book data.

[0160] Specifically, the program 810 may include program codes, which include computer operation instructions.

[0161] The processor 802 may be a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention. The one or more processors included in the computing device may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.

[0162] The memory 806 is used to store the program 810. The memory 806 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0163] The program 810 may be specifically configured to enable the processor 802 to perform the following operations:

[0164] Obtaining an electronic book text, and inputting the electronic book text into a generative recognition model;

[0165] Generate a dialogue text recognition instruction and a role recognition instruction, and input the dialogue text recognition instruction and the role recognition instruction into the generative recognition model;

[0166] At least one dialogue text in the electronic book text output by the generative recognition model and a role corresponding to any dialogue text are obtained.

[0167] In an optional implementation, the program 810 may be specifically configured to enable the processor 802 to perform the following operations:

[0168] Determine a timbre that matches any character, and convert the dialogue text corresponding to the character into dialogue audio presented in the matching timbre;

[0169] According to the dialogue audio corresponding to the character in the e-book text, an e-book audio corresponding to the e-book text is generated.

[0170] In an optional implementation, the program 810 may be specifically configured to enable the processor 802 to perform the following operations:

[0171] Generate a text position marking instruction, input the text position marking instruction into the generative recognition model, and obtain the position information of the dialogue text output by the generative recognition model;

[0172] The electronic book audio corresponding to the electronic book text is generated according to the dialogue audio corresponding to the character in the electronic book text and the position information of the dialogue text.

[0173] In an optional implementation, the program 810 may be specifically configured to enable the processor 802 to perform the following operations:

[0174] Generate a character portrait generation instruction, input the character portrait generation instruction into the generative recognition model, and obtain character portrait information of any character output by the generative recognition model;

[0175] According to the character portrait information of any character, determine the timbre that matches the character.

[0176] In an optional implementation, the program 810 may be specifically configured to enable the processor 802 to perform the following operations:

[0177] Get the timbre label of the candidate timbre;

[0178] For any character, the similarity between the timbre label of each candidate timbre and the character portrait information of the character is calculated, and the timbre matching the character is selected from the candidate timbre according to the similarity.

[0179] In an optional implementation, the program 810 may be specifically configured to enable the processor 802 to perform the following operations:

[0180] In response to a timbre configuration operation for any character, a timbre matching the character is determined.

[0181] In an optional implementation, the program 810 may be specifically configured to enable the processor 802 to perform the following operations:

[0182] Generate a scene recognition instruction, input the scene recognition instruction into the generative recognition model, and obtain scene information corresponding to the dialogue text output by the generative recognition model;

[0183] Generate scene audio corresponding to the dialogue text according to the scene information;

[0184] The electronic book audio corresponding to the electronic book text is generated according to the dialogue audio corresponding to the character in the electronic book text and the scene audio corresponding to the dialogue text.

[0185] In an optional implementation, the program 810 may be specifically configured to enable the processor 802 to perform the following operations:

[0186] A dialogue novel is generated according to at least one dialogue text in the electronic book text and a character corresponding to any dialogue text.

[0187] In an optional implementation, the dialogue text recognition instruction includes:

[0188] Instructions for recognizing conversation text based on preset symbols;

[0189] and / or, instructions indicating the output range of the dialogue text;

[0190] and / or, instructions indicating the output format of the dialogue text;

[0191] And / or, the dialogue text identifies instructions for processing rules under abnormal conditions.

[0192] In an optional implementation, the role identification instruction includes:

[0193] Instructions for identifying roles based on preset rules;

[0194] and / or, instructions indicating the scope of character identification;

[0195] and / or, instructions indicating the output format of the character;

[0196] And / or, instructions for processing rules under role recognition abnormal conditions.

[0197] It can be seen that in the computer storage medium provided by the embodiment of the present invention, the electronic book text, dialogue text recognition instructions and role recognition instructions are input into the generative recognition model, and the generative recognition model outputs the dialogue text in the electronic book text and the role corresponding to the dialogue text, thereby improving the recognition accuracy of the dialogue text and the role in the electronic book text; and different timbres are matched for different roles, and the dialogue text corresponding to each role is converted into dialogue audio presented with the matched timbre, and based on the dialogue audio corresponding to the role in the electronic book text, the electronic book audio corresponding to the electronic book text is generated, so that the dialogues of different roles are presented in the electronic book audio through different timbres, which is convenient for users to distinguish the dialogues of different roles, improves the user's reading experience and improves reading efficiency, and is conducive to improving the user retention rate of the electronic book platform.

[0198] The algorithm or display provided herein is not inherently related to any particular computer, virtual system or other equipment. Various general purpose systems can also be used together with the teachings based on this. According to the above description, it is obvious to construct the structure required for this type of system. In addition, the embodiment of the present invention is not directed to any specific programming language yet. It should be understood that various programming languages ​​can be utilized to realize the content of the present invention described herein, and the description made to specific languages ​​above is for disclosing the best mode of the present invention.

[0199] In the description provided herein, a large number of specific details are described. However, it is understood that embodiments of the present invention can be practiced without these specific details. In some instances, well-known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this description.

[0200] Similarly, it should be understood that in order to streamline the present invention and aid in understanding one or more of the various inventive aspects, in the above description of exemplary embodiments of the present invention, the various features of the embodiments of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, this disclosed method should not be interpreted as reflecting the following intention: that the claimed invention requires more features than the features explicitly recited in each claim. More specifically, as reflected in the claims below, inventive aspects lie in less than all the features of the individual embodiments disclosed above. Therefore, the claims that follow the specific embodiment are hereby expressly incorporated into the specific embodiment, with each claim itself serving as a separate embodiment of the present invention.

[0201] Those skilled in the art will appreciate that the modules in the devices in the embodiments may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments may be combined into one module or unit or component, and in addition they may be divided into a plurality of submodules or subunits or subcomponents. Except that at least some of such features and / or processes or units are mutually exclusive, all features disclosed in this specification (including the accompanying claims, abstracts and drawings) and all processes or units of any method or device disclosed in this manner may be combined in any combination. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstracts and drawings) may be replaced by an alternative feature providing the same, equivalent or similar purpose.

[0202] In addition, those skilled in the art will appreciate that, although some embodiments herein include certain features included in other embodiments but not other features, the combination of features of different embodiments is meant to be within the scope of the present invention and form different embodiments. For example, in the claims below, any one of the claimed embodiments may be used in any combination.

[0203] The various component embodiments of the present invention can be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It should be understood by those skilled in the art that a microprocessor or digital signal processor (DSP) can be used in practice to implement some or all functions of some or all components according to embodiments of the present invention. The present invention can also be implemented as a device or apparatus program (e.g., computer program and computer program product) for executing part or all of the methods described herein. Such a program implementing the present invention can be stored on a computer-readable medium, or can have the form of one or more signals. Such a signal can be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0204] It should be noted that the above embodiments illustrate the present invention rather than limit it, and that those skilled in the art may devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference symbol between brackets shall not be construed as a limitation on the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "one" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention may be implemented by means of hardware comprising a number of different elements and by means of a suitably programmed computer. In a unit claim enumerating a number of devices, several of these devices may be embodied by the same hardware item. The use of the words first, second, and third, etc. does not indicate any order. These words may be interpreted as names. The steps in the above embodiments, unless otherwise specified, should not be understood as limitations on the order of execution.

[0205] The present invention discloses: A1. An electronic book data processing method, comprising:

[0206] Obtaining an electronic book text, and inputting the electronic book text into a generative recognition model;

[0207] Generate a dialogue text recognition instruction and a role recognition instruction, and input the dialogue text recognition instruction and the role recognition instruction into the generative recognition model;

[0208] At least one dialogue text in the electronic book text output by the generative recognition model and a role corresponding to any dialogue text are obtained.

[0209] A2. The method according to A1, wherein the method further comprises:

[0210] Determine a timbre that matches any character, and convert the dialogue text corresponding to the character into dialogue audio presented in the matching timbre;

[0211] According to the dialogue audio corresponding to the character in the e-book text, an e-book audio corresponding to the e-book text is generated.

[0212] A3. The method according to A2, wherein the method further comprises: generating a text position marking instruction, inputting the text position marking instruction into the generative recognition model, and obtaining the position information of the dialogue text output by the generative recognition model;

[0213] Generating the e-book audio corresponding to the e-book text according to the dialogue audio corresponding to the character in the e-book text further includes: generating the e-book audio corresponding to the e-book text according to the dialogue audio corresponding to the character in the e-book text and the position information of the dialogue text.

[0214] A4. The method according to A2 or A3, wherein the method further comprises: generating a character portrait generation instruction, inputting the character portrait generation instruction into the generative recognition model, and obtaining the character portrait information of any character output by the generative recognition model;

[0215] The step of determining the timbre that matches any character further includes: determining the timbre that matches any character based on character portrait information of the character.

[0216] A5. The method according to A4, wherein the step of determining the timbre matching the character according to the character portrait information of any character further comprises:

[0217] Get the timbre label of the candidate timbre;

[0218] For any character, the similarity between the timbre label of each candidate timbre and the character portrait information of the character is calculated, and the timbre matching the character is selected from the candidate timbre according to the similarity.

[0219] A6. The method according to A2 or A3, wherein determining the timbre that matches any character further comprises: determining the timbre that matches any character in response to a timbre configuration operation for the character.

[0220] A7. The method according to any one of A2 to A6, wherein the method further comprises: generating a scene recognition instruction, inputting the scene recognition instruction into the generative recognition model, obtaining scene information corresponding to the dialogue text output by the generative recognition model, and generating scene audio corresponding to the dialogue text according to the scene information;

[0221] Generating the e-book audio corresponding to the e-book text according to the dialogue audio corresponding to the character in the e-book text further includes: generating the e-book audio corresponding to the e-book text according to the dialogue audio corresponding to the character in the e-book text and the scene audio corresponding to the dialogue text.

[0222] A8. The method according to A1, wherein the method further comprises:

[0223] A dialogue novel is generated according to at least one dialogue text in the electronic book text and a character corresponding to any dialogue text.

[0224] A9. A method according to any one of A1-A8, wherein the dialogue text recognition instruction comprises:

[0225] Instructions for recognizing conversation text based on preset symbols;

[0226] and / or, instructions indicating the output range of the dialogue text;

[0227] and / or, instructions indicating the output format of the dialogue text;

[0228] And / or, the dialogue text identifies instructions for processing rules under abnormal conditions.

[0229] A10. A method according to any one of A1-A9, wherein the role identification instruction comprises:

[0230] Instructions for identifying roles based on preset rules;

[0231] and / or, instructions indicating the scope of character identification;

[0232] and / or, instructions indicating the output format of the character;

[0233] And / or, instructions for processing rules under role recognition abnormal conditions.

[0234] The present invention also discloses: B11. A computing device, comprising: a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus;

[0235] The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform the following operations:

[0236] Obtaining an electronic book text, and inputting the electronic book text into a generative recognition model;

[0237] Generate a dialogue text recognition instruction and a role recognition instruction, and input the dialogue text recognition instruction and the role recognition instruction into the generative recognition model;

[0238] At least one dialogue text in the electronic book text output by the generative recognition model and a role corresponding to any dialogue text are obtained.

[0239] B12. The computing device according to B11, wherein the executable instructions cause the processor to perform the following operations:

[0240] Determine a timbre that matches any character, and convert the dialogue text corresponding to the character into dialogue audio presented in the matching timbre;

[0241] According to the dialogue audio corresponding to the character in the e-book text, an e-book audio corresponding to the e-book text is generated.

[0242] B13. The computing device according to B12, wherein the executable instructions cause the processor to perform the following operations:

[0243] Generate a text position marking instruction, input the text position marking instruction into the generative recognition model, and obtain the position information of the dialogue text output by the generative recognition model;

[0244] The electronic book audio corresponding to the electronic book text is generated according to the dialogue audio corresponding to the character in the electronic book text and the position information of the dialogue text.

[0245] B14. The computing device according to B12 or B13, wherein the executable instructions cause the processor to perform the following operations:

[0246] Generate a character portrait generation instruction, input the character portrait generation instruction into the generative recognition model, and obtain character portrait information of any character output by the generative recognition model;

[0247] According to the character portrait information of any character, determine the timbre that matches the character.

[0248] B15. The computing device according to B14, wherein the executable instructions cause the processor to perform the following operations:

[0249] Get the timbre label of the candidate timbre;

[0250] For any character, the similarity between the timbre label of each candidate timbre and the character portrait information of the character is calculated, and the timbre matching the character is selected from the candidate timbre according to the similarity.

[0251] B16. The computing device according to B12 or B13, wherein the executable instructions cause the processor to perform the following operations:

[0252] In response to a timbre configuration operation for any character, a timbre matching the character is determined.

[0253] B17. A computing device according to any one of B12-B16, wherein the executable instructions cause the processor to perform the following operations:

[0254] Generate a scene recognition instruction, input the scene recognition instruction into the generative recognition model, and obtain scene information corresponding to the dialogue text output by the generative recognition model;

[0255] Generate scene audio corresponding to the dialogue text according to the scene information;

[0256] The electronic book audio corresponding to the electronic book text is generated according to the dialogue audio corresponding to the character in the electronic book text and the scene audio corresponding to the dialogue text.

[0257] B18. The computing device according to B11, wherein the executable instructions cause the processor to perform the following operations:

[0258] A dialogue novel is generated according to at least one dialogue text in the electronic book text and a character corresponding to any dialogue text.

[0259] B19. A computing device according to any one of B11-B18, wherein the conversation text recognition instruction comprises:

[0260] Instructions for recognizing conversation text based on preset symbols;

[0261] and / or, instructions indicating the output range of the dialogue text;

[0262] and / or, instructions indicating the output format of the dialogue text;

[0263] And / or, the dialogue text identifies instructions for processing rules under abnormal conditions.

[0264] B20. A computing device according to any one of B11-B19, wherein the role identification instruction comprises:

[0265] Instructions for identifying roles based on preset rules;

[0266] and / or, instructions indicating the scope of character identification;

[0267] and / or, instructions indicating the output format of the character;

[0268] And / or, instructions for processing rules under role recognition abnormal conditions.

[0269] The present invention also discloses: C21. A computer storage medium, wherein at least one executable instruction is stored in the storage medium, and the executable instruction enables a processor to execute an operation corresponding to the electronic book data processing method described in any one of A1-A10.

Claims

1. An electronic book data processing method, comprising: Obtaining an electronic book text, and inputting the electronic book text into a generative recognition model; Generate a dialogue text recognition instruction and a role recognition instruction, and input the dialogue text recognition instruction and the role recognition instruction into the generative recognition model; At least one dialogue text in the electronic book text output by the generative recognition model and a role corresponding to any dialogue text are obtained.

2. The method according to claim 1, wherein: The method further comprises: Determine a timbre that matches any character, and convert the dialogue text corresponding to the character into dialogue audio presented in the matching timbre; According to the dialogue audio corresponding to the character in the e-book text, an e-book audio corresponding to the e-book text is generated.

3. The method according to claim 2, wherein: The method further comprises: generating a text position marking instruction, inputting the text position marking instruction into the generative recognition model, and obtaining position information of the dialogue text output by the generative recognition model; Generating the e-book audio corresponding to the e-book text according to the dialogue audio corresponding to the character in the e-book text further includes: generating the e-book audio corresponding to the e-book text according to the dialogue audio corresponding to the character in the e-book text and the position information of the dialogue text.

4. The method according to claim 2 or 3, wherein: The method further comprises: generating a character portrait generation instruction, inputting the character portrait generation instruction into the generative recognition model, and acquiring character portrait information of any character output by the generative recognition model; The step of determining the timbre that matches any character further includes: determining the timbre that matches any character based on character portrait information of the character.

5. The method according to claim 4, wherein: Determining the timbre matching any character according to the character portrait information of the character further includes: Get the timbre label of the candidate timbre; For any character, the similarity between the timbre label of each candidate timbre and the character portrait information of the character is calculated, and the timbre matching the character is selected from the candidate timbre according to the similarity.

6. The method according to claim 2 or 3, wherein: Determining the timbre matching any character further includes: determining the timbre matching any character in response to a timbre configuration operation for the character.

7. The method according to any one of claims 2 to 6, wherein: The method further includes: generating a scene recognition instruction, inputting the scene recognition instruction into the generative recognition model, acquiring scene information corresponding to the dialogue text output by the generative recognition model, and generating scene audio corresponding to the dialogue text according to the scene information; Generating the e-book audio corresponding to the e-book text according to the dialogue audio corresponding to the character in the e-book text further includes: generating the e-book audio corresponding to the e-book text according to the dialogue audio corresponding to the character in the e-book text and the scene audio corresponding to the dialogue text.

8. The method according to claim 1, wherein: The method further comprises: A dialogue novel is generated according to at least one dialogue text in the electronic book text and a character corresponding to any dialogue text.

9. A computing device comprising: A processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus; The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform the following operations: Obtaining an electronic book text, and inputting the electronic book text into a generative recognition model; Generate a dialogue text recognition instruction and a role recognition instruction, and input the dialogue text recognition instruction and the role recognition instruction into the generative recognition model; At least one dialogue text in the electronic book text output by the generative recognition model and a role corresponding to any dialogue text are obtained.

10. A computer storage medium, wherein at least one executable instruction is stored in the storage medium, and the executable instruction enables a processor to execute an operation corresponding to the electronic book data processing method according to any one of claims 1 to 8.

Images