Audio annotation method, device, computer readable medium and electronic device

Abstract

The present application belongs to the field of artificial intelligence technology, and specifically relates to an audio annotation method, an annotation device, a computer-readable medium, and an electronic device; the present application establishes an audio progress bar, and automatically identifies machine voice segments and user speech segments in an audio file, then obtains annotation tags corresponding to the user speech segments, and automatically forms annotation points with annotation tags on the audio progress bar. By using the method of the present application to annotate audio, users can selectively listen to user speech segments for annotation, and the annotation efficiency and accuracy are high; at the same time, the annotated content of the present application can be intuitively displayed on the audio progress bar, which is convenient for users to quickly locate the content in the audio and convenient for manual rapid review of AI dialogues to achieve optimization of AI robots.

Description

Technical Field

[0001] The present application belongs to the field of artificial intelligence technology, and specifically relates to an audio annotation method, an audio annotation device, a computer-readable medium, and an electronic device. Background Art

[0002] With the advancement of technology, the field of artificial intelligence (AI) has developed rapidly. In the field of voice dialogue of artificial intelligence, the existing customer service platform has been able to realize voice dialogue between robots and users. During this dialogue process, the robot needs to recognize the user's words and understand the true meaning of the user's words in order to communicate effectively.

[0003] In order to improve the AI ​​robot's ability to understand user speech, it is often necessary to manually review the generated conversation audio and annotate it based on the review results, and then optimize the AI ​​robot based on the annotated content. However, the existing conversation audio annotation requires manual listening to the recording while pausing and annotating. The whole process is cumbersome and the annotation efficiency is low. Moreover, since pausing requires reaction time and machine reception requires response time, there are often errors between the marked points and the actual user speech points, resulting in inaccurate annotation. At the same time, the existing annotation method cannot directly display the annotated content on the progress bar, which is not conducive to the positioning of user speech in the audio file.

[0004] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present application, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0005] The purpose of the present application is to provide an audio annotation method, an audio annotation device, a computer-readable medium and an electronic device, which at least to some extent overcome the technical problems in the related art such as low audio annotation efficiency, poor annotation accuracy, and the inability to intuitively display the annotated content on a progress bar.

[0006] Other features and advantages of the present application will become apparent from the following detailed description, or may be learned in part by the practice of the present application.

[0007] According to one aspect of an embodiment of the present application, there is provided an audio tagging method, the method comprising:

[0008] Acquire an audio file, and establish an audio progress bar according to the audio duration of the audio file, wherein the audio file includes a machine voice segment and a user speech segment in a human-computer interaction dialogue;

[0009] Identify the machine speech segment and the user speech segment in the audio file, and record the start time of the user speech segment in the audio file;

[0010] Obtaining a label of the user speech segment; the label includes the label content of the user's speech segment;

[0011] Based on the start time, a marking point is generated on the audio progress bar, and the marking tag is embedded in the marking point.

[0012] According to one aspect of an embodiment of the present application, there is provided an audio tagging device, including:

[0013] A progress bar creation module is configured to obtain an audio file and create an audio progress bar according to the audio duration of the audio file, wherein the audio file includes a machine voice segment and a user speech segment in a human-computer interaction dialogue;

[0014] A recognition module, configured to recognize machine speech segments and user speech segments in the audio file, and record the start time of the user speech segment in the audio file;

[0015] A tag acquisition module is configured to acquire a tag of the user speech segment; the tag includes the tag content made by the user on the user speech segment;

[0016] The marking module is configured to generate a marking point on the audio progress bar based on the start time, and embed the marking tag into the marking point.

[0017] In some embodiments of the present application, based on the above technical solution, the progress bar establishment module includes:

[0018] a progress bar feature acquisition unit, configured to acquire an audio file, a first touch point, and a second touch point;

[0019] a connection unit, configured to use a connection line between the first touch point and the second touch point as an initial progress bar;

[0020] The mapping unit is configured to map the audio duration of the audio file to the initial progress bar to obtain the audio progress bar.

[0021] In some embodiments of the present application, based on the above technical solution, the identification module includes:

[0022] A parsing unit is configured to parse the audio file to obtain voiceprint features to be identified, wherein the voiceprint features include frequency features, timbre features, and pitch features;

[0023] A similarity calculation unit is configured to calculate the similarity between the voiceprint feature to be identified and a preset machine voiceprint feature;

[0024] The segmentation unit is configured to use the audio segment corresponding to the voiceprint feature to be identified that meets the similarity threshold as the machine voice segment; and use the audio segment other than the machine voice segment in the audio file as the user speech segment.

[0025] In some embodiments of the present application, based on the above technical solution, the similarity calculation unit includes:

[0026] A first calculation unit is configured to calculate the similarity between the frequency feature in the voiceprint feature to be identified and the frequency feature in the preset machine speech voiceprint feature to obtain a first similarity;

[0027] A second calculation unit is configured to calculate the similarity between the timbre feature in the voiceprint feature to be identified and the timbre feature in the preset machine speech voiceprint feature to obtain a second similarity;

[0028] The third calculation unit is configured to calculate the similarity between the tone feature in the voiceprint feature to be identified and the tone feature in the preset machine speech voiceprint feature to obtain a third similarity.

[0029] In some embodiments of the present application, based on the above technical solution, the segmentation unit includes:

[0030] The machine speech segmentation unit is configured to use the audio segment corresponding to the voiceprint feature to be identified as the machine speech segment when the first similarity of the voiceprint feature to be identified is greater than a first preset threshold, the second similarity is greater than a second preset threshold, and the third similarity is greater than a third preset threshold.

[0031] In some embodiments of the present application, based on the above technical solution, the identification module further includes a time recording unit, and the time recording unit includes:

[0032] A time acquisition unit, configured to acquire a feature change point in the audio file, wherein the feature change point includes a time point when the voiceprint feature to be identified changes;

[0033] The screening unit is configured to screen the feature change points, and record the time point when the voiceprint feature to be identified corresponding to the machine speech segment changes to the voiceprint feature to be identified corresponding to the user speech segment as the starting time.

[0034] In some embodiments of the present application, based on the above technical solution, the annotation module includes:

[0035] A pull-up box setting unit, configured to set a pull-up box on the marked point, wherein the pull-up box is triggered to be displayed based on a click on the marked point;

[0036] The input unit is configured to input the annotation label into the pull-up box.

[0037] According to one aspect of an embodiment of the present application, a computer-readable medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the audio tagging method in the above technical solution is implemented.

[0038] According to one aspect of an embodiment of the present application, an electronic device is provided, comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute the audio tagging method in the above technical solution by executing the executable instructions.

[0039] According to one aspect of the embodiments of the present application, a computer program product or a computer program is provided, the computer program product or the computer program including computer instructions, the computer instructions being stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the audio tagging method in the above technical solution.

[0040] In the technical solution provided in the embodiment of the present application, the present application establishes an audio progress bar and automatically identifies machine voice segments and user speech segments in the audio file, then obtains the annotation tags corresponding to the user speech segments, and automatically forms annotation points with annotation tags on the audio progress bar. By using the method of the present application for audio annotation, users can selectively listen to the user speech segments to annotate without listening to the entire audio file, and the annotation efficiency is high; moreover, after the user has completed the annotation, the annotation will automatically appear on the progress bar, and the annotation points are based on the start time of the user speech segment in the audio file. The annotation points on the progress bar are accurate, and do not require pause reaction time and machine reception response time. Therefore, the annotation accuracy is high; at the same time, the annotated content of the present application can be intuitively displayed on the audio progress bar, which is convenient for users to quickly locate the content in the audio and facilitate manual rapid review of AI dialogues to achieve optimization of AI robots.

[0041] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present application, and together with the specification are used to explain the principles of the present application. Obviously, the drawings described below are only some embodiments of the present application, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0043] Figure 1 The exemplary system architecture block diagram applying the technical solution of the present application is schematically shown.

[0044] Figure 2 The flowchart of the audio annotation method of the present application is schematically shown.

[0045] Figure 3 A schematic diagram of an audio progress bar in an embodiment of the present application is schematically shown.

[0046] Figure 4 A flowchart of a method for establishing an audio progress bar in an embodiment of the present application is schematically shown.

[0047] Figure 5 A schematic diagram of establishing an audio progress bar using touch points in an embodiment of the present application is schematically shown.

[0048] Figure 6 The flowchart of the method for identifying machine speech segments and user speech segments of the present application is schematically shown.

[0049] Figure 7 The flowchart of the method for segmenting an audio file in the present application is schematically shown.

[0050] Figure 8 The schematic diagram shows the effect of displaying the audio file after segmentation in the present application.

[0051] Fig. 9 A flowchart of the method for obtaining user speech fragments in the present application is schematically shown.

[0052] Fig.10 The display effect diagram after labeling of this application is schematically shown.

[0053] Fig.11 The structural block diagram of the audio annotation device of the present application is schematically shown.

[0054] Fig.12 The structure block diagram of a computer system suitable for implementing an electronic device of an embodiment of the present application is schematically shown. DETAILED DESCRIPTION

[0055] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this application will be more comprehensive and complete and fully convey the concept of the example embodiments to those skilled in the art.

[0056] In addition, described feature, structure or characteristic can be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present application. However, those skilled in the art will appreciate that the technical scheme of the present application can be put into practice without one or more of the specific details, or other methods, components, devices, steps, etc. can be adopted. In other cases, known methods, devices, realizations or operations are not shown or described in detail to avoid blurring the various aspects of the application.

[0057] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0058] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0059] User conversation refers to the questions users ask AI robots based on their needs and the content of their conversations with AI robots. By analyzing user conversations, we can evaluate the intelligence of AI robots. At the same time, we can also adjust the answer strategy of AI robots by annotating user conversations to optimize AI robots.

[0060] However, the existing dialogue audio annotation requires manual listening to the recording to find the user's speech point, and then pausing to annotate when reaching the user's speech point, and recording the pause time at the same time. The existing method has the following problems: first, it is necessary to manually listen to the entire recording audio, and the audio of the user's speech needs to be manually identified, so the annotation efficiency is low; second, because the reaction time of the person and the response time of the machine are certain influencing factors, when the person hears the content of the user's speech and wants to annotate it, there will be deviations in the time of clicking the pause button and the time of the machine's response, which will cause the time point recorded by the user to deviate from the actual time point, and the annotation is inaccurate; third, the user annotation method is to record the time and the annotation content, and it cannot be directly displayed on the progress bar. Moreover, the existing recording playback is to automatically generate the progress bar of the audio playback software itself through the audio playback software. The progress bar cannot be effectively edited, and the annotated content cannot be directly displayed on the progress bar, which is not conducive to the positioning of the user's speech in the audio file.

[0061] In order to solve the above technical problems, the present application discloses an audio annotation method, an audio annotation device, a computer-readable medium and an electronic device. The content of the present application will be further explained in various aspects below.

[0062] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0063] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0064] Figure 1 The exemplary system architecture block diagram applying the technical solution of the present application is schematically shown.

[0065] like Figure 1 As shown, the system architecture 100 may include a terminal device 110, a network 120, and a server 130. The terminal device 110 may include various electronic devices such as a smart phone, a tablet computer, a laptop computer, and a desktop computer. The server 130 may be an independent physical server, or a server cluster or a distributed system composed of multiple physical servers, or a cloud server providing cloud computing services. The network 120 may be a communication medium of various connection types that can provide a communication link between the terminal device 110 and the server 130, such as a wired communication link or a wireless communication link.

[0066] According to the implementation requirements, the system architecture in the embodiment of the present application can have any number of terminal devices, networks and servers. For example, the server 130 can be a server group composed of multiple server devices. In addition, the technical solution provided in the embodiment of the present application can be applied to the terminal device 110, can also be applied to the server 130, or can be implemented by the terminal device 110 and the server 130 together, and the present application does not make any special restrictions on this.

[0067] The above section introduces the contents of an exemplary system architecture that applies the technical solution of the present application. Next, the audio annotation method of the present application will be introduced.

[0068] According to one aspect of the embodiment of the present application, Figure 2 As shown, Figure 2The flowchart of the audio tagging method of the present application is schematically shown. The present application provides an audio tagging method, comprising steps S210 to S240.

[0069] In step S210: an audio file is obtained, and an audio progress bar is established according to the audio duration of the audio file, wherein the audio file includes machine voice segments and user speech segments in a human-computer interaction dialogue.

[0070] An audio file is a file in audio format, which is a pre-recorded or saved human-computer interaction dialogue file. The audio file of this application includes machine voice clips and user speech clips in the human-computer interaction dialogue. The machine voice clip is the voice sent by the AI ​​robot, and the user speech clip is the user's answer to the AI ​​or the voice content of the user's question based on the needs.

[0071] Audio files exist in different formats. Therefore, when acquiring an audio file, the present application will identify the format of the audio file. If the format cannot be played on the current terminal device 110 or server 130, the audio file can be converted to a format suitable for playing on the current terminal device 110 or server 130 by decoding and re-encoding. For example, if the acquired audio file is in MP3 format, and the current terminal device 110 can only play WAV format, then the decoding library can be customized according to the encoding type of the MP3 format, and then the customized decoding library can be compiled into webassembly code (a technical solution that can be written in a non-JavaScript programming language and can be run on a browser) by Emscripten (a compiler) for the current terminal device 110 or server 130 to call, and the decoding is performed by enabling multi-threading in webassembly on the current terminal device 110 or server 130, and the audio can be played without delay.

[0072] After obtaining the audio file, the present application can establish an audio progress bar according to the audio duration of the audio file. The audio progress bar of the present application is not generated when the audio player software in the current terminal device 110 or the server 130 plays the audio, but is independently established. The present application uses an independently established audio progress bar to mark the audio, and sets a style for the audio progress bar that can be selected by the user. The user can adjust the audio progress bar according to the actual situation, for example, the color, size and width of the audio progress bar can be set, so as to achieve a more personalized representation of the audio progress bar.

[0073] For establishing the audio progress bar, this application may adopt the following method.

[0074] In one embodiment of the present application, based on the above technical solution, a method for establishing an audio progress bar according to the audio duration of an audio file includes:

[0075] A progress bar of a specific length is drawn and displayed at any position of the display interface; the audio duration of the audio file is mapped to the progress bar to form an audio progress bar.

[0076] The audio progress bar of the present application is established according to the audio duration of the audio file, so the present application can draw a progress bar of a specific length and then display it at any position on the display interface. The display interface corresponds to the interface used to display information in the current terminal device 110 or server 130. For example, when the current terminal device 110 is a mobile phone, the display interface corresponds to the mobile phone screen, and when the current terminal device 110 is a computer, the display interface corresponds to the computer monitor.

[0077] like Figure 3 As shown, Figure 3 A schematic diagram of an audio progress bar in an embodiment of the present application is schematically shown.

[0078] in, Figure 3 There are two audio progress bars in the display interface 310, namely the first audio progress bar L1 and the second audio progress bar L2. The length of the two audio progress bars is fixed to 20 cm. The audio duration of the audio file is 20 minutes. If the 20-minute audio is mapped to the 20-cm progress bar, then the duration corresponding to each centimeter is 1 minute. Among them, the first marked point T1 of the first audio progress bar L1 is located at the starting position, and the playback has not started at this time, while the second marked point T2 of the second audio progress bar L2 is located at the midpoint of the progress bar, and the playback duration is exactly 10 minutes at this time.

[0079] The above method can be used to establish an audio progress bar. However, since the length of the above audio progress bar is fixed, it will not be displayed in a specific display interface. For example, when the current terminal device 110 is a mobile phone, the screen size is small. Therefore, if the length is still fixed, the progress bar may not be fully displayed on the screen. Therefore, in order to solve this problem, the present application proposes the following method for establishing an audio progress bar.

[0080] In one embodiment of the present application, before drawing a progress bar of a specific length and displaying it at any position on the display interface, the method further includes: acquiring the size of the display interface, and setting the length of the audio progress bar according to the size of the display interface.

[0081] This application can set the length of the audio progress bar according to the acquired display interface size. For example, mobile phones are generally small in size, so the audio progress bar is set shorter. In order to achieve the best display, this application sets the length of the audio progress bar to 4 / 5 of the display interface size. For example, if the display interface size of the mobile phone is 10cm*5cm, then the corresponding vertical screen progress bar is 4cm, and the corresponding size of the horizontal screen is 8cm.

[0082] When the above method is used to establish an audio progress bar, the audio progress bar has great limitations. When the display interface size is too small, the audio progress bar cannot be displayed well. Therefore, the present application also provides the following solution to establish an audio progress bar.

[0083] like Figure 4 As shown, Figure 4 The flowchart of the method for establishing an audio progress bar in an embodiment of the present application is schematically shown. In one embodiment of the present application, the method for obtaining an audio file and establishing an audio progress bar according to the audio duration of the audio file includes steps S410 to S430.

[0084] Step S410: Acquire an audio file, a first touch point, and a second touch point.

[0085] The first touch point and the second touch point are the start and end points of the audio progress bar determined by the user touching the display interface. The position of the audio progress bar can be determined by acquiring the first touch point and the second touch point.

[0086] Step S420: using the line connecting the first touch point and the second touch point as an initial progress bar.

[0087] After determining the start and end points of the audio progress bar, the line connecting the first touch point and the second touch point can be directly used as the initial progress bar, and then step S430 is performed.

[0088] Step S430: Map the audio duration of the audio file to the initial progress bar to obtain the audio progress bar.

[0089] Step S430 directly maps the audio duration to the corresponding initial progress bar, so that the length of the audio progress bar represents the corresponding duration. This part is similar to the above audio progress bar mapping process and will not be repeated here.

[0090] The following example further illustrates the contents of step S410 to step S430. Figure 5 As shown, Figure 5The schematic diagram of establishing an audio progress bar using touch points in an embodiment of the present application is schematically shown. When the screen size of the terminal device 110 is small, the present application can implement the custom setting of the audio progress bar using the above solution. The user clicks the first touch point 510 and the second touch point 520 on the diagonal of the screen respectively, and then the corresponding audio progress bar is a diagonal line. This helps to improve the utilization rate of the screen, can lengthen the length of the audio progress bar, and can be applied to various different terminal devices 110.

[0091] After the audio progress bar is established using the above method, step S220 can be continued.

[0092] In step S220: identify the machine speech segments and user speech segments in the audio file, and record the starting time of the user speech segment in the audio file.

[0093] The audio file includes machine voice segments and user speech segments, so the audio file needs to be identified, and there are three strategies for identifying audio files. The first is to identify machine voice segments and user speech segments separately. If this is done, it is necessary to pre-train a sufficient number of machine voice samples and user speech samples to realize the recognition of audio files. As for user speech, the voiceprints corresponding to each user are different, so there are many forms, and it is impossible to fully form a sample that can cover all users. Therefore, this method cannot well identify machine voice segments and user speech segments. The second way is to directly identify the user speech segments summarized in the audio file, and then use other segments in the audio file as machine voice segments. This method also requires enough user speech segment samples, so it is also difficult to accurately identify user speech segments in audio files. Therefore, this application uses the third method, that is, this application identifies the machine voice segments in the audio file, and then uses other sound segments except the machine voice segments as user speech segments. The specific method is as follows:

[0094] In one embodiment of the present application, Figure 6 As shown, Figure 6 The flowchart of the method for identifying machine speech segments and user speech segments of the present application is schematically shown. The method for identifying machine speech segments and user speech segments in an audio file of the present application includes steps S610 to S640.

[0095] In step S610: the audio file is parsed to obtain the voiceprint features to be identified, which include frequency features, timbre features and pitch features.

[0096] The machine voice segments and user speech segments in the audio files of the present application correspond to the AI ​​robot and the user, and the sound characteristics of the AI ​​robot are preset and fixed. Therefore, for audio files, there are segments with multiple different sound characteristics in the audio files. For a recording file, there are generally two sound characteristics. Therefore, by parsing the audio files, the sound characteristics of different segments in all audio files can be obtained.

[0097] The sound characteristics are generally measured by voiceprints. Voiceprint is a sound wave spectrum that carries speech information and is displayed by an electroacoustic instrument. The production of human language is a complex physiological and physical process between the human body's language center and the vocal organs. The vocal organs (e.g., tongue, teeth, larynx, lungs, nasal cavity) used by people when speaking vary greatly in size and shape, so the voiceprints of any two people are different. This application needs to parse the audio file and obtain the voiceprint features of all audio clips in the audio file as the voiceprint features to be identified. The voiceprint features to be identified include frequency features, timbre features and pitch features. A person's voice is generally composed of frequency, timbre and pitch. Therefore, the subject of the sound can be identified through these three parts. Correspondingly, the voiceprint features of the AI ​​robot of this application are pre-set, so through step S610, two voiceprint features corresponding to different clips can be parsed from the audio file as the voiceprint features to be identified, and then continue to step S620.

[0098] In step S620: the similarity between the voiceprint feature to be identified and the preset machine voiceprint feature is calculated.

[0099] The process of similarity calculation in this application is the process of voiceprint recognition, which can be called speaker recognition (Speaker Recognition), including: speaker identification (Speaker Identification) and speaker verification (Speaker Verification). The former is used to determine which of several people said a certain speech, which is a "multiple-choice" problem; while the latter is used to confirm whether a certain speech is said by a specified person, which is a "one-to-one discrimination" problem. This application mainly uses the latter speaker confirmation, which corresponds to machine voice confirmation in this application. Before confirming the machine voice, the voiceprint of the machine voice needs to be modeled first. This is the so-called "training" or "learning" process.

[0100] However, this application requires training of preset machine voiceprint features by inputting a sufficient number of machine voices into the convolutional neural network for training. During the training process, the voiceprint features in the machine voice are mainly trained, and the frequency features, timbre features and pitch features are trained separately. By training with a sufficient number of samples, the frequency features, timbre features and pitch features corresponding to the machine voice can be obtained. However, this application can also directly initially define the frequency features, timbre features and pitch features of the machine voice as the object of comparison without the need for training. The preset machine voiceprint features can be obtained through the above method.

[0101] After obtaining the preset machine voiceprint features, it is necessary to calculate the similarity between the voiceprint features to be identified and the preset machine voiceprint features. In one embodiment of the present application, the method for calculating the similarity between the voiceprint features to be identified and the preset machine voiceprint features includes: calculating the similarity between the frequency features in the voiceprint features to be identified and the frequency features in the preset machine voice voiceprint features to obtain a first similarity; calculating the similarity between the timbre features in the voiceprint features to be identified and the timbre features in the preset machine voice voiceprint features to obtain a second similarity; calculating the similarity between the pitch features in the voiceprint features to be identified and the pitch features in the preset machine voice voiceprint features to obtain a third similarity.

[0102] Through the above method, the similarity between the voiceprint feature to be identified and the preset machine voiceprint feature in three dimensions is obtained. In this application, the similarity of the three dimensions can be averaged as the overall similarity between the voiceprint feature to be identified and the preset machine voiceprint feature. Therefore, through the above steps, the similarity of the voiceprint feature to be identified and the preset machine voiceprint feature in three dimensions or the overall similarity can be obtained, and then step S630 can be performed.

[0103] In step S630: the audio segment corresponding to the voiceprint feature to be identified that meets the similarity threshold is used as a machine voice segment;

[0104] The present application has two methods for determining whether the similarity threshold is satisfied, which correspond to the two similarities obtained in step S620 respectively.

[0105] The first method: When the similarity between the voiceprint feature to be identified and the preset machine voiceprint feature in three dimensions is obtained, in one embodiment of the present application, the method of using the audio segment corresponding to the voiceprint feature to be identified that meets the similarity threshold as the machine voice segment includes: when the first similarity of the voiceprint feature to be identified is greater than the first preset threshold, the second similarity is greater than the second preset threshold, and the third similarity is greater than the third preset threshold, the audio segment corresponding to the voiceprint feature to be identified is used as the machine voice segment.

[0106] The first method of the present application calculates the similarity of the three dimensions separately. Only when the similarity of the three dimensions meets the similarity threshold, the audio segment corresponding to the voiceprint feature to be identified is used as a machine voice segment. This method refers to the similarity of the three dimensions and is more accurate in identifying machine voice segments. However, due to the harsh conditions, the number of voiceprints corresponding to the obtained machine voice segments may be small. Therefore, the present application can also use the following second method.

[0107] The second method: when the overall similarity is obtained, the overall similarity is directly compared with the similarity threshold. When the overall similarity is greater than the first similarity, the audio segment corresponding to the corresponding voiceprint feature to be identified is used as a machine voice segment. This method uses the average value of the three dimensions as the basis for comparison, and can refer to the comprehensive three dimensions. Therefore, for some dimensions that do not meet some similarity thresholds, but the overall similarity meets the similarity threshold, they can all be used as machine voice segments. Therefore, this solution can obtain more machine voice segments.

[0108] The above method can be used to obtain the voiceprint features to be identified that meet the similarity threshold. The audio segment corresponding to the voiceprint features to be identified is the machine voice segment. Therefore, all machine voice segments in the audio file are identified through step S630, and then continue to step S640.

[0109] In step S640: the audio segments in the audio file other than the machine voice segments are taken as user speech segments.

[0110] For audio files, the audio segments other than the machine voice segments are regarded as user speech segments.

[0111] The following example further illustrates the contents of step S610 to step S640: An audio file is obtained, and the specific voice content of the audio file is as follows, where AI represents an AI robot and PR represents a user.

[0112] AI: Hello, are you ***?

[0113] PR: Yes.

[0114] AI: This is ** Bank. We are calling because we saw that you are interested in our credit card. Thank you for your support. We would like to do a simple survey with you regarding this credit card?

[0115] PR: Okay.

[0116] AI: Do you have a car? Do you need to buy one?

[0117] PR: No, there is no demand for it at the moment.

[0118] AI: Okay, thank you for your patience, goodbye.

[0119] When the present application parses the audio file, seven audio segments will be obtained. The number of audio segments is generally segmented according to the change of the voiceprint. For example, the user said three sentences in a row, but his voiceprint did not change. Therefore, even if more content was said during the period, it still corresponds to an audio segment. Each audio segment corresponds to a voiceprint feature. In step S610 of the present application, seven audio features are obtained as audio features to be identified. At this time, these seven audio features can be input respectively with the preset machine voiceprint features for similarity calculation, which corresponds to the content of step S620 of the present application. The similarity between the seven audio features and the machine voiceprint features in three dimensions is calculated. For example, in the first paragraph, "AI: Hello, are you ***?" The first similarity corresponding to the audio feature obtained by calculation is 98%, the second similarity corresponding to the timbre feature is 97%, and the third similarity corresponding to the pitch feature is 99%.

[0120] At this point, you can continue with step S630, and compare the similarities of the above seven audio clips with the similarity thresholds respectively. If the similarities of the voiceprint features to be identified in the corresponding audio clips in the three dimensions of audio, timbre and pitch are greater than the specific similarity threshold, then the clip corresponding to the corresponding voiceprint features to be identified will be used as a machine audio clip. For example, through identification, the four clips spoken by the above AI robot are used as machine voice clips, and the other three clips are user speech clips.

[0121] Through the above steps, the machine voice segments and user speech segments in the audio file can be identified. In order to make it more convenient for users to mark each segment, this application can also split the audio file. The specific method is as follows.

[0122] In one embodiment of the present application, Figure 7 As shown, Figure 7 The flowchart of the method for segmenting an audio file in the present application is schematically shown. The method includes steps S710 to S720.

[0123] Step S710: Identify feature change points in the audio file, where the feature change points include time points when voiceprint features change.

[0124] For audio files, if there is a pause or no sound in the middle, there is no voiceprint. Therefore, the voiceprint will not change when the user or AI robot is speaking. The feature change point is the time point when the voiceprint changes.

[0125] The following example illustrates the above content. If the time corresponding to an audio file is 30 seconds (s), the corresponding voice content and time are as follows:

[0126] (0s-5s)AI: Hello, are you ***?

[0127] (6s-7s)PR: Yes.

[0128] (9s-18s) AI: This is ** Bank. We are calling because we saw that you are interested in our credit card. Thank you for your support. We would like to do a simple survey with you on this credit card?

[0129] (20s-21s)PR: OK.

[0130] (23s-27s) AI: Do you have a car? Do you need to buy one?

[0131] (28s-30s)PR: No, there is no demand at the moment.

[0132] The characteristic change points in the above audio files are the time points corresponding to 6s, 9s, 20s, 23s and 28s respectively. After obtaining these time points, step S720 can be performed.

[0133] Step S720: segmenting the audio file using the feature change points as segmentation points.

[0134] By segmenting these characteristic change points, six audio segments can be obtained. Correspondingly, the six audio segments can be displayed on the display interface 310, such as Figure 8 As shown, Figure 8 The present application schematically shows the effect diagram of displaying the audio file after segmentation. Through the above segmentation, six audio clips can be displayed on the display interface P in the form of playback controls 810. Among them, different playback controls 810 have different audio lengths, and the user can randomly click on any audio segment to play, and then mark the audio.

[0135] In one embodiment of the present application, a marking label option is provided on the playback control 810 of the present application for the user to select, wherein the marking label option can be customized according to the user to place commonly used marking labels here, for example, commonly used marking labels include: accurate characters with text, accurate characters without text, accurate answer, wrong answer, unrecognizable dialect, regression test, etc. Therefore, by using this method to pre-set labels, the user can save the time of label classification and input, and can directly obtain them by clicking.

[0136] The above method can obtain machine voice segments and user speech segments, and segment the two. However, this application also needs to obtain the start time of the user speech segment in the audio file. In one embodiment of the present application, Fig. 9As shown, Fig. 9 The method for recording the start time of the user speech segment in the audio file includes steps S910 to S920:

[0137] Step S910: Acquire feature change points in the audio file, where the feature change points include time points when the voiceprint features to be identified change.

[0138] This step is the same as step S710 and will not be repeated here.

[0139] Step S920: Filter the feature change points, and record the time point when the voiceprint feature to be identified corresponding to the machine speech segment changes to the voiceprint feature to be identified corresponding to the user speech segment as the starting time.

[0140] Step S920 of this application needs to filter out the time point when the voiceprint feature to be identified corresponding to the machine voice segment changes to the voiceprint feature to be identified corresponding to the user speech segment as the starting time point. There are two cases for feature change points, one is from user speech to machine voice, and the other is from machine speech to user speech. Therefore, step S920 of this application only filters out the time point from machine speech to user speech as the starting time. Continuing with the audio file of step S710 as an example, the starting time includes the 6th second, the 20th second, and the 28th second of the entire audio file. These three time points are the corresponding starting times.

[0141] After identifying the machine voice segment and the user speech segment and obtaining the corresponding start time of the user speech segment, you can proceed to step S230.

[0142] In step S230: obtain the annotation label of the user speech segment; the annotation label includes the annotation content made by the user on the user speech segment.

[0143] After the machine voice segment and the user speech segment are divided through step S720, the user can directly play the user speech segment, and then mark it accordingly, and can choose different marking tags. After the user marking is completed, the corresponding method of the present application can obtain the marking tag, and then proceed to step S240.

[0144] In step S240: based on the start time, a marking point is generated on the audio progress bar, and a marking tag is embedded in the marking point.

[0145] After obtaining the starting time of the user's speech and the corresponding annotation label, you can generate an annotation point on the audio progress bar. You can use the relationship between the duration and the progress bar to locate the position of the annotation point. This application can accurately locate the time to ms. For example, a recording is 20 minutes, the audio progress bar is 20 cm, and the corresponding starting time of a user's speech is 10 minutes, 10 seconds, and 10 milliseconds. Then the annotation point of the user's speech on the progress bar is located at 36610 / 72000 of the progress bar, that is, the corresponding position on the progress bar is 20cm*36610 / 72000.

[0146] After forming the annotation point, the annotation label can be embedded in the annotation point. Specifically, the method of embedding the annotation label in the annotation point in the present application includes: setting a pull-up box on the annotation point, the pull-up box is triggered to display based on the click of the annotation point, and the annotation label is input into the pull-up box. The pull-up box can also be replaced by setting a floating box, and the present application does not limit this.

[0147] The following will be Fig.10 To further illustrate the content of step S240 of this application. Fig.10 As shown, Fig.10 The display effect diagram after labeling of this application is schematically shown.

[0148] Continuing with the voice content corresponding to the audio file in step S710 as an example, the starting time includes the 6th second, 20th second and 28th second of the entire audio file, corresponding to the formation of three annotation points P1, P2, and P3. Among them, P1 is located at 20 / 30 of the entire progress bar, P1 is located at 28 / 30 of the entire progress bar, and P1 is located at 6 / 30 of the entire progress bar. Then enter the received annotation tags into the pull-up box. For example, the annotations for the audio file in step S710 are, P1 is affirmative, P2 is affirmative, and P3 is negative, respectively representing the attitude of the user's speech. Therefore, the annotation note can be embedded in the pull-up box. The user can intuitively see the annotation content of the entire audio file, which is convenient for the user to locate the audio file and adjust the speech of the AI ​​robot later.

[0149] In the technical solution provided in the embodiment of the present application, the present application establishes an audio progress bar and automatically identifies machine voice segments and user speech segments in the audio file, then obtains the annotation tags corresponding to the user speech segments, and automatically forms annotation points with annotation tags on the audio progress bar. By using the method of the present application for audio annotation, users can selectively listen to the user speech segments to annotate without listening to the entire audio file, and the annotation efficiency is high; moreover, after the user has completed the annotation, the annotation will automatically appear on the progress bar, and the annotation points are based on the start time of the user speech segment in the audio file. The annotation points on the progress bar are accurate, and do not require pause reaction time and machine reception response time. Therefore, the annotation accuracy is high; at the same time, the annotated content of the present application can be intuitively displayed on the audio progress bar, which is convenient for users to quickly locate the content in the audio and facilitate manual rapid review of AI dialogues to achieve optimization of AI robots.

[0150] It should be noted that although the steps of the method in the present application are described in a specific order in the drawings, this does not require or imply that the steps must be performed in this specific order, or that all the steps shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps, etc.

[0151] The above part introduces the content of the data audio annotation method of this application, and the following part will continue to introduce the content of the audio annotation device of this application.

[0152] According to one aspect of an embodiment of the present application, an audio tagging device 1100 is provided, including:

[0153] The progress bar establishment module 1110 is configured to obtain an audio file and establish an audio progress bar according to the audio duration of the audio file, wherein the audio file includes a machine voice segment and a user speech segment in a human-computer interaction dialogue;

[0154] The recognition module 1120 is configured to recognize machine speech segments and user speech segments in the audio file, and record the start time of the user speech segment in the audio file;

[0155] The tag acquisition module 1130 is configured to acquire the annotated tags of the user speech segment; the annotated tags include the annotated content of the user's speech segment;

[0156] The marking module 1140 is configured to generate a marking point on the audio progress bar based on the start time, and embed the marking tag into the marking point.

[0157] In some embodiments of the present application, based on the above technical solution, the progress bar establishment module 1110 includes:

[0158] a progress bar feature acquisition unit, configured to acquire an audio file, a first touch point, and a second touch point;

[0159] A connection unit, configured to use a connection line between the first touch point and the second touch point as an initial progress bar;

[0160] The mapping unit is configured to map the audio duration of the audio file to the initial progress bar to obtain the audio progress bar.

[0161] In some embodiments of the present application, based on the above technical solution, the identification module 1120 includes:

[0162] The parsing unit is configured to parse the audio file to obtain the voiceprint features to be identified, where the voiceprint features include frequency features, timbre features, and pitch features;

[0163] A similarity calculation unit is configured to calculate the similarity between the voiceprint feature to be identified and the preset machine voiceprint feature;

[0164] The segmentation unit is configured to use the audio segments corresponding to the voiceprint features to be identified that meet the similarity threshold as machine voice segments; and use the audio segments in the audio file other than the machine voice segments as user speech segments.

[0165] In some embodiments of the present application, based on the above technical solution, the similarity calculation unit includes:

[0166] A first calculation unit is configured to calculate the similarity between the frequency feature in the voiceprint feature to be identified and the frequency feature in the preset machine speech voiceprint feature to obtain a first similarity;

[0167] A second calculation unit is configured to calculate the similarity between the timbre feature in the voiceprint feature to be identified and the timbre feature in the preset machine speech voiceprint feature to obtain a second similarity;

[0168] The third calculation unit is configured to calculate the similarity between the tone feature in the voiceprint feature to be identified and the tone feature in the preset machine speech voiceprint feature to obtain a third similarity.

[0169] In some embodiments of the present application, based on the above technical solution, the segmentation unit includes:

[0170] The machine speech segmentation unit is configured to use the audio segment corresponding to the voiceprint feature to be identified as a machine speech segment when the first similarity of the voiceprint feature to be identified is greater than a first preset threshold, the second similarity is greater than a second preset threshold, and the third similarity is greater than a third preset threshold.

[0171] In some embodiments of the present application, based on the above technical solution, the identification module 1120 further includes a time recording unit, and the time recording unit includes:

[0172] A time acquisition unit is configured to acquire a feature change point in the audio file, where the feature change point includes a time point when a feature of the voiceprint to be identified changes;

[0173] The screening unit is configured to screen feature change points, and record the time point when the voiceprint feature to be identified corresponding to the machine speech segment changes to the voiceprint feature to be identified corresponding to the user speech segment as the starting time.

[0174] In some embodiments of the present application, based on the above technical solution, the marking module 1140 includes:

[0175] A pull-up box setting unit is configured to set a pull-up box on the marked point, and the pull-up box is triggered to display based on a click on the marked point;

[0176] The input unit is configured to input the annotation label into the drop-down box.

[0177] The specific details of the audio annotation device provided in each embodiment of the present application have been described in detail in the corresponding method embodiment, and will not be repeated here.

[0178] The above part introduces the contents of the audio annotation device of the present application, and the following will continue to introduce the contents of other aspects of the present application.

[0179] According to one aspect of an embodiment of the present application, a computer-readable medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the audio tagging method in the above technical solution is implemented.

[0180] According to one aspect of an embodiment of the present application, an electronic device is provided, comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute the audio tagging method in the above technical solution by executing the executable instructions.

[0181] According to one aspect of the embodiments of the present application, a computer program product or a computer program is provided, the computer program product or the computer program including computer instructions, the computer instructions being stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the audio tagging method in the above technical solution.

[0182] Fig.12 The structure block diagram of a computer system for implementing an electronic device according to an embodiment of the present application is schematically shown.

[0183] It should be noted that Fig.12 The computer system 1200 of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0184] like Fig.12 As shown, the computer system 1200 includes a central processing unit 1201 (CPU), which can perform various appropriate actions and processes according to the program stored in the read-only memory 1202 (ROM) or the program loaded from the storage part 1208 to the random access memory 1203 (RAM). Various programs and data required for system operation are also stored in the random access memory 1203. The central processing unit 1201, the read-only memory 1202 and the random access memory 1203 are connected to each other through a bus 1204. The input / output interface 1205 (Input / Output interface, i.e., I / O interface) is also connected to the bus 1204.

[0185] The following components are connected to the input / output interface 1205: an input section 1206 including a keyboard, a mouse, etc.; an output section 1207 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 1208 including a hard disk, etc.; and a communication section 1209 including a network interface card such as a LAN card, a modem, etc. The communication section 1209 performs communication processing via a network such as the Internet. A drive 1210 is also connected to the input / output interface 1205 as needed. A removable medium 1211, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 1210 as needed so that a computer program read therefrom is installed into the storage section 1208 as needed.

[0186] In particular, according to an embodiment of the present application, the process described in each method flow chart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer readable medium, and the computer program contains a program code for executing the method shown in the flow chart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 1209, and / or installed from the removable medium 1211. When the computer program is executed by the central processor 1201, various functions defined in the system of the present application are executed.

[0187] It should be noted that the computer-readable medium shown in the embodiment of the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by an instruction execution system, device or device or used in combination with it. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, wherein a computer-readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer readable signal media may also be any computer readable medium other than computer readable storage media, which may send, propagate, or transmit programs for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0188] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the above-mentioned module, program segment or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0189] It should be noted that, although several modules or units of the equipment for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present application, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into being embodied by multiple modules or units.

[0190] Through the description of the above implementation methods, it is easy for those skilled in the art to understand that the example implementation methods described here can be implemented by software or by combining software with necessary hardware. Therefore, the technical solution according to the implementation methods of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a touch terminal, or a network device, etc.) to execute the method according to the implementation methods of the present application.

[0191] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary technical means in the art that are not disclosed in the present application.

[0192] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. An audio tagging method, characterized in that: The method comprises: Acquire an audio file, and establish an audio progress bar according to the audio duration of the audio file, wherein the audio file includes a machine voice segment and a user speech segment in a human-computer interaction dialogue; Identify the machine speech segment and the user speech segment in the audio file, and record the start time of the user speech segment in the audio file; Obtaining a label of the user speech segment; the label includes the label content of the user's speech segment; Based on the start time, generating a mark point on the audio progress bar, setting a pull-up box on the mark point, and triggering display of the pull-up box based on clicking the mark point; The annotation label is input into the pull-up box.

2. The audio tagging method according to claim 1, characterized in that: Acquire an audio file, and establish an audio progress bar according to the audio duration of the audio file, including: Acquire the audio file, the first touch point, and the second touch point; Using the line connecting the first touch point and the second touch point as an initial progress bar; The audio duration of the audio file is mapped to the initial progress bar to obtain the audio progress bar.

3. The audio tagging method according to claim 1, characterized in that: Identifying machine speech segments and user speech segments in the audio file includes: Parsing the audio file to obtain voiceprint features to be identified, wherein the voiceprint features to be identified include frequency features, timbre features, and pitch features; Calculating the similarity between the voiceprint feature to be identified and the preset machine voiceprint feature; The audio segment corresponding to the voiceprint feature to be identified that meets the similarity threshold is used as the machine voice segment; The audio segments in the audio file other than the machine voice segments are used as the user speech segments.

4. The audio tagging method according to claim 3, characterized in that: Calculating the similarity between the voiceprint feature to be identified and the preset machine voiceprint feature includes: Calculating the similarity between the frequency feature in the voiceprint feature to be identified and the frequency feature in the preset machine speech voiceprint feature to obtain a first similarity; Calculating the similarity between the timbre feature in the voiceprint feature to be identified and the timbre feature in the preset machine speech voiceprint feature to obtain a second similarity; The similarity between the tone feature in the voiceprint feature to be identified and the tone feature in the preset machine speech voiceprint feature is calculated to obtain a third similarity.

5. The audio tagging method according to claim 4, characterized in that: The audio segment corresponding to the voiceprint feature to be identified that meets the similarity threshold is used as the machine voice segment, including: When the first similarity of the voiceprint feature to be identified is greater than the first preset threshold, the second similarity is greater than the second preset threshold, and the third similarity is greater than the third preset threshold, the audio segment corresponding to the voiceprint feature to be identified is used as the machine voice segment.

6. The audio tagging method according to claim 3, characterized in that: Recording the start time of the user speech segment in the audio file includes: Acquire a feature change point in the audio file, wherein the feature change point includes a time point when the voiceprint feature to be identified changes; The feature change point is filtered, and the time point when the voiceprint feature to be identified corresponding to the machine speech segment is changed to the voiceprint feature to be identified corresponding to the user speech segment is recorded as the starting time.

7. An audio tagging device, characterized in that: include: A progress bar creation module is configured to obtain an audio file and create an audio progress bar according to the audio duration of the audio file, wherein the audio file includes a machine voice segment and a user speech segment in a human-computer interaction dialogue; A recognition module, configured to recognize machine speech segments and user speech segments in the audio file, and record the start time of the user speech segment in the audio file; A tag acquisition module is configured to acquire the annotated tags of the user speech fragment; The annotation tag includes the annotation content made by the user on the user speech segment; The marking module is configured to generate a marking point on the audio progress bar based on the start time, set a pull-up box on the marking point, and trigger the display of the pull-up box based on a click on the marking point; and input the marking tag into the pull-up box.

8. A computer readable medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the audio tagging method according to any one of claims 1 to 6 is implemented.

9. An electronic device, characterized in that: include: processor; as well as A memory, configured to store executable instructions of the processor; The processor is configured to perform the audio tagging method according to any one of claims 1 to 6 by executing the executable instructions.

Images