Method and apparatus for identifying interactive audio, device, and storage medium

Abstract

Embodiments of the present application disclose a method and device for identifying interactive audio, equipment and a storage medium. The method comprises: obtaining interactive audio; determining an identification result of the interactive audio based on an audio identification service in an audio identification service system, the audio identification service system comprising at least an invalid identification service and a text processing service, the processing efficiency of the text processing service being higher than that of the invalid identification service, and the identification result comprising valid interaction and invalid interaction; and determining the identification result of the interactive audio as valid interaction when the identification result of the text processing service is valid interaction. The terminal equipment can output the result more quickly and efficiently.

Description

Technical Field

[0001] This application relates to audio processing technology, including but not limited to a method, apparatus, device, and storage medium for recognizing interactive audio. Background Technology

[0002] In the process of artificial intelligence working, it is usually determined whether a response is needed based on the audio of the user's speech.

[0003] In related technologies, the identification result is usually determined directly based on the invalid rejection service, and then the identification result is used to determine whether an answer is required.

[0004] However, due to the possible presence of noise or other external influencing factors, the invalid rejection service cannot quickly obtain identification results, thus preventing the artificial intelligence from determining in a timely manner whether a response is needed, resulting in slow output speed and low efficiency. Summary of the Invention

[0005] In view of this, the interactive audio recognition method, apparatus, device, and storage medium provided in the embodiments of this application enable terminal devices to output results more quickly and efficiently. The interactive audio recognition method, apparatus, device, and storage medium provided in the embodiments of this application are implemented as follows:

[0006] This application provides an embodiment of an interactive audio recognition method, which is applied to a terminal device. The method includes:

[0007] Get interactive audio;

[0008] Based on the audio recognition service in the audio recognition service system, the recognition result of the interactive audio is determined. The audio recognition service system includes at least an invalid rejection service and a text processing service. The processing efficiency of the text processing service is higher than that of the invalid rejection service. The recognition result includes valid interactions and invalid interactions.

[0009] If the text processing service identifies the interaction as valid, the audio recognition result will also be considered valid.

[0010] The interactive audio recognition device provided in this application embodiment is applied to a terminal device and includes: an acquisition module and a recognition module;

[0011] The acquisition module is used to acquire interactive audio.

[0012] The recognition module is used to determine the recognition result of interactive audio based on the audio recognition service in the audio recognition service system. The audio recognition service system includes at least an invalid rejection service and a text processing service. The processing efficiency of the text processing service is higher than that of the invalid rejection service. The recognition result includes valid interaction and invalid interaction.

[0013] If the text processing service identifies the interaction as valid, the audio recognition result will also be considered valid.

[0014] The computer device provided in this application includes a memory and a processor. The memory stores a computer program that can run on the processor, and the processor executes the program to implement the method of this application.

[0015] The computer-readable storage medium provided in this application embodiment stores a computer program thereon, which, when executed by a processor, implements the method provided in this application embodiment.

[0016] The interactive audio recognition method, apparatus, computer device, and computer-readable storage medium provided in this application embodiment can determine the recognition result of interactive audio based on the audio recognition service in the audio recognition service system. Since the processing efficiency of the text processing service is higher than that of the invalid rejection service, the recognition result of the interactive audio can be determined as valid interaction if the recognition result of the text processing service is valid interaction. Compared with using only the invalid rejection service to determine the recognition result of interactive audio, the recognition result of interactive audio can be obtained more quickly and efficiently. Furthermore, based on the text processing service, the accuracy of the recognition result can be guaranteed to be high and the consumption of computing power can be reduced, which means that the terminal device can output the recognition result of the interactive audio more promptly. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with this application and, together with the specification, serve to explain the technical solutions of this application.

[0018] Figure 1 A schematic diagram illustrating an application scenario of the interactive audio recognition method provided in this application embodiment;

[0019] Figure 2 A flowchart illustrating the interactive audio recognition method provided in this application embodiment;

[0020] Figure 3 A schematic diagram illustrating the process of obtaining the recognition result of interactive audio provided in an embodiment of this application;

[0021] Figure 4 Another schematic diagram illustrating the process of obtaining the recognition result of interactive audio provided in an embodiment of this application;

[0022] Figure 5 Another schematic diagram illustrating the process of obtaining the recognition result of interactive audio provided in an embodiment of this application;

[0023] Figure 6 Another schematic diagram illustrating the process of obtaining the recognition result of interactive audio provided in an embodiment of this application;

[0024] Figure 7 Another schematic diagram illustrating the process of obtaining the recognition result of interactive audio provided in an embodiment of this application;

[0025] Figure 8 A logical diagram illustrating the results of obtaining interactive audio recognition, provided in an embodiment of this application;

[0026] Figure 9 Another schematic diagram illustrating the process of obtaining the recognition result of interactive audio provided in an embodiment of this application;

[0027] Figure 10 Another logical diagram illustrating the recognition result of interactive audio provided in an embodiment of this application;

[0028] Figure 11 A flowchart illustrating the process of determining the identification result based on the invalid rejection service provided in this application embodiment;

[0029] Figure 12 Another flowchart illustrating the process of determining the identification result based on the invalid rejection service provided in this application embodiment;

[0030] Figure 13 A schematic diagram showing the output of the terminal device based on the recognition result provided in the embodiments of this application;

[0031] Figure 14 Another schematic diagram showing the output of the terminal device provided in this application embodiment based on the recognition result;

[0032] Figure 15 A schematic diagram of the structure of the interactive audio recognition device provided in the embodiments of this application;

[0033] Figure 16 A schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the specific technical solutions of this application will be further described in detail below with reference to the accompanying drawings of the embodiments of this application. The following embodiments are used to illustrate this application, but are not intended to limit the scope of this application.

[0035] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0036] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0037] It should be noted that the terms "first, second, third" used in the embodiments of this application are used to distinguish similar or different objects and do not represent a specific order of objects. It can be understood that "first, second, third" can be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0038] During the operation of a voice assistant, it typically determines whether a response is needed based on the audio of the user's spoken words. For example, after a user speaks a sentence or a paragraph to the voice assistant on their phone, the voice assistant can determine whether a response is required based on the received audio.

[0039] For example, when the interactive audio is a question or request, such as "Please help me open software A" or "What does B mean?", the voice assistant can determine that these interactive audios are valid interactions and can provide corresponding answers. When the interactive audio is certain types of declarative or exclamatory sentences, such as "It's raining today" or "This meal was delicious", the voice assistant can determine that these interactive audios are invalid interactions and does not require a response.

[0040] In actual implementation, if it is necessary to use a voice assistant to determine whether the interactive audio is a valid or invalid interaction, the technical solution used in related technologies is usually to determine the recognition result based on the invalid rejection service, and then determine whether an answer is required based on the recognition result.

[0041] However, in practical applications, due to possible noise or other external influencing factors, the invalid rejection service cannot quickly obtain recognition results, which makes it impossible for the voice assistant to determine whether a response is needed in a timely manner. This results in slow output speed and low efficiency. Consequently, if the audio requires a response, the user cannot receive a timely reply.

[0042] To address the aforementioned problems in the existing technology, this application proposes an interactive audio recognition method, and the practical application scenarios of this method will be explained below.

[0043] Figure 1 Please refer to the following diagram illustrating the application scenario of the interactive audio recognition method provided in this application embodiment. Figure 1 The terminal device 110 may include, but is not limited to, mobile phones, wearable devices (such as smartwatches, smart bracelets, smart glasses, etc.), tablets, laptops, in-vehicle terminals, PCs (Personal Computers), virtual reality (VR) devices, augmented reality (AR) devices, smart screens, artificial intelligence (AI) speakers, headphones, terminals in industrial control, terminals in self-driving, terminals in remote medical surgery, terminals in smart grids, terminals in transportation safety, terminals in smart cities, terminals in smart homes, personal digital assistants (PDAs), etc. The functions implemented by this method can be achieved by the processor in the terminal device calling program code. Of course, the program code can be stored in computer storage media. Therefore, the terminal device includes at least a processor and storage media.

[0044] In this scenario, the terminal device 110 can identify the interactive audio after acquiring it, thereby determining whether the interactive audio is valid. The interactive audio can be, for example, words spoken by the user, such as "Please turn on the light" or "Turn off the speaker," etc., without any specific restrictions.

[0045] After receiving the interactive audio, the terminal device can execute the interactive audio recognition method provided in this embodiment to obtain the recognition result of the interactive audio, that is, to determine whether the terminal device needs to output a result (i.e., whether it needs to respond to the interactive audio), such as outputting voice or performing an operation. The output voice can be audio with preset content or a response audio obtained after semantic recognition based on the interactive audio; no specific limitations are imposed here. Performing an operation can be, for example, displaying an indicator light, emitting a prompt sound, or activating a function of the terminal device; no specific limitations are imposed here.

[0046] For example, if the recognition result of the interactive audio is determined to be a valid interaction, then the terminal device can perform the corresponding operation; if the recognition result of the interactive audio is determined to be an invalid interaction, then the terminal device does not need to perform the corresponding operation.

[0047] The following explains the implementation process of the interactive audio recognition method provided in the embodiments of this application based on the above-mentioned practical application scenarios.

[0048] Figure 2 Please refer to the flowchart illustrating the interactive audio recognition method provided in this application embodiment. Figure 2 The method includes:

[0049] S210: Acquire interactive audio.

[0050] The execution entity of this method can be, for example, Figure 1 In the application scenario shown, the terminal device can acquire interactive audio through a recording device, such as a microphone or other recording device. Alternatively, the interactive audio can be audio input to the terminal device from other devices, such as an audio file or other files containing audio, and the terminal device obtains the interactive audio based on that audio file or other files containing audio.

[0051] In some embodiments, the terminal device can first acquire audio from the environment or through other devices, and determine whether the audio is human voice. If so, the audio can be used as interactive audio; if not, the audio can be treated as background noise and left unprocessed or discarded. Alternatively, the terminal device can directly acquire human voice audio and filter out ambient audio using a preset algorithm.

[0052] Taking the example of a terminal device automatically collecting human voice audio from the environment, it can start recording when it detects human voice audio, and stop recording when no human voice audio is present in the environment for a certain period of time, thus obtaining interactive audio. Alternatively, it can start recording in response to user actions and use the audio collected within a preset time period as interactive audio.

[0053] Alternatively, during the recording process, voice judgment can be performed based on the current recording content to determine whether a complete sentence has been obtained. If so, the recording can be stopped to obtain interactive audio; if a complete sentence has not been obtained, the recording can continue until a complete sentence is obtained or no voice is recorded.

[0054] S220: Determine the recognition result of the interactive audio based on the audio recognition service in the audio recognition service system.

[0055] The audio recognition service system includes at least an invalid rejection service and a text processing service, and the recognition results include valid interactions and invalid interactions.

[0056] If the text processing service identifies the interaction as valid, the audio recognition result will also be considered valid.

[0057] It should be noted that the audio recognition service system can be a system used to identify interactive audio. Based on the services in the system, it can determine whether the interactive audio is a valid interaction or an invalid interaction. Valid interaction means that the interactive audio requires a response, while invalid interaction means that the interactive audio does not require a response.

[0058] For example, if the recognition result of the interactive audio is determined to be a valid interaction, a response audio can be output for the interactive audio; if the recognition result of the interactive audio is determined to be an invalid interaction, then no response audio needs to be output.

[0059] In some embodiments, the audio recognition service system may include an invalid rejection service and a text processing service. The invalid rejection service may identify interactive audio according to an invalid rejection algorithm to obtain a recognition result. For example, the interactive audio may be input into a neural network model in the invalid rejection algorithm, and the recognition result may be determined based on the output of the neural network model.

[0060] Text processing services can be service solutions that process the text corresponding to interactive audio. For example, interactive audio can be converted into text to obtain interactive text, and then the recognition result can be determined based on the content of the interactive text.

[0061] Text processing services can include various feasible methods, one of which is determining the recognition result of the interactive text based on the presence or absence of certain words. It is evident that because text processing services do not require the participation of neural network models during execution, but only need to determine the occurrence of words, their processing efficiency is higher than that of invalid rejection services.

[0062] Furthermore, for text processing services, depending on their type, it may not be possible to directly determine whether the interactive audio is a valid or invalid interaction; for example, it may be determined as a valid or invalid interaction, or it may be determined as an invalid or not invalid interaction. Different determination methods can be applied based on different types to obtain the corresponding results.

[0063] It should be noted that, due to the relatively high efficiency of text processing services, recognition results can be obtained faster than invalid rejection services in most cases, thus improving recognition efficiency.

[0064] In the embodiments of this application, the recognition result of interactive audio can be determined based on the recognition result of at least one audio recognition service.

[0065] For example, if the text processing service identifies the interaction as valid, the recognition result of the interactive audio can be determined as valid, thus eliminating the need for additional identification based on the invalid rejection service, thereby improving the efficiency of determining the recognition result of the interactive audio.

[0066] In the interactive audio recognition method provided in this application embodiment, the recognition result of interactive audio can be determined based on at least one audio recognition service in the audio recognition service system. Since the processing efficiency of the text processing service is higher than that of the invalid rejection service, the recognition result of the interactive audio can be determined as a valid interaction if the recognition result of the text processing service is a valid interaction. Compared with using only the invalid rejection service to determine the recognition result of interactive audio, the recognition result of interactive audio can be obtained more quickly and efficiently. Furthermore, based on the text processing service, the accuracy of the recognition result can be guaranteed to be high and the consumption of computing power can be reduced. That is, the terminal device can output the recognition result of the interactive audio more promptly.

[0067] It should be noted that the execution order of the various services in the audio recognition service system is not unique. The order in the implementation process can be determined according to the actual needs. Under different execution orders, the method for determining the recognition result of interactive audio can also be different.

[0068] The implementation process of the interactive audio recognition method when the recognition result of the text processing service is a valid interaction has been explained above. The following will explain the implementation process of the interactive audio recognition method when the recognition result of the text processing service is an invalid interaction, based on different execution orders.

[0069] Figure 3 Please refer to the flowchart illustrating the process of obtaining the recognition result of interactive audio provided in the embodiments of this application. Figure 3 The execution sequence may include: a text processing service, an invalidation rejection service, and a rhythm controller. The recognition results output by both the text processing service and the invalidation rejection service can be transmitted to the rhythm controller, which then determines the recognition result of the interactive audio. Additionally, the recognition result of the text processing service can also be transmitted to the invalidation rejection service.

[0070] The text processing service and the invalid rejection service can be performed sequentially. For example, the recognition result can be determined first based on the text processing service. If the recognition result of the text processing service is a valid interaction, then the recognition result of the interactive audio can be determined as a valid interaction. If the recognition result of the text processing service is not a valid interaction, then the recognition result can be determined based on the invalid rejection service, and the recognition result determined by the invalid rejection service can be used as the recognition result of the interactive audio.

[0071] In one embodiment, determining the recognition result of interactive audio based on the audio recognition service in the audio recognition service system includes: recognizing the interactive audio based on the text processing service; and determining the recognition result of the interactive audio based on the recognition result of the invalid rejection service if the recognition result of the text processing service is not a valid interaction.

[0072] It should be noted that the recognition result of interactive audio can be determined by a rhythm controller in the terminal device. The rhythm controller can receive recognition results from at least one audio recognition service in the audio recognition service system. For example, it can receive recognition results from a text processing service or recognition results from an invalid rejection service. The rhythm controller can then determine the recognition result of interactive audio based on these recognition results.

[0073] For example, please refer to Figure 3 If the text processing service identifies the interaction as valid, then invalid rejection service is not required, and the identification result of the interactive audio can be determined as valid interaction. If the text processing service does not identify the interaction as valid, then invalid rejection service can be performed, and the identification result of invalid rejection service can be obtained. The identification result of invalid rejection service can be used as the identification result of interactive audio. In other words, if the identification result of invalid rejection service is valid interaction, then the identification result of interactive audio can be determined as valid interaction; if the identification result of invalid rejection service is invalid interaction, then the identification result of interactive audio can be determined as invalid interaction.

[0074] In the interactive audio recognition method provided in this application embodiment, when the recognition result of the text processing service is not a valid interaction and the recognition result of the invalid rejection service is obtained, the recognition result of the interactive audio can be determined based on the recognition result of the invalid rejection service. Specifically, when the interaction is determined to be invalid based on the text processing service, further judgment is made based on the recognition result of the invalid rejection service, which increases the accuracy of the obtained interactive audio recognition result compared to using only the text processing service.

[0075] The following explains the process of obtaining recognition results based on a text processing service that includes a high-frequency word matching service.

[0076] Figure 4 For another flowchart illustrating the process of obtaining the recognition result of interactive audio provided in this application embodiment, please refer to... Figure 4 The text processing service includes a high-frequency word matching service, which is used to determine the recognition result based on the matching of words in the interactive audio with the high-frequency word list.

[0077] In some embodiments, the high-frequency vocabulary may include multiple words, each of which may be determined based on historical interaction data. For example, historical interaction data may be traversed to determine the frequency of words appearing in the text corresponding to each interaction audio in the case of effective interaction, and words with a frequency greater than a threshold may be included in the high-frequency vocabulary.

[0078] It should be noted that the high-frequency vocabulary list can be determined and entered by the user into the terminal device, transmitted to the terminal device by other devices, or obtained by the terminal device based on historical interaction data; no specific restrictions are imposed here.

[0079] In determining the recognition result based on the high-frequency word matching service, the following approach can be used: if there are words in the interactive audio that match the high-frequency word list, the recognition result of the high-frequency word matching service is determined to be a valid interaction; if there are no words in the interactive audio that match the high-frequency word list, the recognition result of the high-frequency word matching service is determined to be a invalid interaction.

[0080] In other words, if the words in the text corresponding to the interactive audio appear in the high-frequency vocabulary list, the recognition result of the high-frequency vocabulary matching service is determined to be a valid interaction; if the words in the text corresponding to the interactive audio do not appear in the high-frequency vocabulary list, the recognition result of the high-frequency vocabulary matching service is determined to be a invalid interaction, without specific restrictions.

[0081] It should be noted that the above matching method is only an example. In actual implementation, for example, if the hit rate of words in the high-frequency vocabulary list in the interactive audio is higher than the preset high-frequency threshold, the recognition result of the high-frequency vocabulary matching service can be determined as a valid interaction; if the hit rate of words in the high-frequency vocabulary list in the interactive audio is lower than the preset high-frequency threshold, the recognition result of the high-frequency vocabulary matching service can be determined as a non-valid interaction, etc. The above example is not a limitation.

[0082] For example, if the preset high-frequency threshold is 20%, and more than 20% of the words in the text corresponding to the interactive audio appear in the high-frequency vocabulary list, then the recognition result of the high-frequency vocabulary matching service can be determined as a valid interaction; correspondingly, if less than 20% of the words in the text corresponding to the interactive audio appear in the high-frequency vocabulary list, then the recognition result of the high-frequency vocabulary matching service can be determined as a invalid interaction.

[0083] In some embodiments, the following technical means can be used to determine the text corresponding to the interactive audio:

[0084] After the terminal device receives the interactive audio, it can perform ASR (Automatic Speech Recognition) processing on the audio to obtain the corresponding text. For example, if the interactive audio is a sentence entered by the user, the corresponding text can be obtained after ASR processing.

[0085] Figure 4 The diagram illustrates the process of determining the recognition results of interactive audio based on a high-frequency word matching service. Figure 4 The upper subgraph represents the case where the high-frequency word matching service is a valid interaction. If the recognition result of the high-frequency word matching service is a valid interaction, then the recognition result of the interactive audio can be determined to be a valid interaction. Figure 4 The lower half of the diagram represents the case where the high-frequency word matching service is not a valid interaction. If the recognition result of the high-frequency word matching service is not a valid interaction, an invalid rejection service can be performed. If the recognition result of the invalid rejection service is obtained within a preset time, the recognition result of the invalid rejection service will be used as the recognition result of the interactive audio.

[0086] In the interactive audio recognition method provided in this application embodiment, if the hit rate of words in the high-frequency vocabulary list in the interactive audio is higher than a preset high-frequency threshold, the recognition result of the high-frequency vocabulary matching service is determined to be a valid interaction; if the hit rate of words in the high-frequency vocabulary list in the interactive audio is lower than the preset high-frequency threshold, the recognition result of the high-frequency vocabulary matching service is determined to be a invalid interaction. By using the hit rate of the high-frequency vocabulary list, the recognition result can be obtained more quickly and the computational power consumption can be reduced, thereby improving the efficiency of obtaining the recognition result of the interactive audio.

[0087] In some embodiments, determining the recognition result of interactive audio based on the audio recognition service in the audio recognition service system includes: recognizing the interactive audio based on the text processing service; and determining the recognition result of the interactive audio based on the recognition result of the invalid rejection service if the recognition result of the text processing service is not an invalid interaction.

[0088] It should be noted that the above situation is different from the aforementioned method. It is an implementation method under the premise that the recognition result determined by the text processing service is an invalid interaction or not an invalid interaction. Taking the text processing service as a low-frequency word matching service as an example.

[0089] The following explains the process of obtaining recognition results based on a text processing service that includes a low-frequency word matching service.

[0090] Figure 5 For another flowchart illustrating the process of obtaining the recognition result of interactive audio provided in this application embodiment, please refer to... Figure 5The text processing service includes a low-frequency word matching service, which is used to determine the recognition result based on the matching of words in the interactive audio with the low-frequency word list.

[0091] In some embodiments, the low-frequency vocabulary may include multiple words, each of which may be determined based on historical interaction data. For example, historical interaction data may be traversed to determine the frequency of words appearing in the text corresponding to each interaction audio in the case of invalid interaction, and words with a frequency greater than a threshold may be included in the low-frequency vocabulary.

[0092] It should be noted that, similar to the acquisition method of the high-frequency vocabulary list, the low-frequency vocabulary list can be determined and entered by the user into the terminal device, or transmitted to the terminal device by other devices, or obtained by the terminal device based on historical interaction data, without specific restrictions here.

[0093] The process of determining the recognition result based on low-frequency word matching service can be achieved in the following ways:

[0094] If all words in the interactive audio match words from the low-frequency vocabulary list, the low-frequency vocabulary matching service is determined to be an invalid interaction; if there are words in the interactive audio that do not match words from the low-frequency vocabulary list, the high-frequency vocabulary matching service is determined to be a valid interaction.

[0095] For example, all words in the text corresponding to the interactive audio can be identified, and it can be determined whether each word has appeared in the low-frequency vocabulary list. If all words have appeared in the low-frequency vocabulary list, the recognition result of the low-frequency vocabulary matching service can be determined as an invalid interaction; correspondingly, as long as there are words that do not appear in the low-frequency vocabulary list, the recognition result of the low-frequency vocabulary matching service can be determined as not an invalid interaction.

[0096] It should be noted that the above matching method is only an example. In actual implementation, if the hit rate of words in the low-frequency vocabulary list in the interactive audio is higher than the preset low-frequency threshold, the recognition result of the low-frequency vocabulary matching service can be determined as an invalid interaction. For example, the preset low-frequency threshold can be 80%, that is, if more than 80% of the words in the text corresponding to the interactive audio appear in the low-frequency vocabulary list, the recognition result of the low-frequency vocabulary matching service can be determined as an invalid interaction; if less than 80% of the words in the text corresponding to the interactive audio appear in the low-frequency vocabulary list, the recognition result of the low-frequency vocabulary matching service can be determined as not an invalid interaction.

[0097] Figure 5The diagram illustrates the process of determining the recognition result of interactive audio based on low-frequency word matching service. If the recognition result of low-frequency word matching service is invalid interaction, invalid rejection service can be performed. If the recognition result of invalid rejection service is obtained within a preset time period, the recognition result of invalid rejection service is used as the recognition result of interactive audio. If the recognition result of invalid rejection service is not obtained within the preset time period, the recognition result of interactive audio is determined to be invalid interaction.

[0098] Alternatively, in actual implementation, if the recognition result of the low-frequency word matching service is invalid, the invalid rejection service can be omitted, and the recognition result of the interactive audio can be determined as invalid. One of the implementation methods can be selected to determine the recognition result of the interactive audio according to actual needs.

[0099] In the interactive audio recognition method provided in this application embodiment, if all words in the interactive audio match words from the low-frequency vocabulary list, the recognition result of the low-frequency vocabulary matching service is determined to be an invalid interaction; if there are words in the interactive audio that do not match words from the low-frequency vocabulary list, the recognition result of the high-frequency vocabulary matching service is determined to be not an invalid interaction. By utilizing the low-frequency vocabulary list matching results, recognition results can be obtained more quickly and computational power consumption can be reduced, thereby improving the efficiency of obtaining interactive audio recognition results.

[0100] The following explains another implementation process of the interactive audio recognition method when the recognition result of the text processing service is invalid interaction.

[0101] Figure 6 For another flowchart illustrating the process of obtaining the recognition result of interactive audio provided in this application embodiment, please refer to... Figure 6 The execution sequence may include: a text processing service, an invalid rejection service, and a rhythm controller. The recognition results output by the text processing service and the invalid rejection service can be transmitted to the rhythm controller, which then determines the recognition result of the interactive audio.

[0102] The text processing service and the invalid rejection service can be performed independently. For example, the recognition result can be determined based on the text processing service and the invalid rejection service respectively. If the recognition result of the text processing service is a valid interaction, the recognition result of the interactive audio can be determined as a valid interaction. If the recognition result of the text processing service is an invalid interaction, the recognition result of the interactive audio can be determined based on whether the recognition result of the invalid rejection service is received within a preset time.

[0103] In other words, based on the audio recognition service in the audio recognition service system, the recognition result of the interactive audio is determined, including: recognizing the interactive audio according to the text processing service and the invalid rejection service respectively; and determining the recognition result of the interactive audio based on whether the recognition result of the invalid rejection service is obtained within a preset time period when the recognition result of the text processing service is not a valid interaction.

[0104] It should be noted that this solution can also determine the recognition result of interactive audio through a rhythm controller.

[0105] For example, please refer to Figure 6 If the recognition result of the text processing service is a valid interaction, the recognition result of the interactive audio can be determined as a valid interaction regardless of whether the invalid rejection service is received; if the recognition result of the text processing service is an invalid interaction and no recognition result of the invalid rejection service is obtained within the preset time, the recognition result of the interactive audio is determined as an invalid interaction.

[0106] The preset duration can be set according to actual needs, such as 0.1 seconds; the starting point for timing can be set according to actual needs, such as after obtaining interactive audio or after receiving the recognition result from the text processing service, without specific restrictions.

[0107] In one embodiment, determining the recognition result of the interactive audio based on whether the recognition result of the invalid rejection service is obtained within a preset time period includes: if the recognition result of the invalid rejection service is obtained within the preset time period, determining the recognition result of the interactive audio based on the recognition result of the invalid rejection service; if the recognition result of the invalid rejection service is not obtained within the preset time period, determining the recognition result of the interactive audio based on the recognition result of the text processing service.

[0108] If the recognition result of the interactive audio is determined based on the recognition result of the text processing service, then if the recognition result of the text processing service is not a valid interaction, the recognition result of the interactive audio is determined to be an invalid interaction.

[0109] In the interactive audio recognition method provided in this application embodiment, if the recognition result of the text processing service is invalid and no recognition result of the invalid rejection service is obtained within a preset time period, the recognition result of the interactive audio is determined to be invalid. Specifically, if no recognition result of the invalid rejection service is received within the preset time period, the recognition result of the interactive audio can be directly determined to be invalid, avoiding a situation where the determination of the interactive audio recognition result is slow due to the slow processing of the invalid rejection service. In other words, compared to using only the text processing service, the efficiency of determining the recognition result of the interactive audio can be improved.

[0110] In some embodiments, the audio recognition service system further includes a dialogue management service, and the method further includes: when no recognition result of the text processing service and the recognition result of the invalid rejection service are obtained within a preset time period, but the recognition result of the dialogue management service is obtained, determining the recognition result of the interactive audio based on the recognition result of the dialogue management service.

[0111] The dialogue management service can identify the interactive audio based on the semantics of the text corresponding to it, thereby obtaining the recognition result. The following two scenarios illustrate how to determine the recognition result of the interactive audio when all three services are present simultaneously.

[0112] For one of the processes for determining the recognition result of interactive audio when dialogue management service, high-frequency word matching service and invalid rejection service are present at the same time, the following scheme can be adopted.

[0113] Figure 7 For another flowchart illustrating the process of obtaining the recognition result of interactive audio provided in this application embodiment, please refer to... Figure 7 The audio recognition service system in this solution may include dialogue management service, high-frequency word matching service, and invalid rejection service.

[0114] Figure 7 The diagram shows the logic of one feasible service implementation. After obtaining the interactive audio, the recognition result can be determined by the dialogue management service and the high-frequency word matching service. If the recognition result of the high-frequency word matching service is invalid, the recognition result can be determined by the invalid rejection service.

[0115] Figure 8 This is a logical diagram illustrating the recognition result of the interactive audio provided in the embodiments of this application. Please refer to the following: Figure 7 and Figure 8 In this scheme, after obtaining the interactive audio, the corresponding text can be obtained based on the aforementioned ASR service. Then, the recognition result can be determined by the dialogue management service and the high-frequency word matching service, respectively. If the high-frequency word matching service determines that the recognition result is a valid interaction, then the recognition result of the interactive audio can be determined as a valid interaction. If the high-frequency word matching service determines that the recognition result is not a valid interaction, then the invalid rejection service can determine the recognition result, and the recognition result of the invalid rejection service can be used as the recognition result of the interactive audio. If the recognition result of the high-frequency word matching service and the recognition result of the invalid rejection service are not obtained within a preset time, the recognition result of the interactive audio can be determined based on the recognition result of the dialogue management service.

[0116] for Figure 7 as well as Figure 8In one implementation shown, the slow recognition result of the invalid rejection service can be effectively avoided, which leads to a slow recognition result of the interactive audio. This improves the efficiency of the recognition result of the interactive audio. Furthermore, the high-frequency word matching service can also ensure the accuracy of the recognition result and reduce the consumption of computing power.

[0117] For another process of determining the recognition result of interactive audio when dialogue management service, high-frequency word matching service and invalid rejection service exist simultaneously, the following scheme can be adopted.

[0118] Figure 9 For another flowchart illustrating the process of obtaining the recognition result of interactive audio provided in this application embodiment, please refer to... Figure 9 The audio recognition service system in this solution may include dialogue management service, high-frequency word matching service, and invalid rejection service.

[0119] Figure 9 The diagram shows the logic of one possible service implementation. After obtaining the interactive audio, the recognition result can be determined through dialogue management service, high-frequency word matching service, and invalid rejection service, and priorities can be set for each service. Within a preset time period, the recognition result of the service with higher priority is taken as the recognition result of the interactive audio.

[0120] The priorities, from highest to lowest, are as follows: invalid rejection service, high-frequency word matching service, and dialogue management service.

[0121] Figure 10 This is another logical diagram illustrating the acquisition of interactive audio recognition results provided in this application embodiment. Please refer to the reference [reference needed]. Figure 9 and Figure 10 In this scheme, after obtaining the interactive audio, the corresponding text can be obtained based on the aforementioned ASR service. The recognition result can then be determined by the dialogue management service, the high-frequency word matching service, and the invalid rejection service. If the invalid rejection service's recognition result is obtained within a preset time period, it is used as the recognition result of the interactive audio. If the high-frequency word matching service's recognition result is valid within the preset time period and no invalid rejection service's recognition result is obtained, the interactive audio's recognition result is determined to be valid. If no high-frequency word matching service's recognition result is obtained within the preset time period, or if the obtained high-frequency word matching service's recognition result is not valid, the dialogue management service's recognition result is used as the interactive audio's recognition result.

[0122] for Figure 9 as well as Figure 10In one implementation shown, the slow recognition result of the invalid rejection service can be effectively avoided, which leads to a slow recognition result of the interactive audio. This improves the efficiency of the recognition result of the interactive audio. Furthermore, the high-frequency word matching service can also ensure the accuracy of the recognition result and reduce the consumption of computing power.

[0123] It should be noted that for both of the above feasible solutions, there will be different service output speeds. Based on the order of the output results, there are several feasible implementation methods.

[0124] In one embodiment, if the recognition results of the text processing service, the invalid rejection service, and the dialogue management service are not obtained within a preset time period, the recognition result of the interactive audio is determined to be an invalid interaction.

[0125] In another embodiment, if the recognition result of the text processing service is obtained as a valid interaction within a preset time period, but the recognition result of the dialogue management service is not obtained, the recognition result of the interactive audio is determined to be a valid interaction.

[0126] In another embodiment, if the identification result of the invalid rejection service is obtained within a preset time period, but the identification result of the dialogue management service is not obtained, the identification result of the interactive audio is determined based on the identification result of the invalid rejection service.

[0127] Based on the above embodiments, the situations that exist during implementation can be categorized as follows:

[0128] Scenario 1: If no recognition result is received from any service within the preset time, in order to avoid no output for a long time, the recognition result of the interactive audio can be determined as invalid interaction.

[0129] Scenario 2: If only the recognition result of the dialogue management service is received within the preset time, the recognition result of the dialogue management service can be used as the recognition result of the interactive audio since the recognition results of the other two services are not received.

[0130] Scenario 3: If only the recognition result of the text processing service is received within the preset time, and the recognition result is a valid interaction, since the recognition results of the other two services are not received, it can be determined that the recognition result of the interactive audio is a valid interaction.

[0131] Scenario 4: Only the recognition result of the text processing service is received within the preset time, and the recognition result is not a valid interaction. Since the recognition results of the other two services are not received, it can be determined that the recognition result of the interactive audio is an invalid interaction.

[0132] Scenario 5: If only the identification result of the invalid rejection service is received within the preset time, and no identification results of the other two services are received, the identification result of the invalid rejection service can be used as the identification result of the interactive audio.

[0133] Scenario 6: If two or three service recognition results are received within a preset time, the recognition result of the interactive audio can be determined based on the recognition result of the service with higher priority.

[0134] In actual implementation, the recognition result of interactive audio can be determined according to the specific circumstances and methods.

[0135] The following explains the implementation process of determining the identification result based on the invalid rejection service provided in the embodiments of this application.

[0136] Figure 11 Please refer to the flowchart illustrating the process of determining the identification result based on the invalid rejection service provided in this application embodiment. Figure 11 The steps for determining the identification result based on the invalid rejection service include:

[0137] S310: Obtain the speech features of the interactive audio and the text features of the interactive text.

[0138] Interactive text includes text obtained from interactive audio recognition.

[0139] In some embodiments, the interactive audio can be processed by the ASR service described above to obtain interactive text, and then the speech features of the interactive audio and the text features of the interactive text can be obtained.

[0140] It should be noted that after processing by the ASR service, audio recognition can be achieved, thereby enabling audio-to-text transcription and obtaining the text transcribed from the interactive audio.

[0141] In the process of determining the corresponding features, neural networks can be used for feature extraction. For example, interactive audio can be input into a speech neural network to obtain speech features; interactive text and other related interactive text can be input into a text neural network to obtain text features.

[0142] Specifically, speech neural networks can use convolutional neural networks (CNNs), while text neural networks can use long short-term memory (LSTM) networks. It should be noted that in actual implementation, the appropriate neural network can be chosen based on specific needs, such as deep neural networks, and is not limited to the two examples mentioned above.

[0143] S320: Obtain the rejection probability based on speech features and text features.

[0144] In some embodiments, the rejection probability can be the probability that the recognition result of the interactive audio is a valid interaction, as determined by the invalid rejection service.

[0145] After obtaining the speech features and text features respectively, a linear weighted sum can be performed to obtain the rejection probability. The specific calculation formula is as follows:

[0146] P = X * P1 + (1 - X) * P2;

[0147] Where X represents the weight of the speech features, P1 represents the speech features, P2 represents the text features, and P represents the final weighted result. The speech features can be obtained based on a speech neural network, and the text features can be obtained based on a text neural network. The weights of the speech features can be determined based on historical experience values, without specific restrictions here.

[0148] It should be noted that during the linear weighting process, the feature selection can be weighted frame by frame. After obtaining the above weighting result, the weighted result P can be input into the linear prediction layer to predict the rejection probability, and finally obtain the rejection probability.

[0149] S330: If the rejection probability is less than the target threshold, the recognition result is determined to be a valid interaction.

[0150] S340: If the rejection probability is greater than or equal to the target threshold, the recognition result is determined to be an invalid interaction.

[0151] In some embodiments, after obtaining the rejection probability, the recognition result can be determined based on the magnitude of the rejection probability. For example, the rejection probability can be a percentage, and the target threshold can be 50%. If the rejection probability is greater than or equal to 50%, the recognition result can be determined as an invalid interaction; correspondingly, if the rejection probability is less than 50%, the recognition result can be determined as a valid interaction.

[0152] In the interactive audio recognition method provided in this application embodiment, during the process of determining the recognition result based on the invalid rejection service, the speech features of the interactive audio and the text features of the interactive text can be obtained; based on the speech features and text features, the rejection probability is obtained; if the rejection probability is less than a target threshold, the recognition result is determined to be a valid interaction; if the rejection probability is greater than or equal to the target threshold, the recognition result is determined to be an invalid interaction. By combining the calculation of speech features and text features and performing linear weighting processing, the recognition result of the invalid rejection service can be determined more accurately, improving the accuracy of the output recognition result.

[0153] Let's explain below. Figure 11The scheme shown illustrates the process steps used to determine the probability of rejection during its implementation.

[0154] Figure 12 This is another flowchart illustrating the process of determining the identification result based on the invalid rejection service provided in this application embodiment. Please refer to the attached diagram. Figure 11 and Figure 12 , can be based on Figure 12 The specific steps shown are used to obtain the rejection probability.

[0155] It should be understood that although the steps in the above flowcharts are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the above flowcharts may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0156] The following explains the implementation process of the terminal device determining the recognition result of the interactive audio and then outputting it accordingly in the embodiments of this application.

[0157] Figure 13 Please refer to the schematic diagram of the terminal device outputting the recognition result according to the embodiment of this application. Figure 13 , Figure 13 The diagram illustrates one possible output method for a terminal device with a display interface, which can display interactive audio and responses to the interactive audio.

[0158] Example, Figure 13 The message "Please turn on the living room light" shown is the interactive audio; "Already on" is the text output based on the recognition result of the interactive audio. In addition to outputting the above text, it can also output the corresponding voice and perform the corresponding operation, that is, control the living room light to turn on via the Internet.

[0159] The following explains another implementation process in this application embodiment where the terminal device outputs the corresponding result after determining the recognition result of the interactive audio.

[0160] Figure 14 Please refer to another schematic diagram of the terminal device outputting the recognition result according to the embodiment of this application. Figure 14 , Figure 14 The diagram shows a feasible way for a terminal device without a display interface to output results, which can be done by illuminating an indicator light.

[0161] Example, Figure 14 If the indicator light is on, it means that the interaction audio recognition result is a valid interaction; if the indicator light is off, it means that the interaction audio recognition result is an invalid interaction.

[0162] It should be noted that, Figure 14 The example shown is only one example. The meaning of the flashing pattern of the light can be set according to actual needs. Devices such as buzzers, motors, and speakers can also be used to output the result. The above-mentioned indicator light is not a limitation.

[0163] Based on the foregoing embodiments, this application provides an interactive video recognition device, which includes various modules and units included in each module, and can be implemented by a processor; of course, it can also be implemented by specific logic circuits; in the implementation process, the processor can be a central processing unit (CPU), microprocessor (MPU), digital signal processor (DSP) or field programmable gate array (FPGA), etc.

[0164] Figure 15 Please refer to the schematic diagram of the interactive audio recognition device provided in the embodiments of this application. Figure 15 The device includes: an acquisition module 111 and an identification module 112;

[0165] Module 111 is used to acquire interactive audio;

[0166] The recognition module 112 is used to determine the recognition result of interactive audio based on the audio recognition service in the audio recognition service system. The audio recognition service system includes at least an invalid rejection service and a text processing service. The processing efficiency of the text processing service is higher than that of the invalid rejection service. The recognition result includes valid interaction and invalid interaction. When the recognition result of the text processing service is a valid interaction, the recognition result of the interactive audio is determined to be a valid interaction.

[0167] In some embodiments, the recognition module 112 is specifically used to recognize interactive audio based on a text processing service; if the recognition result of the text processing service is not a valid interaction, the recognition result of the interactive audio is determined based on the recognition result of the invalid rejection service.

[0168] In some embodiments, the text processing service includes a high-frequency word matching service, which is used to determine the recognition result based on the matching status of words in the interactive audio with words in the high-frequency word list. When there are words in the interactive audio that match words in the high-frequency word list, the recognition result of the high-frequency word matching service is determined to be a valid interaction; when there are no words in the interactive audio that match words in the high-frequency word list, the recognition result of the high-frequency word matching service is determined to be a invalid interaction.

[0169] In some embodiments, the recognition module 112 is specifically used to recognize interactive audio based on a text processing service; if the recognition result of the text processing service is not an invalid interaction, the recognition result of the interactive audio is determined based on the recognition result of the invalid rejection service.

[0170] In some embodiments, the text processing service includes a low-frequency word matching service, which determines the recognition result based on the matching status of words in the interactive audio with words in the low-frequency word list. When all words in the interactive audio match words in the low-frequency word list, the recognition result of the low-frequency word matching service is determined to be an invalid interaction; when there are words in the interactive audio that do not match words in the low-frequency word list, the recognition result of the high-frequency word matching service is determined to be not an invalid interaction.

[0171] In some embodiments, the recognition module 112 is specifically used to recognize interactive audio based on the text processing service and the invalid rejection service respectively; if the recognition result of the text processing service is not a valid interaction, the recognition result of the interactive audio is determined based on whether the recognition result of the invalid rejection service is obtained within a preset time period.

[0172] In some embodiments, the identification module 112 is specifically used to determine the identification result of the interactive audio based on the identification result of the invalid rejection service if the identification result of the invalid rejection service is obtained within a preset time period; and to determine the identification result of the interactive audio based on the identification result of the text processing service if the identification result of the invalid rejection service is not obtained within a preset time period.

[0173] In some embodiments, the recognition module 112 is specifically used to determine that the recognition result of the interactive audio is an invalid interaction when the recognition result of the text processing service is not a valid interaction.

[0174] In some embodiments, the identification module 112 is further configured to determine the identification result of the interactive audio based on the identification result of the dialogue management service when no identification result of the text processing service or the invalid rejection service is obtained within a preset time period, but the identification result of the dialogue management service is obtained.

[0175] In some embodiments, the identification module 112 is further configured to determine that the identification result of the interactive audio is invalid interaction if the identification result of the text processing service, the identification result of the invalid rejection service, and the identification result of the dialogue management service are not obtained within a preset time period.

[0176] In some embodiments, the recognition module 112 is further configured to determine that the recognition result of the interactive audio is a valid interaction when the recognition result of the text processing service is obtained within a preset time period and the recognition result of the dialogue management service is not obtained.

[0177] In some embodiments, the identification module 112 is further configured to determine the identification result of the interactive audio based on the identification result of the invalid rejection service when the identification result of the invalid rejection service is obtained within a preset time period and the identification result of the dialogue management service is not obtained.

[0178] In some embodiments, when performing identification based on the invalid rejection service, the identification module 112 can obtain the speech features of the interactive audio and the text features of the interactive text, the interactive text including the text obtained from the interactive audio identification; obtain the rejection probability based on the speech features and the text features; if the rejection probability is less than the target threshold, determine the identification result as a valid interaction; if the rejection probability is greater than or equal to the target threshold, determine the identification result as an invalid interaction.

[0179] The descriptions of the above device embodiments are similar to those of the above method embodiments, and have similar beneficial effects. For technical details not disclosed in the device embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0180] It should be noted that, in the embodiments of this application... Figure 15 The module division of the interactive audio recognition device shown is illustrative and represents only one logical functional division; in actual implementation, other division methods may be used. Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, exist as separate physical units, or be integrated into one unit with two or more units. The integrated units can be implemented in hardware, as software functional units, or a combination of both.

[0181] It should be noted that, in the embodiments of this application, if the above-described methods are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, or the parts that contribute to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause an electronic device to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), magnetic disks, or optical disks. Thus, the embodiments of this application are not limited to any specific hardware and software combination.

[0182] Figure 16 Please refer to the schematic diagram of the computer device provided in the embodiments of this application. Figure 16 This application provides a computer device, which can be the aforementioned terminal device, and its internal structure diagram can be as follows. Figure 16 As shown, the computer device includes a processor 420, memory, and a network interface 440 connected via a system bus 410. The processor 420 provides computing and control capabilities. The memory includes a non-volatile storage medium 431 and internal memory 432. The non-volatile storage medium 431 stores the operating system, computer programs, and a database. The internal memory 432 provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database is used to store data. The network interface 440 is used to communicate with external terminals via a network connection. When the computer program is executed by the processor 420, it implements the methods described above.

[0183] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the method provided in the above embodiments.

[0184] This application also provides a computer program product containing instructions that, when run on a computer, cause the computer to perform the steps in the method provided in the above-described method embodiments.

[0185] Those skilled in the art will understand that Figure 16 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0186] In one embodiment, the interactive audio recognition device provided in this application can be implemented as a computer program, which can be configured as follows: Figure 16 The device operates on the computer device shown. The memory of the computer device can store the various program modules that make up the above-described apparatus. The computer program, composed of the various program modules, causes the processor to execute the steps of the methods in the various embodiments of this application described in this specification.

[0187] It should be noted that the descriptions of the storage medium and device embodiments above are similar to the descriptions of the method embodiments above, and have similar beneficial effects. For technical details not disclosed in the storage medium, storage medium, and device embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0188] It should be understood that the phrases "one embodiment," "an embodiment," or "some embodiments" mentioned throughout the specification mean that a specific feature, structure, or characteristic related to an embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment," "in one embodiment," or "in some embodiments" appearing throughout the specification do not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the above-described embodiments are merely for descriptive purposes and do not represent the superiority or inferiority of the embodiments. The descriptions of the various embodiments above tend to emphasize the differences between the various embodiments; their similarities or commonalities can be referred to mutually, and for the sake of brevity, they will not be repeated here.

[0189] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that there can be three kinds of relationships. For example, object A and / or object B can represent three situations: object A exists alone, object A and object B exist simultaneously, and object B exists alone.

[0190] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0191] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The embodiments described above are merely illustrative. For example, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple modules or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or modules can be electrical, mechanical, or other forms.

[0192] The modules described above as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules. They may be located in one place or distributed across multiple network units. Some or all of the modules may be selected to achieve the purpose of this embodiment according to actual needs.

[0193] In addition, each functional module in the various embodiments of this application can be integrated into one processing unit, or each module can be a separate unit, or two or more modules can be integrated into one unit; the integrated modules can be implemented in hardware or in the form of hardware plus software functional units.

[0194] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as mobile storage devices, read-only memory (ROM), magnetic disks, or optical disks.

[0195] Alternatively, if the integrated units described above are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, or the parts that contribute to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause an electronic device to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0196] The methods disclosed in the several method embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments.

[0197] The features disclosed in the several product embodiments provided in this application can be arbitrarily combined without conflict to obtain new product embodiments.

[0198] The features disclosed in the several method or device embodiments provided in this application can be arbitrarily combined without conflict to obtain new method or device embodiments.

[0199] The above description is merely an embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for recognizing interactive audio, characterized in that, The method is applied to a terminal device, and the method includes: Get interactive audio; Based on the audio recognition service in the audio recognition service system, the recognition result of the interactive audio is determined. The audio recognition service system includes at least an invalid rejection service and a text processing service. The processing efficiency of the text processing service is higher than that of the invalid rejection service. The recognition result includes valid interactions and invalid interactions. The text processing service includes a high-frequency word matching service and a low-frequency word matching service. If the recognition result of the text processing service is a valid interaction, the recognition result of the interactive audio will be determined as a valid interaction.

2. The method according to claim 1, characterized in that, The audio recognition service in the audio recognition service system determines the recognition result of the interactive audio, including: Based on the text processing service, the interactive audio is recognized; If the recognition result of the text processing service is not a valid interaction, the recognition result of the interactive audio is determined based on the recognition result of the invalid rejection service.

3. The method according to claim 2, characterized in that, When the text processing service is a high-frequency word matching service, the high-frequency word matching service is used to determine the recognition result based on the matching status of words in the interactive audio with the high-frequency word list. When there are words in the interactive audio that match the high-frequency word list, the recognition result of the high-frequency word matching service is determined to be a valid interaction; when there are no words in the interactive audio that match the high-frequency word list, the recognition result of the high-frequency word matching service is determined to be a invalid interaction.

4. The method according to claim 1, characterized in that, The audio recognition service in the audio recognition service system determines the recognition result of the interactive audio, including: Based on the text processing service, the interactive audio is recognized; If the recognition result of the text processing service is not an invalid interaction, the recognition result of the interactive audio is determined based on the recognition result of the invalid rejection service.

5. The method according to claim 4, characterized in that, When the text processing service is a low-frequency word matching service, the low-frequency word matching service is used to determine the recognition result based on the matching status of words in the interactive audio with words in the low-frequency word list. When all words in the interactive audio match words in the low-frequency word list, the recognition result of the low-frequency word matching service is determined to be an invalid interaction; when there are words in the interactive audio that do not match words in the low-frequency word list, the recognition result of the low-frequency word matching service is determined to be a non-invalid interaction.

6. The method according to claim 1, characterized in that, The audio recognition service in the audio recognition service system determines the recognition result of the interactive audio, including: The interactive audio is identified based on the text processing service and the invalid rejection service, respectively. If the recognition result of the text processing service is not a valid interaction, the recognition result of the interactive audio is determined based on whether the recognition result of the invalid rejection service is obtained within a preset time period.

7. The method according to claim 6, characterized in that, The step of determining the recognition result of the interactive audio based on whether the recognition result of the invalid rejection service is obtained within a preset time period includes: If the identification result of the invalid rejection service is obtained within a preset time period, the identification result of the interactive audio is determined based on the identification result of the invalid rejection service; If the invalid rejection service does not obtain a recognition result within a preset time period, the recognition result of the interactive audio is determined based on the recognition result of the text processing service.

8. The method according to claim 7, characterized in that, Determining the recognition result of the interactive audio based on the recognition result of the text processing service includes: If the recognition result of the text processing service is not a valid interaction, the recognition result of the interactive audio is determined to be an invalid interaction.

9. The method according to any one of claims 1-8, characterized in that, The audio recognition service system also includes a dialogue management service, and the method further includes: If the recognition result of the text processing service and the recognition result of the invalid rejection service are not obtained within the preset time period, but the recognition result of the dialogue management service is obtained, the recognition result of the interactive audio is determined according to the recognition result of the dialogue management service.

10. The method according to any one of claims 1-8, characterized in that, The audio recognition service system also includes a dialogue management service, and the method further includes: If the recognition results of the text processing service, the invalid rejection service, and the dialogue management service are not obtained within a preset time period, the recognition result of the interactive audio is determined to be an invalid interaction.

11. The method according to any one of claims 1-8, characterized in that, The audio recognition service system also includes a dialogue management service, and the method further includes: If the recognition result of the text processing service is obtained within a preset time period and the recognition result of the dialogue management service is not obtained, the recognition result of the interactive audio is determined to be a valid interaction.

12. The method according to any one of claims 1-8, characterized in that, The audio recognition service system also includes a dialogue management service, and the method further includes: If the identification result of the invalid rejection service is obtained within a preset time period, but the identification result of the dialogue management service is not obtained, the identification result of the interactive audio is determined based on the identification result of the invalid rejection service.

13. The method according to any one of claims 1-8, characterized in that, Determining the identification result based on the invalid rejection service includes: The speech features of the interactive audio and the text features of the interactive text are obtained, wherein the interactive text includes the text obtained from the interactive audio recognition. Based on the aforementioned speech and text features, the rejection probability is obtained; If the rejection probability is less than the target threshold, the recognition result is determined to be a valid interaction; if the rejection probability is greater than or equal to the target threshold, the recognition result is determined to be an invalid interaction.

14. An interactive audio recognition device, characterized in that, The device is applied to a terminal device, and the device includes: an acquisition module and an identification module; The acquisition module is used to acquire interactive audio; The recognition module is used to determine the recognition result of the interactive audio based on the audio recognition service in the audio recognition service system. The audio recognition service system includes at least an invalid rejection service and a text processing service. The processing efficiency of the text processing service is higher than that of the invalid rejection service. The recognition result includes valid interactions and invalid interactions. The text processing service includes a high-frequency word matching service and a low-frequency word matching service. If the recognition result of the text processing service is a valid interaction, the recognition result of the interactive audio will be determined as a valid interaction.

15. A computer device comprising a memory and a processor, the memory storing a computer program executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the method according to any one of claims 1 to 13.

16. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 13.

Images