Voice interaction module awakening method and device, equipment, medium and product

Abstract

The invention discloses a voice interaction module awakening method and device, equipment, a medium and a product. The method comprises the steps of obtaining voice information of a user in response to voice input of the user; performing feature extraction on the voice information to obtain wake-up words, voiceprint features and semantic features of the voice information; under the condition that the wake-up word is a preset wake-up word, determining a user type of the user according to voiceprint features and semantic features of the voice information; and based on the user type and a preset mapping relationship between the user type and a voice interaction module, waking up the voice interaction module corresponding to the user type. According to the embodiment of the invention, under the condition that one electronic device has a plurality of voice interaction models, the corresponding voice interaction model is flexibly awakened, and the user experience is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle control and particularly to a method, device, equipment, medium and product for waking up a voice interaction module. Background Art

[0002] With the development of technology, more and more electronic devices are integrated with multiple voice interaction modules to improve the intelligence of electronic devices.

[0003] In the related art, there may be multiple interaction models on a vehicle, such as a large model for adult chatting and a large model for child dialogue. The adult chatting model is used for voice interaction between adults and voice interaction models, and the large model for child dialogue is used for language interaction between children and voice interaction models. To wake up each interaction model, a corresponding wake-up word is required to wake up the interaction model.

[0004] In the above-mentioned related technologies, when there are multiple voice interaction models on an electronic device, each voice interaction model requires a specific wake-up word to wake up the voice interaction model. The voice interaction model can only be started after receiving the specific wake-up word. The wake-up method is single and the user experience is poor. Summary of the invention

[0005] The embodiments of the present application provide a method, apparatus, device, medium and product for waking up a voice interaction module, which can flexibly wake up the corresponding voice interaction model when there are multiple voice interaction models in an electronic device, thereby improving the user experience.

[0006] On the one hand, an embodiment of the present application provides a method for waking up a voice interaction module, the method comprising:

[0007] In response to a user's voice input, obtaining the user's voice information;

[0008] Extracting features of the voice information to obtain a wake-up word, a voiceprint feature, and a semantic feature of the voice information;

[0009] When the wake-up word is a preset wake-up word, determining the user type of the user according to the voiceprint features and semantic features of the voice information;

[0010] Based on the user type and the preset mapping relationship between the user type and the voice interaction module, the voice interaction module corresponding to the user type is awakened.

[0011] Optionally, according to the semantic features, determining a relevance weight and a coherence weight corresponding to the voice information, the coherence weight being used to represent the degree of coherence of the text in the voice information, and the relevance weight being used to represent the degree of relevance of the text in the voice information;

[0012] Determining a voiceprint weight corresponding to the voice information according to the voiceprint feature;

[0013] A user type of the user is determined based on the relevance weight, the consistency weight, and the voiceprint weight.

[0014] Optionally, determining the user type of the user according to the voiceprint feature and the semantic feature of the voice information includes:

[0015] Determining the relevance weight corresponding to the voice information according to the semantic feature includes:

[0016] Extracting keyword information of the semantic feature;

[0017] Comparing the keyword information with a first preset text in a preset text library to obtain a comparison result;

[0018] In the case where the comparison result determines that there is a first preset text corresponding to the keyword information, determining an arrangement position of the keyword information according to the first preset text sequence;

[0019] Determining first weight information corresponding to the keyword information according to the arrangement position;

[0020] A correlation weight is determined according to the first weight information and a first preset weight coefficient.

[0021] Optionally, determining the coherence weight corresponding to the speech information according to the semantic feature includes:

[0022] Acquire the word interval time, the total word count and the number of occurrences of a second preset text of the keyword information, where the second preset text is used to represent words that interrupt the coherence of the voice information;

[0023] Determining the occurrence probability of the second preset text according to the total number of words and the number of occurrences of the second preset text;

[0024] Determining a first coherence score according to the word interval time and the preset interval time;

[0025] determining a second coherence score according to the occurrence probability and a preset probability threshold;

[0026] A coherence weight is determined based on the first coherence score, the second coherence score, and a second preset weight coefficient.

[0027] Optionally, determining a voiceprint weight corresponding to the voice information according to the voiceprint feature includes:

[0028] Inputting the voiceprint feature into a preset voiceprint model, and using the preset voiceprint model to determine the sound frequency corresponding to the voiceprint feature;

[0029] Determining third weight information of the sound frequency by using the preset voiceprint model;

[0030] The voiceprint weight is determined according to the third weight information and the third preset weight coefficient.

[0031] Optionally, before inputting the voiceprint feature into a preset voiceprint model and using the preset voiceprint model to determine the sound frequency corresponding to the voiceprint feature, the method further includes:

[0032] Obtain voiceprint information for multiple preset age groups;

[0033] Extracting voiceprint features from the voiceprint information to obtain voiceprint features for each preset age group;

[0034] A preset voiceprint model is established based on each preset age group and its corresponding voiceprint features.

[0035] Optionally, determining the user type of the user based on the relevance weight, the coherence weight, and the voiceprint weight includes:

[0036] Determining a total voice weight of the voice information based on a sum of the relevance weight, the coherence weight, and the voiceprint weight;

[0037] When the total voice weight is greater than the preset voice weight, the user type of the user is determined to be a child;

[0038] When the total voice weight is not greater than the preset voice weight, the user type of the user is determined to be an adult.

[0039] Optionally, before extracting features from the voice information to obtain voiceprint features and semantic features of the voice information, the method further includes:

[0040] Performing signal preprocessing on the voice information to obtain processed voice information, wherein the signal preprocessing includes at least one of amplification, filtering, denoising and automatic gain control;

[0041] The extracting features of the voice information to obtain voiceprint features and semantic features of the voice information includes:

[0042] Feature extraction is performed on the processed voice information to obtain voiceprint features and semantic features of the voice information.

[0043] On the other hand, an embodiment of the present application provides a device for waking up a voice interaction module, the device comprising:

[0044] An acquisition module, used for acquiring the user's voice information in response to the user's voice input;

[0045] A feature extraction module, used to extract features from the voice information to obtain a wake-up word, voiceprint features, and semantic features of the voice information;

[0046] a determination module, configured to determine the user type of the user according to the voiceprint features and semantic features of the voice information when the wake-up word is a preset wake-up word;

[0047] A wake-up module is used to wake up the voice interaction module corresponding to the user type based on the user type and a preset mapping relationship between the user type and the voice interaction module.

[0048] In another aspect, an embodiment of the present application provides an electronic device, the device comprising: a processor and a memory storing computer program instructions;

[0049] When the processor executes the computer program instructions, the method for waking up the voice interaction module as described in the first aspect is implemented.

[0050] On the other hand, an embodiment of the present application provides a computer storage medium, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the method for waking up the voice interaction module as described in the first aspect is implemented.

[0051] On the other hand, an embodiment of the present application provides a computer program product. When the instructions in the computer program product are executed by a processor of an electronic device, the electronic device executes the wake-up method of the voice interaction module as described in the first aspect.

[0052] The wake-up method, device, equipment, medium and product of the voice interaction module in the embodiments of the present application can, during the process of waking up the voice interaction module, first input voice information through the user's voice, and then perform feature extraction on the voice information to obtain the wake-up word, voiceprint feature and semantic feature of the voice information. Among them, when the wake-up word is a preset wake-up word, it means that the voice interaction module needs to be woken up, and the user type of the user needs to be determined based on the voiceprint feature and the semantic feature, and then the corresponding voice interaction module is woken up according to the user type. That is, in the process of waking up the corresponding voice interaction module, the voice interaction module can be flexibly woken up according to the user's voice information, thereby improving the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solution of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0054] Figure 1 It is a flowchart of a method for waking up a voice interaction module provided by an embodiment of the present application;

[0055] Figure 2 is a flowchart of a method for determining a user type provided by an embodiment of the present application;

[0056] Figure 3 is a structural schematic diagram of a wake-up device for a voice interaction module provided in another embodiment of the present application;

[0057] Figure 4 It is a structural schematic diagram of an electronic device provided in yet another embodiment of the present application. DETAILED DESCRIPTION

[0058] The features and exemplary embodiments of various aspects of the present application will be described in detail below. In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application, rather than to limit the present application. For those skilled in the art, the present application can be implemented without the need for some of these specific details. The following description of the embodiments is only to provide a better understanding of the present application by illustrating the examples of the present application.

[0059] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the text "include..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0060] In order to solve the problems of the prior art, the embodiments of the present application provide a method, device, equipment, medium and product for waking up a voice interaction module. In the process of waking up the voice interaction module, the user first inputs voice information, and then extracts features of the voice information to obtain the wake-up word, voiceprint features and semantic features of the voice information. In the case where the wake-up word is a preset wake-up word, it means that the voice interaction module needs to be woken up, and the user type of the user needs to be determined based on the voiceprint features and semantic features, and then the corresponding voice interaction module is woken up according to the user type, that is, in the process of waking up the corresponding voice interaction module, the voice interaction module can be flexibly woken up according to the user's voice information, thereby improving the user experience.

[0061] The following first introduces the wake-up method of the voice interaction module provided in the embodiment of the present application.

[0062] Figure 1 FIG. 1 is a flow chart showing a method for waking up a voice interaction module according to an embodiment of the present application. Figure 1 As shown, the wake-up method of the voice interaction module may include S101-S104:

[0063] S101, in response to a user's voice input, obtaining the user's voice information.

[0064] In an embodiment of the present application, for ease of introduction, an electronic device including at least two voice interaction modules is taken as an example for explanation. In the process of waking up the voice interaction module, the user first performs voice input so that the electronic device can obtain the user's voice intention, that is, the voice information spoken by the user, so as to wake up the corresponding voice interaction module according to the user's voice information.

[0065] S102, extracting features from the voice information to obtain a wake-up word, voiceprint features, and semantic features of the voice information.

[0066] In some embodiments, in order to facilitate recognition of voice information and to more accurately wake up the voice interaction module, before S102, the method may further include:

[0067] Performing signal preprocessing on the voice information to obtain processed voice information;

[0068] S102 may specifically include:

[0069] The processed speech information is subjected to feature extraction to obtain the voiceprint features and semantic features of the speech information.

[0070] In this embodiment, after acquiring the voice information, the voice signal of the voice information can be preprocessed to obtain processed voice information, wherein the signal preprocessing includes at least one of amplification, filtering, denoising and automatic gain control.

[0071] Specifically, in order to improve the performance and accuracy of the speech recognition system, the speech information can be filtered for subsequent processing; in addition, since the speech information needs to be sampled during the digitization process, filtering can be performed to prevent high-frequency signals from being mixed with low-frequency signals to ensure that the processed speech information does not contain other audio signals; the gain of the speech signal can also be adjusted to ensure that a relatively stable amplitude can be maintained at different volumes.

[0072] In some other embodiments, analog-to-digital conversion may be performed or the high-frequency portion of the voice information may be emphasized through pre-emphasis to remove the influence of lip radiation and increase the high-frequency resolution of the voice information.

[0073] S103: When the wake-up word is a preset wake-up word, determine the user type of the user according to the voiceprint features and semantic features of the voice information.

[0074] In this embodiment, when the wake-up word is determined to be a preset wake-up word, it means that the voice information contains a specific wake-up word, that is, multiple voice interaction modules integrated in an electronic device can be awakened by the specific wake-up word.

[0075] As an example, the specific wake-up word may be "hello" or a user-defined name.

[0076] Reference Figure 2 In some embodiments, S103 may include:

[0077] S1031, determining a relevance weight and a coherence weight corresponding to the speech information according to the semantic features;

[0078] S1032, determining a voiceprint weight corresponding to the voice information according to the voiceprint feature;

[0079] S1033, determining the user type of the user based on the relevance weight, the consistency weight, and the voiceprint weight.

[0080] In this embodiment, for the voice interaction module, the user type can be determined by determining the voiceprint weight, relevance weight and coherence weight in the voice information, so as to accurately determine the voice interaction module that needs to be awakened later.

[0081] Specifically, the user type can be users of different age groups, for example, users can be divided into children and adults. For these two types of users, different voiceprints are reflected in the voice information. The relevance and coherence of each sentence in the voice information are different, so the voiceprint weight, relevance weight and coherence weight can be calculated separately. Finally, the user type of the user is comprehensively judged by the voiceprint weight, relevance weight and coherence weight.

[0082] The coherence weight is used to indicate the degree of coherence of the text in the speech information, and the relevance weight is used to indicate the degree of relevance of the text in the speech information.

[0083] In some embodiments, S1031 may specifically include:

[0084] Extract keyword information of semantic features;

[0085] Comparing the keyword information with the first preset text in the preset text library to obtain a comparison result;

[0086] If the comparison result determines that there is a first preset text corresponding to the keyword information, determining an arrangement position of the keyword information according to the first preset text sequence;

[0087] Determine first weight information corresponding to the keyword information according to the arrangement position;

[0088] The correlation weight is determined according to the first weight information and the first preset weight coefficient.

[0089] In this embodiment, when determining the relevance weight, it is first necessary to extract the semantic features to obtain multiple keyword information in the semantic features, and then compare the multiple keyword information with the first preset text in the preset text library to obtain a comparison result, that is, the preset text library pre-stores the first preset text commonly used by the target user. For example, when the target user is a child, the commonly used first preset text may be the name of a cartoon or a children's song. Then, by performing a text comparison between the extracted keyword information and the first preset text, when the same keyword as the first preset text exists in the multiple keyword information, it is also necessary to determine the arrangement position of the keyword according to the first preset text sequence. Since each first preset text in the preset text library has arranged the first preset text in a certain order, after obtaining the first preset text corresponding to the keyword information, the first weight information corresponding to the keyword information can be determined by the arrangement position of the first preset text in the first preset text sequence, and then the relevance weight of the voice information is determined according to the first weight information and the first preset weight coefficient.

[0090] As an example, the certain order may be pre-sorted according to the frequency of occurrence of the words collected and belonging to the target user, that is, the higher the frequency of occurrence, the more likely it is that the first preset text belongs to the target user.

[0091] To make it easier to understand, here are some specific examples;

[0092] For example, the target user is a child, the voice information includes the name of a cartoon, the name of the cartoon exists in the preset text library, and the arrangement position is ranked fifth in the first preset text sequence. At this time, the first weight information can be 0.5, and the first pre-weight coefficient corresponding to the relevance weight is 0.2. At this time, the relevance weight is 0.5*0.2=0.1.

[0093] It is worth noting that the preset text library can be continuously updated, and the first preset text sequence can also be continuously updated.

[0094] In some other embodiments, S1031 may also include:

[0095] Obtain the word interval time, total word count and number of occurrences of the second preset text of the keyword information;

[0096] Determining the occurrence probability of the second preset text according to the total number of words and the number of occurrences of the second preset text;

[0097] determining a first coherence score according to the word interval time and the preset interval time;

[0098] determining a second coherence score according to the occurrence probability and a preset probability threshold;

[0099] A coherence weight is determined based on the first coherence score, the second coherence score, and the second preset weight coefficient.

[0100] In this embodiment, for two user types, children and adults, since children cannot fully express the semantics when inputting voice, there may be pauses or words irrelevant to the semantics. Therefore, the coherence of the voice information can be judged by obtaining the interval time between each word, the total number of words and the number of times the second preset text appears in the keyword information, thereby determining the coherence weight.

[0101] As an example, the second preset text is used to represent words that interrupt the continuity of the voice information, and the second preset text may include at least one of “eh” and “hmm”.

[0102] Specifically, the total number of words in the second preset text can be counted according to the number of occurrences of the second preset text, and then the occurrence frequency of the second preset text can be determined according to the total number of words and the total number of words in the second preset text, and the second coherence score can be determined according to the occurrence frequency and a preset probability threshold.

[0103] In some other embodiments, it is also necessary to determine the first coherence score based on the word interval time and the preset interval time; wherein the preset interval time is the normal speaking speed time of the target user, if the target user is an adult, the preset interval time may be 0.5S, if the target user is a child, the preset interval time may be 1S.

[0104] For example, if the average value of the time interval between each word is 2S and the preset time interval is 1S, it indicates that the speech information is incoherent, and the corresponding first coherence score is 0.3.

[0105] If the occurrence frequency of the second preset text is 0.2 at this time, the second coherence score is determined to be 0.2 at this time.

[0106] When determining the coherence weight, the coherence weight may be determined according to the product of the sum of the first coherence score and the second coherence score and the second preset weight coefficient.

[0107] For example, the first coherence score is 0.3, the second coherence score is 0.2, and the second preset weight coefficient is 0.3. In this case, the coherence weight may be (0.3+0.2)*0.3=0.15.

[0108] In some other embodiments, S1032 may include:

[0109] Input the voiceprint feature into a preset voiceprint model, and use the preset voiceprint model to determine the sound frequency corresponding to the voiceprint feature;

[0110] Determine third weight information of the sound frequency using a preset voiceprint model;

[0111] The voiceprint weight is determined according to the third weight information and the third preset weight coefficient.

[0112] In this embodiment, when determining the voiceprint weight, the extracted voiceprint can be input into a preset voiceprint model, and then the user's voice frequency can be determined by the preset voiceprint model, and the third weight information within can be output, and then the voiceprint weight can be determined by the third weight information and the third preset weight coefficient to determine the user type.

[0113] In some embodiments, in order to more quickly determine the voiceprint weight, before inputting the voiceprint feature into a preset voiceprint model and using the preset voiceprint model to determine the sound frequency corresponding to the voiceprint feature, the method may further include:

[0114] Obtain voiceprint information for multiple preset age groups;

[0115] Extracting voiceprint features from the voiceprint information to obtain voiceprint features for each preset age group;

[0116] A preset voiceprint model is established based on each preset age group and its corresponding voiceprint features.

[0117] In this embodiment, when constructing a preset voiceprint model, voiceprint information of a preset number of preset age groups is first obtained, that is, the voiceprint information of a preset number of target users aged 5-12 and above 12 years old can be obtained respectively, and then the neural network model is used to extract features of the voiceprint information to obtain the voiceprint features of each preset age group, and the neural network model is continuously trained. When the loss function value meets the preset convergence condition, the training of the preset voiceprint model is completed.

[0118] As an example, the preset number may be 100.

[0119] When performing preset voiceprint model training, a sound frequency greater than a preset frequency threshold may be determined as a child, wherein the preset sound frequency threshold may be 1000 Hz.

[0120] In this embodiment, when the voiceprint feature is input into the preset voiceprint model, the preset voiceprint model analyzes the sound frequency of the voiceprint feature to obtain the sound frequency corresponding to the voiceprint feature. When the sound frequency corresponding to the voiceprint feature is greater than 1000HZ, it can be determined that the third weight information corresponding to the voiceprint feature is 0.8. If the third preset weight coefficient is 0.5, the voiceprint weight at this time is 0.8*0.5=0.4.

[0121] It is worth noting that the above-mentioned first preset weight coefficient, second preset weight coefficient and third preset weight coefficient are all determined artificially based on experiments.

[0122] In some embodiments, S1033 may include:

[0123] Determining a total voice weight of the voice information based on the sum of the relevance weight, the coherence weight, and the voiceprint weight;

[0124] When the total voice weight is greater than the preset voice weight, the user type of the user is determined to be a child;

[0125] When the total voice weight is not greater than the preset voice weight, the user type of the user is determined to be an adult.

[0126] In this embodiment, after respectively determining the voiceprint weight, the coherence weight and the correlation weight by the above method, the voiceprint weight, the coherence weight and the correlation weight may be summed to determine the total voice weight of the voice information, that is, 0.1+0.15+0.4=0.65.

[0127] A preset voice weight can be pre-set in the electronic device. When the total voice weight is greater than the preset voice weight, the user type can be determined to be a child. When the total voice weight is not greater than the preset voice weight, the user type can be determined to be an adult.

[0128] As an example, the preset voice weight may be 0.5.

[0129] S104, based on the user type and the preset mapping relationship between the user type and the voice interaction module, wake up the voice interaction module corresponding to the user type.

[0130] In this embodiment, a preset user type and voice interaction module mapping relationship is pre-stored in the electronic device. When the user type is determined by the user's voice information, the voice interaction module corresponding to the user type can be determined by the preset user type and voice interaction module mapping relationship.

[0131] For example, when the user type is determined to be a child, the first voice interaction module corresponding to the child can be awakened, and the first voice interaction module can understand the semantics of the child; when the user type is determined to be an adult, the second voice interaction module corresponding to the adult can be awakened; the second voice interaction module can understand the semantics of the adult.

[0132] In some other embodiments, when a voice interaction module has already been awakened, the user type can be determined by the above method, and the voice interaction module to be awakened can be determined by the user type. If the voice interaction module to be awakened is inconsistent with the currently awakened voice interaction module, the voice interaction module to be awakened is awakened, thereby realizing flexible switching of voice interaction modules and improving user experience.

[0133] In addition, if there is no need to switch the voice interaction module at this time, the current voice interaction module can directly understand the voice information and perform corresponding operations.

[0134] In an embodiment of the present application, in the process of waking up the voice interaction module, the user first inputs voice information, and then performs feature extraction on the voice information to obtain the wake-up word, voiceprint feature and semantic feature of the voice information. Among them, when the wake-up word is a preset wake-up word, it means that the voice interaction module needs to be woken up. It is necessary to determine the user type of the user based on the voiceprint feature and the semantic feature, and then wake up the corresponding voice interaction module based on the user type. That is, in the process of waking up the corresponding voice interaction module, the voice interaction module can be flexibly woken up according to the user's voice information, thereby improving the user experience.

[0135] Reference Figure 3 An embodiment of the present application further provides a wake-up device 300 for a voice interaction module. The wake-up device 300 for a voice interaction module may include:

[0136] The acquisition module 301 is used to acquire the user's voice information in response to the user's voice input;

[0137] A feature extraction module 302 is used to extract features from the voice information to obtain a wake-up word, voiceprint features, and semantic features of the voice information;

[0138] A determination module 303, configured to determine the user type of the user according to the voiceprint features and semantic features of the voice information when the wake-up word is a preset wake-up word;

[0139] The wake-up module 304 is used to wake up the voice interaction module corresponding to the user type based on the user type and the preset mapping relationship between the user type and the voice interaction module.

[0140] As an optional implementation, the determination module 303 may also be used to:

[0141] Determine the relevance weight and coherence weight corresponding to the speech information according to the semantic features, the coherence weight is used to indicate the degree of coherence of the text in the speech information, and the relevance weight is used to indicate the degree of relevance of the text in the speech information;

[0142] Determine the voiceprint weight corresponding to the voice information according to the voiceprint features;

[0143] Based on the relevance weight, the coherence weight, and the voiceprint weight, the user type of the user is determined.

[0144] As an optional implementation, the determination module 303 may also be used to:

[0145] Extract keyword information of semantic features;

[0146] Comparing the keyword information with the first preset text in the preset text library to obtain a comparison result;

[0147] If the comparison result determines that there is a first preset text corresponding to the keyword information, determining an arrangement position of the keyword information according to the first preset text sequence;

[0148] Determine first weight information corresponding to the keyword information according to the arrangement position;

[0149] The correlation weight is determined according to the first weight information and the first preset weight coefficient.

[0150] As an optional implementation, the determination module 303 may also be used to:

[0151] Obtaining the word interval time, the total word count and the number of occurrences of a second preset text of the keyword information, where the second preset text is used to indicate words that interrupt the coherence of the voice information;

[0152] Determining the occurrence probability of the second preset text according to the total number of words and the number of occurrences of the second preset text;

[0153] determining a first coherence score according to the word interval time and the preset interval time;

[0154] determining a second coherence score according to the occurrence probability and a preset probability threshold;

[0155] A coherence weight is determined based on the first coherence score, the second coherence score, and the second preset weight coefficient.

[0156] As an optional implementation, the determination module 303 may also be used to:

[0157] Input the voiceprint feature into a preset voiceprint model, and use the preset voiceprint model to determine the sound frequency corresponding to the voiceprint feature;

[0158] Determine third weight information of the sound frequency using a preset voiceprint model;

[0159] The voiceprint weight is determined according to the third weight information and the third preset weight coefficient.

[0160] As an optional implementation, the determination module 303 may also be used to:

[0161] Obtain voiceprint information for multiple preset age groups;

[0162] Extracting voiceprint features from the voiceprint information to obtain voiceprint features for each preset age group;

[0163] A preset voiceprint model is established based on each preset age group and its corresponding voiceprint features.

[0164] As an optional implementation, the determination module 303 may also be used to:

[0165] Determining a total voice weight of the voice information based on the sum of the relevance weight, the coherence weight, and the voiceprint weight;

[0166] When the total voice weight is greater than the preset voice weight, the user type of the user is determined to be a child;

[0167] When the total voice weight is not greater than the preset voice weight, the user type of the user is determined to be an adult.

[0168] As an optional implementation, the feature extraction module 302 may also be used for:

[0169] Performing signal preprocessing on the voice information to obtain processed voice information, the signal preprocessing comprising at least one of amplification, filtering, denoising and automatic gain control;

[0170] Extract features from speech information to obtain voiceprint features and semantic features of speech information, including:

[0171] The processed speech information is subjected to feature extraction to obtain the voiceprint features and semantic features of the speech information.

[0172] Figure 4 A schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present application is shown.

[0173] The electronic device may include a processor 401 and a memory 402 storing computer program instructions.

[0174] Specifically, the processor 401 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.

[0175] The memory 402 may include a large capacity memory for data or instructions. By way of example and not limitation, the memory 402 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive or a combination of two or more of these. In one example, the memory 402 may include a removable or non-removable (or fixed) medium, or the memory 402 is a non-volatile solid-state memory. The memory 402 may be inside or outside the integrated gateway disaster recovery device.

[0176] In one example, the memory 402 may be a read-only memory (ROM). In one example, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically rewritable ROM (EAROM), or a flash memory, or a combination of two or more of these.

[0177] The memory 402 may include a read-only memory (ROM), a random access memory (RAM), a disk storage medium device, an optical storage medium device, a flash memory device, an electrical, optical or other physical / tangible memory storage device. Therefore, typically, the memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., a memory device) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the wake-up method of the voice interaction module according to the first aspect of the present disclosure.

[0178] The processor 401 reads and executes the computer program instructions stored in the memory 402 to implement Figure 1 A method for waking up a voice interaction module in the illustrated embodiment.

[0179] In one example, the electronic device may further include a communication interface 403 and a bus 404. Figure 4 As shown, the processor 401, the memory 402, and the communication interface 403 are connected via a bus 404 and communicate with each other.

[0180] The communication interface 403 is mainly used to implement communication between various modules, devices, units and / or equipment in the embodiments of the present application.

[0181] Bus 404 includes hardware, software or both, and the parts of electronic equipment are coupled to each other. For example, but not limitation, bus may include accelerated graphics port (Accelerated Graphics Port, AGP) or other graphics bus, enhanced industry standard architecture (Extended Industry Standard Architecture, EISA) bus, front side bus (Front Side Bus, FSB), hypertransmission (Hyper Transport, HT) interconnection, industry standard architecture (IndustryStandard Architecture, ISA) bus, infinite bandwidth interconnection, low pin count (LPC) bus, memory bus, micro channel architecture (MCA) bus, peripheral component interconnection (PCI) bus, PCI-Express (PCI-X) bus, serial advanced technology attachment (SATA) bus, video electronics standard association local (VLB) bus or other suitable bus or two or more of these combinations. In appropriate cases, bus 404 may include one or more buses. Although the present application embodiment describes and shows a specific bus, the application considers any suitable bus or interconnection.

[0182] The electronic device can execute the wake-up method of the voice interaction module in the embodiment of the present application, thereby realizing the combination Figure 1-Figure 3 A method and device for waking up a voice interaction module are described.

[0183] In addition, in combination with the wake-up method of the voice interaction module in the above embodiment, the embodiment of the present application can provide a computer storage medium for implementation. The computer storage medium stores computer program instructions; when the computer program instructions are executed by the processor, any of the wake-up methods of the voice interaction module in the above embodiment is implemented.

[0184] In an optional embodiment, in combination with the wake-up method of the voice interaction module in the above-mentioned embodiment, the embodiment of the present application can provide a computer program product for implementation, and the instructions in the computer program product are executed by the processor of the electronic device, so that the electronic device can implement any one of the wake-up methods of the voice interaction module in the above-mentioned embodiment.

[0185] It should be clear that the present application is not limited to the specific configuration and processing described above and shown in the figures. For the sake of simplicity, a detailed description of the known method is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present application is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between the steps after understanding the spirit of the present application.

[0186] The functional blocks shown in the above block diagram can be implemented as hardware, software, firmware or a combination thereof. When implemented in hardware, it can be, for example, an electronic circuit, an application specific integrated circuit (ASIC), appropriate firmware, a plug-in, a function card, etc. When implemented in software, the elements of the present application are programs or code segments that are used to perform the required tasks. The program or code segment can be stored in a machine-readable medium, or transmitted on a transmission medium or a communication link by a data signal carried in a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, optical fiber media, radio frequency (RF) links, etc. The code segment can be downloaded via a computer network such as the Internet, an intranet, etc.

[0187] It should also be noted that the exemplary embodiments mentioned in this application describe some methods or systems based on a series of steps or devices. However, this application is not limited to the order of the above steps, that is, the steps can be performed in the order mentioned in the embodiment, or in a different order from the embodiment, or several steps can be performed simultaneously.

[0188] Aspects of the present disclosure are described above with reference to the flowchart and / or block diagram of the method, device (system) and computer program product according to the embodiment of the present disclosure. It should be understood that each box in the flowchart and / or block diagram and the combination of each box in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device to produce a machine so that these instructions executed by the processor of the computer or other programmable data processing device enable the implementation of the function / action specified in one or more boxes of the flowchart and / or block diagram. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field programmable logic circuit. It can also be understood that each box in the block diagram and / or flowchart and the combination of boxes in the block diagram and / or flowchart can also be implemented by dedicated hardware that performs a specified function or action, or can be implemented by a combination of dedicated hardware and computer instructions.

[0189] The above are only specific implementation methods of the present application. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, modules and units described above can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here. It should be understood that the protection scope of the present application is not limited to this. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in this application, and these modifications or replacements should be included in the protection scope of this application.

Claims

1. A method for waking up a voice interaction module, characterized in that: include: In response to a user's voice input, obtaining the user's voice information; Extracting features of the voice information to obtain a wake-up word, a voiceprint feature, and a semantic feature of the voice information; When the wake-up word is a preset wake-up word, determining the user type of the user according to the voiceprint features and semantic features of the voice information; Based on the user type and the preset mapping relationship between the user type and the voice interaction module, the voice interaction module corresponding to the user type is awakened.

2. The method according to claim 1, characterized in that The determining the user type of the user according to the voiceprint features and semantic features of the voice information includes: Determining, according to the semantic features, a relevance weight and a coherence weight corresponding to the voice information, wherein the coherence weight is used to indicate the degree of coherence of the text in the voice information, and the relevance weight is used to indicate the degree of relevance of the text in the voice information; Determining a voiceprint weight corresponding to the voice information according to the voiceprint feature; A user type of the user is determined based on the relevance weight, the consistency weight, and the voiceprint weight.

3. The method according to claim 2, characterized in that Determining the relevance weight corresponding to the voice information according to the semantic feature includes: Extracting keyword information of the semantic feature; Comparing the keyword information with a first preset text in a preset text library to obtain a comparison result; In the case where the comparison result determines that there is a first preset text corresponding to the keyword information, determining an arrangement position of the keyword information according to the first preset text sequence; Determining first weight information corresponding to the keyword information according to the arrangement position; A correlation weight is determined according to the first weight information and a first preset weight coefficient.

4. The method according to claim 3, characterized in that: Determining the coherence weight corresponding to the speech information according to the semantic feature includes: Acquire the word interval time, the total word count and the number of occurrences of a second preset text of the keyword information, where the second preset text is used to represent words that interrupt the coherence of the voice information; Determining the occurrence probability of the second preset text according to the total number of words and the number of occurrences of the second preset text; Determining a first coherence score according to the word interval time and the preset interval time; determining a second coherence score according to the occurrence probability and a preset probability threshold; A coherence weight is determined based on the first coherence score, the second coherence score, and a second preset weight coefficient.

5. The method according to any one of claims 2 to 4, characterized in that: The determining, according to the voiceprint feature, a voiceprint weight corresponding to the voice information includes: Inputting the voiceprint feature into a preset voiceprint model, and using the preset voiceprint model to determine the sound frequency corresponding to the voiceprint feature; Determining third weight information of the sound frequency by using the preset voiceprint model; The voiceprint weight is determined according to the third weight information and the third preset weight coefficient.

6. The method according to claim 5, characterized in that Before inputting the voiceprint feature into a preset voiceprint model and using the preset voiceprint model to determine the sound frequency corresponding to the voiceprint feature, the method further includes: Obtain voiceprint information for multiple preset age groups; Extracting voiceprint features from the voiceprint information to obtain voiceprint features for each preset age group; A preset voiceprint model is established based on each preset age group and its corresponding voiceprint features.

7. The method according to claim 2, characterized in that: The determining the user type of the user based on the relevance weight, the coherence weight and the voiceprint weight includes: Determining a total voice weight of the voice information based on a sum of the relevance weight, the coherence weight, and the voiceprint weight; When the total voice weight is greater than the preset voice weight, the user type of the user is determined to be a child; When the total voice weight is not greater than the preset voice weight, the user type of the user is determined to be an adult.

8. The method according to any one of claim 6, characterized in that: Before extracting features from the voice information to obtain voiceprint features and semantic features of the voice information, the method further includes: Performing signal preprocessing on the voice information to obtain processed voice information, wherein the signal preprocessing includes at least one of amplification, filtering, denoising and automatic gain control; The extracting features of the voice information to obtain voiceprint features and semantic features of the voice information includes: Feature extraction is performed on the processed voice information to obtain voiceprint features and semantic features of the voice information.

9. A wake-up device for a voice interaction module, characterized in that: The device comprises: An acquisition module, used for acquiring the user's voice information in response to the user's voice input; A feature extraction module, used to extract features from the voice information to obtain a wake-up word, voiceprint features, and semantic features of the voice information; a determination module, configured to determine the user type of the user according to the voiceprint features and semantic features of the voice information when the wake-up word is a preset wake-up word; A wake-up module is used to wake up the voice interaction module corresponding to the user type based on the user type and a preset mapping relationship between the user type and the voice interaction module.

10. An electronic device, characterized in that: The device comprises: a processor and a memory storing computer program instructions; When the processor executes the computer program instructions, the wake-up method of the voice interaction module as described in any one of claims 1-8 is implemented.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer program instructions, and when the computer program instructions are executed by the processor, the wake-up method of the voice interaction module according to any one of claims 1 to 8 is implemented.

12. A computer program product, characterized in that When the instructions in the computer program product are executed by a processor of an electronic device, the electronic device executes the wake-up method of the voice interaction module as described in any one of claims 1 to 8.

Images