Intelligent voice household appliance based on voiceprint detection technology

Abstract

The invention provides an intelligent voice household appliance based on a voiceprint detection technology, and the household appliance comprises a voice receiving module which is configured to obtain target voice data of at least one user; the voiceprint judgment module is configured to perform voiceprint judgment according to the target voice data to obtain target user binding data; and the music playing module is configured to play the audio data according to the target user binding data. According to the scheme, the problem that in the prior art, 'private customization 'of intelligent voice home appliances is tedious in use is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of smart home appliances, and in particular to a smart voice home appliance based on voiceprint detection technology. Background Art

[0002] The development of Internet communication technology has made people's lives more convenient. In recent years, the smart home industry has developed rapidly, and smart home life is spreading to thousands of households. At present, many smart voice devices have Internet access after being equipped with the network, which can realize online dialogue between people and machines. For example, when the user wakes up the device and asks to play a song, the voice device will automatically search for a song from the cloud music library to play, but this song may not necessarily be the user's favorite. Therefore, this "clumsy" human-computer interaction needs to be further optimized.

[0003] With the in-depth research of voiceprint recognition technology, this technology is expected to be promoted to smart home appliances. Functions such as "recognizing people by voice" and "personal customization" will give home appliances a more intelligent and customized experience. Existing technologies use the Internet of Things technology to establish a communication connection between smart lighting devices and smart speakers, and associate lighting effect tags and situation type tags with specific voice control commands to build an associated voice command database, and then perform semantic matching analysis on the user voice control signal received by the smart speaker and each voice control command in the associated voice command database, so as to achieve intelligent adjustment of lighting effects based on the matching results.

[0004] However, although such interactive technology can realize the "personal customization" of smart voice appliances to a certain extent, due to the setting of a large number of different types of tags, users need to associate all the specified tags of different types each time they use it in order to achieve more accurate music playback, which makes the use of smart furniture cumbersome. Summary of the invention

[0005] The present application provides a smart voice appliance based on voiceprint detection technology to solve the problem of cumbersome use of "personal customization" of smart voice appliances in the prior art.

[0006] The intelligent voice home appliance comprises:

[0007] A voice receiving module, wherein the voice receiving module is configured to obtain target voice data of at least one user;

[0008] A voiceprint determination module, wherein the voiceprint determination module is configured to perform voiceprint determination according to the target voice data to obtain target user binding data, wherein the target user binding data corresponds to a user;

[0009] A music playing module, wherein the music playing module is configured to play audio data according to the target user binding data;

[0010] The voiceprint determination module comprises:

[0011] A feature extraction unit, wherein the feature extraction unit is configured to extract features according to the target speech data to obtain feature segments;

[0012] A feature matching unit, wherein the feature matching unit is configured to:

[0013] Calculate the feature overlap between each feature segment and a user-preset feature segment corresponding to each feature segment to obtain a feature similarity corresponding to each feature segment;

[0014] When the number of the feature segments whose feature similarity is greater than the second preset threshold is greater than the third preset threshold, the target user binding data corresponding to the target voice data is extracted.

[0015] Preferably, the voice receiving module includes:

[0016] a sound receiving unit, wherein the sound receiving unit is configured to obtain voice data of at least one user;

[0017] A signal conversion unit is configured to convert all the voice data from sound signals into electrical signals to obtain the target voice data.

[0018] Preferably, the voice receiving module further includes:

[0019] A wake-up judgment unit, wherein the wake-up judgment unit is configured to:

[0020] Calculating the voice overlap between the target voice data and the preset arousing voice data to obtain the voice similarity;

[0021] When the voice similarity is greater than a first preset threshold, the target voice data is sent to the voiceprint determination module.

[0022] Preferably, the voiceprint determination module further includes:

[0023] A user identification generating unit, the user identification generating unit being configured to extract the target user binding data and generate user identification information according to the target user binding data, the user identification information including a first user identification;

[0024] A user identification confirmation unit, wherein the user identification confirmation unit is configured to send the target user binding data corresponding to the first user identification to the music playing module.

[0025] Preferably, the user identification generating unit is further configured to generate user identification information according to all the target user binding data when at least two target user binding data are extracted, and the user identification information further includes a second user identification;

[0026] The user identification confirmation unit is further configured to:

[0027] Playing confirmation prompt tones of the first user identification and the second user identification according to the user identification information;

[0028] When the user's first confirmation information is received, the target user binding data corresponding to the first user identifier is sent to the music playing module; when the user's second confirmation information is received, the target user binding data corresponding to the second user identifier is sent to the music playing module.

[0029] Preferably, the intelligent voice appliance further includes:

[0030] A user binding module, wherein the user binding module is configured to:

[0031] The target user binding data of the corresponding user is established according to the user's personalized settings, and the target user binding data at least includes user identification information, user voiceprint data, and singer information.

[0032] Preferably, the feature matching unit is further configured as follows:

[0033] When the feature segment is received, extracting the user voiceprint data;

[0034] The user voiceprint data is feature segmented according to the feature segment to obtain the user preset feature segment.

[0035] Preferably, the music playing module is further configured as follows:

[0036] Extracting the singer information and the music information according to the target user binding data;

[0037] Randomly select target singer information from all the singer information, and randomly select corresponding target music information according to the target singer information, wherein the target music information corresponds to a piece of music;

[0038] Play the music corresponding to the target music information.

[0039] Preferably, the smart voice appliance also includes a communication module, a cloud platform and a mobile terminal, the communication module is a 4G-CAT1 smart module, the communication module is communicatively connected to the cloud platform, the cloud platform is communicatively connected to the mobile terminal; the mobile terminal synchronizes data with the cloud platform.

[0040] As can be seen from the above content, the present application provides an intelligent voice home appliance based on voiceprint detection technology, the intelligent voice home appliance includes a voice receiving module, the voice receiving module is configured to obtain the target voice data of at least one user; a voiceprint judgment module, the voiceprint judgment module is configured to perform voiceprint judgment according to the target voice data, and obtain the target user binding data, the target user binding data corresponds to a user; a music playing module, the music playing module is configured to play audio data according to the target user binding data; the voiceprint judgment module includes: a feature extraction unit, the feature extraction unit is configured to perform feature extraction according to the target voice data, and obtain a feature segment; a feature matching unit, the feature matching unit is configured to: calculate the feature overlap according to each feature segment and the user preset feature segment corresponding to each feature segment, and obtain the feature similarity corresponding to each feature segment; when the number of the feature segments whose feature similarity is greater than the second preset threshold is greater than the third preset threshold, extract the target user binding data corresponding to the target voice data. The present application solves the problem of cumbersome use of "private customization" of intelligent voice home appliances in the prior art through the above scheme. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solution of the present application, the drawings required for use in the embodiments are briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0042] Figure 1 A schematic diagram of a smart voice home appliance based on voiceprint detection technology for this application;

[0043] Figure 2 This is a schematic diagram of a voice receiving module in an intelligent voice appliance based on voiceprint detection technology in this application;

[0044] Figure 3 This is a schematic diagram of a voiceprint determination module in an intelligent voice appliance based on voiceprint detection technology in the present application. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] Figure 1A schematic diagram of an intelligent voice appliance based on voiceprint detection technology is shown in this application.

[0047] See also Figure 1 It can be seen that this embodiment provides a smart voice home appliance based on voiceprint detection technology, and the smart voice home appliance includes:

[0048] The voice receiving module 100 is configured to obtain target voice data of at least one user. Specifically, in this embodiment, the voice receiving module 100 is used to obtain the target voice data of the user.

[0049] Exemplarily, for example, when the user makes a sound of "turn on the music", the voice receiving module 100 acquires the sound of "turn on the music" and converts it into audio data.

[0050] Figure 2 This is a schematic diagram of a voice receiving module in an intelligent voice appliance based on voiceprint detection technology in the present application.

[0051] See also Figure 2 It can be seen that, further, in some embodiments, the voice receiving module 100 includes:

[0052] The sound receiving unit 110 is configured to obtain voice data of at least one user. Specifically, in this embodiment, the sound receiving unit 110 mainly plays the role of obtaining user audio.

[0053] Exemplarily, for example, the sound receiving unit 110 may be a sound receiving device, and the sound emitted by the user is acquired through the sound receiving unit 110 .

[0054] The voice receiving module 100 also includes:

[0055] The signal conversion unit 120 is configured to convert all the voice data from sound signals into electrical signals to obtain a plurality of the target voice data. Specifically, in this embodiment, since the pure audio data cannot be processed related to voiceprint detection, the signal conversion unit 120 is used to convert all the voice data from sound signals into electrical signals to provide data for subsequent voiceprint detection.

[0056] The voice receiving module 100 also includes:

[0057] The wake-up judgment unit 130 is configured to:

[0058] Calculating the voice overlap between the target voice data and the preset arousing voice data to obtain the voice similarity;

[0059] When the voice similarity is greater than a first preset threshold, the target voice data is sent to the voiceprint determination module 200 .

[0060] Specifically, in this embodiment, the wake-up determination unit 130 is further provided, and the wake-up determination unit 130 is used to determine whether there is a user's wake-up instruction.

[0061] Exemplarily, for example, when the user makes a sound, the invocation determination unit 130 determines whether it is the sound of "turning on music", and thereby determines whether to initiate subsequent voiceprint detection.

[0062] It should be noted that when a user makes a sound, there will generally be some noise, so the sound needs to be denoised. This embodiment does not impose too many restrictions on the specific technology of the denoising process.

[0063] The intelligent voice home appliance also includes:

[0064] The voiceprint judgment module 200 is configured to perform voiceprint judgment according to the target voice data to obtain target user binding data, wherein the target user binding data corresponds to one user. Specifically, in this embodiment, considering that multiple people may speak at the same time, it is necessary to distinguish the human voice. Therefore, the voiceprint judgment module 200 is used to judge whether the human voice is the voice of the corresponding user, thereby ensuring that the smart voice appliance will not be activated by any sound other than the user's voice.

[0065] For example, when the sound of "turn on the music" appears in a film or TV work, a conventional voice appliance will start directly. However, in this embodiment, before turning on the intelligent voice appliance, the voiceprint judgment module 200 needs to identify the human voice to ensure that the intelligent voice appliance will not start without the user's demand.

[0066] Figure 3 This is a schematic diagram of a voiceprint determination module in an intelligent voice appliance based on voiceprint detection technology in the present application.

[0067] See also Figure 3 It can be seen that, further, in some embodiments, the voiceprint determination module 200 includes:

[0068] A feature extraction unit 210, wherein the feature extraction unit 210 is configured to perform feature extraction according to the target speech data to obtain a plurality of feature segments;

[0069] The feature matching unit 220 is configured to:

[0070] Calculate the feature overlap between each feature segment and a user-preset feature segment corresponding to each feature segment to obtain a feature similarity corresponding to each feature segment;

[0071] When the number of the feature segments whose feature similarity is greater than the second preset threshold is greater than the third preset threshold, the target user binding data corresponding to the target voice data is extracted.

[0072] Specifically, in this embodiment, a number of feature segments are extracted from the target voice data by the feature extraction unit 210, and the feature segment is matched with the user preset feature segment set by the user by the feature matching unit 220 to calculate the feature overlap, and when the requirements are met, the target user binding data is extracted and sent to the subsequent modules.

[0073] Exemplarily, for example, when there is a human voice, the features of the human voice are extracted by the feature extraction unit 210, wherein the features of the human voice may include timbre features; the feature matching unit 220 extracts a user-preset feature segment from the data set by the user, wherein the user-preset feature segment is a number of sound segments of the sound data entered in advance by the user; the feature matching unit 220 calculates the feature overlap between each corresponding feature segment and the user-preset feature segment, and when the similarity between the two is greater than the second preset threshold, the feature segment is marked as a "segment that meets the requirements", and the proportion of the "segment that meets the requirements" in all feature segments is counted, and when the proportion is greater than the third preset threshold, the human voice is identified as the sound of "turning on the music" made by the user.

[0074] Furthermore, in some embodiments, the voiceprint determination module 200 further includes:

[0075] A user identification generating unit 230, wherein the user identification generating unit 230 is configured to extract the target user binding data and generate user identification information according to the target user binding data, wherein the user identification information includes a first user identification;

[0076] The user identification confirmation unit 240 is configured to send the target user binding data corresponding to the first user identification to the music playing module 300.

[0077] Specifically, in this embodiment, the user identification generation unit 230 extracts the user identification information of the corresponding customer from the target user binding data set by the user, and each user identification information corresponds to only one user; the target user binding data corresponding to the first user identification is sent to the music playback module 300 through the user identification confirmation unit 240.

[0078] Wherein, when the user identification generating unit 230 is further configured to generate user identification information according to all the target user binding data when at least two target user binding data are extracted, the user identification information further includes a second user identification;

[0079] The user identification confirmation unit 240 is further configured to:

[0080] Playing confirmation prompt tones of the first user identification and the second user identification according to the user identification information;

[0081] When the user's first confirmation information is received, the target user binding data corresponding to the first user identifier is sent to the music playing module 300; when the user's second confirmation information is received, the target user binding data corresponding to the second user identifier is sent to the music playing module 300.

[0082] Exemplarily, for example, when there are two or more human voices at the same time and all the voices have been pre-bound in the smart voice appliance, that is, when the first user identifier and the second user identifier exist at the same time, it is necessary to select which user to play music; the user identifier of each user is generated by the user identifier generation unit 230, and all user identifiers are broadcasted by the user identifier confirmation unit 240 to remind the user to select the corresponding user identifier to play music for the corresponding user; when the user sends the user identifier "Zhang San", the smart voice appliance plays music according to the user "Zhang San", and the "Zhang San" sent by the user is regarded as user identifier confirmation information in this embodiment.

[0083] It should be noted that for the user identification confirmation information sent by the user, there is no need to identify the user's identity. It can be understood that even a human voice that is not bound to the smart voice appliance can send a user identification confirmation information to select the corresponding user for music playback.

[0084] The intelligent voice home appliance also includes:

[0085] The music playing module 300 is configured to play audio data according to the target user binding data. Specifically, in this embodiment, the corresponding audio data of the corresponding user is played through the music playing module 300.

[0086] Furthermore, in some embodiments, the intelligent voice appliance further includes:

[0087] The user binding module 400 is configured to:

[0088] The target user binding data of the corresponding user is established according to the user's personalized settings, and the target user binding data includes at least user identification information, user voiceprint data, and singer information. Specifically, in this embodiment, the user's relevant information is bound through the user binding module 400, wherein the user binds the user information through voice input, text input, and database input, and the target user binding data is generated thereby.

[0089] Furthermore, in some embodiments, the feature matching unit 220 is further configured to:

[0090] When the feature segment is received, extracting the user voiceprint data;

[0091] The user voiceprint data is feature segmented according to the feature segments to obtain a plurality of user preset feature segments. Specifically, in this embodiment, the user preset feature segments are extracted from the target user binding data by the feature matching unit 220 .

[0092] Furthermore, in some embodiments, the music playing module 300 is further configured to:

[0093] Extracting the singer information and the music information according to the target user binding data;

[0094] Randomly select target singer information from all the singer information, and randomly select corresponding target music information according to the target singer information, wherein the target music information corresponds to a piece of music;

[0095] Play the music corresponding to the target music information.

[0096] Specifically, in this embodiment, the music of the corresponding singer set by the corresponding user is played through the music playing module 300, wherein the sound of the music can also be "customized" in this embodiment, and the adjustment method can be determined according to the user's usage habits. There are no excessive restrictions on the specific adjustment method in this embodiment.

[0097] Furthermore, in some embodiments, the smart voice appliance also includes a communication module 500, a cloud platform 600 and a mobile terminal 700, the communication module 500 is a 4G-CAT1 smart module, the communication module 500 is communicatively connected to the cloud platform 600, the cloud platform 600 is communicatively connected to the mobile terminal 700; the mobile terminal 700 synchronizes data with the cloud platform 600.

[0098] Specifically, in this embodiment, the conventional WIFI module is replaced by a 4G-CAT1 smart module (the 4G-CAT1 smart module is a cellular Internet of Things module that uses the currently available 4G network. The module automatically connects to the network after being powered on, and does not require network configuration settings like a WiFi smart module). When the user's home has no network environment or the network environment is poor, the smart voice device can still support online voice recognition and online music playback without network configuration; the user's music preferences can be updated at any time through the cloud platform 600, the communication module 500 and the mobile terminal 700, wherein the user's music preferences can be adjusted by importing music data from the user's mobile terminal 700.

[0099] This embodiment has the following advantages:

[0100] After the smart voice home appliance completes the user's network binding, it registers the user's voiceprint and further sets the singer name, volume, default operation mode and other information matched under the voiceprint on the mobile terminal. In this way, when the user wakes up the smart device with voice and asks to play a song, the system will automatically identify the person based on the voiceprint matching, and further play the music according to the singer name and volume set by the user or operate the device operation mode customized by the user in advance according to the user's setting information. The present invention is conducive to improving the intelligence and accuracy of the smart voice device in playing the songs to be played, and is conducive to improving the matching degree between the smart device playing songs, default state operation and the target user. It has a good application prospect in the future.

Claims

1. An intelligent voice home appliance based on voiceprint detection technology, characterized in that: The intelligent voice home appliance comprises: A voice receiving module (100), the voice receiving module (100) being configured to obtain target voice data of at least one user; A voiceprint determination module (200), the voiceprint determination module (200) being configured to perform voiceprint determination according to the target voice data to obtain target user binding data, the target user binding data corresponding to a user; A music playing module (300), the music playing module (300) being configured to play audio data according to the target user binding data; The voiceprint determination module (200) comprises: A feature extraction unit (210), wherein the feature extraction unit (210) is configured to perform feature extraction based on the target speech data to obtain a feature segment; A feature matching unit (220), wherein the feature matching unit (220) is configured to: Calculate the feature overlap between each feature segment and a user-preset feature segment corresponding to each feature segment to obtain a feature similarity corresponding to each feature segment; When the number of the feature segments whose feature similarity is greater than the second preset threshold is greater than the third preset threshold, the target user binding data corresponding to the target voice data is extracted.

2. According to claim 1, the intelligent voice appliance based on voiceprint detection technology is characterized in that: The voice receiving module (100) comprises: A sound receiving unit (110), wherein the sound receiving unit (110) is configured to obtain voice data of at least one user; A signal conversion unit (120), wherein the signal conversion unit (120) is configured to convert all the voice data from sound signals into electrical signals to obtain a plurality of target voice data.

3. The intelligent voice appliance based on voiceprint detection technology according to claim 2, characterized in that: The voice receiving module (100) further comprises: A wake-up judgment unit (130), wherein the wake-up judgment unit (130) is configured to: Calculating the voice overlap between the target voice data and the preset arousing voice data to obtain the voice similarity; When the voice similarity is greater than a first preset threshold, the target voice data is sent to the voiceprint determination module (200).

4. The intelligent voice appliance based on voiceprint detection technology according to claim 1, characterized in that: The voiceprint determination module (200) further comprises: A user identification generating unit (230), the user identification generating unit (230) being configured to extract the target user binding data and generate user identification information according to the target user binding data, the user identification information comprising a first user identification; A user identification confirmation unit (240), wherein the user identification confirmation unit (240) is configured to send the target user binding data corresponding to the first user identification to the music playing module (300).

5. The intelligent voice home appliance based on voiceprint detection technology according to claim 4, characterized in that: The user identification generating unit (230) is further configured to generate user identification information according to all the target user binding data when at least two target user binding data are extracted, wherein the user identification information further includes a second user identification; A user identification confirmation unit (240), wherein the user identification confirmation unit (240) is further configured to: Playing confirmation prompt tones of the first user identification and the second user identification according to the user identification information; When the user's first confirmation information is received, the target user binding data corresponding to the first user identifier is sent to the music playing module (300); when the user's second confirmation information is received, the target user binding data corresponding to the second user identifier is sent to the music playing module (300).

6. The intelligent voice appliance based on voiceprint detection technology according to claim 1, characterized in that: The intelligent voice home appliance also includes: A user binding module (400), wherein the user binding module (400) is configured to: The target user binding data of the corresponding user is established according to the user's personalized settings, and the target user binding data at least includes user identification information, user voiceprint data, and singer information.

7. The intelligent voice appliance based on voiceprint detection technology according to claim 6, characterized in that: The feature matching unit (220) is further configured to: When the feature segment is received, extracting the user voiceprint data; The user voiceprint data is feature segmented according to the feature segment to obtain the user preset feature segment.

8. The intelligent voice home appliance based on voiceprint detection technology according to claim 6, characterized in that: The music playing module (300) is further configured to: Extracting the singer information and the music information according to the target user binding data; Randomly select target singer information from all the singer information, and randomly select corresponding target music information according to the target singer information, wherein the target music information corresponds to a piece of music; Play the music corresponding to the target music information.

9. The intelligent voice home appliance based on voiceprint detection technology according to claim 6, characterized in that: The intelligent voice home appliance further comprises a communication module (500), a cloud platform (600) and a mobile terminal (700); the communication module (500) is a 4G-CAT1 intelligent module; the communication module (500) is communicatively connected to the cloud platform (600); the cloud platform (600) is communicatively connected to the mobile terminal (700); and the mobile terminal (700) and the cloud platform (600) are data synchronized.

Images