A voiceprint feature updating method and device, electronic equipment and storage medium

Abstract

Embodiments of the present application provide a voiceprint feature updating method and device, electronic equipment and storage medium, relating to the field of biometric identification technology, comprising: obtaining voice data of a user to be verified, extracting voiceprint features of the user to be verified according to the obtained voice data, and obtaining other biological data of the user to be verified, extracting other biological features of the user to be verified according to the obtained other biological data; calculating a first similarity between the extracted voiceprint features and voiceprint features of a registered user, and calculating a second similarity between the extracted other biological features and other biological features of the registered user; in the case that the first similarity is greater than a voiceprint tolerance threshold and the second similarity is greater than an identity representation threshold, updating the voiceprint features of the registered user by using the extracted voiceprint features. The voiceprint feature updating scheme provided by the embodiments of the present application can improve the reliability of identity verification.

Description

Technical Field

[0001] This application relates to the field of biometric technology, and in particular to a method, apparatus, electronic device and storage medium for updating voiceprint features. Background Technology

[0002] In biometrics, user authentication typically relies on user characteristics. For example, facial features can be used for authentication. To ensure reliable authentication, multimodal features are generally employed, such as a combination of voiceprint and facial features.

[0003] In related technologies, the voiceprint and facial features of users are usually obtained and stored during the registration phase. During the verification phase, the voiceprint and facial features of the user to be verified are obtained and matched with the voiceprint and facial features stored during the registration phase. If both the voiceprint and facial features match successfully, the user to be verified is considered to have passed the verification.

[0004] While the above scheme can achieve identity verification using multimodal features, users' vocal organs and vocalization methods may change over time, leading to changes in their voiceprint features. As a result, the voiceprint features extracted during the verification phase may be difficult to match with those extracted during the registration phase, thus reducing the reliability of identity verification. Summary of the Invention

[0005] The purpose of this application is to provide a voiceprint feature updating method, apparatus, electronic device, and storage medium to improve the reliability of identity verification. The specific technical solution is as follows:

[0006] In a first aspect, embodiments of this application provide a voiceprint feature update method, the method comprising:

[0007] The process involves obtaining the voice data of the user to be verified, extracting the voiceprint features of the user based on the obtained voice data, obtaining other biometric data of the user to be verified, and extracting other biometric features of the user based on the obtained other biometric data. The other biometric data refers to biometric data other than voice data that supports identity verification.

[0008] Calculate the first similarity between the extracted voiceprint features and the voiceprint features of a registered user, and calculate the second similarity between the extracted other biometric features and the other biometric features of the registered user;

[0009] If the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold, the voiceprint features of the registered user are updated using the extracted voiceprint features.

[0010] In one embodiment of this application, the voiceprint fault tolerance threshold is less than the voiceprint verification threshold, the identity representation threshold is greater than the other feature verification threshold, and the voiceprint verification threshold and the other feature verification threshold are: thresholds used for identity verification based on voiceprint features and other biometric features;

[0011] The method further includes:

[0012] If the first similarity is greater than the voiceprint verification threshold and the second similarity is greater than the other feature verification thresholds, it is determined that the user to be verified and the registered user are the same user; and / or

[0013] If the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold, the user to be verified and the registered user are determined to be the same user.

[0014] In one embodiment of this application, updating the registered user's voiceprint features using the extracted voiceprint features when the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold includes:

[0015] If the first similarity is greater than the voiceprint fault tolerance threshold and less than the voiceprint verification threshold, and the second similarity is greater than the identity representation threshold, the voiceprint features of the registered user are updated using the extracted voiceprint features.

[0016] In one embodiment of this application, updating the registered user's voiceprint features using the extracted voiceprint features when the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold includes:

[0017] Determine the speech quality of the acquired speech data;

[0018] If the determined voice quality meets the preset voice quality conditions, the first similarity is greater than the voiceprint fault tolerance threshold, and the second similarity is greater than the identity representation threshold, the voiceprint features of the registered user are updated using the extracted voiceprint features.

[0019] In one embodiment of this application, updating the voiceprint features of the registered user using the extracted voiceprint features when the determined voice quality meets preset voice quality conditions, the first similarity is greater than the voiceprint fault tolerance threshold, and the second similarity is greater than the identity representation threshold includes:

[0020] Determine the data quality of other biological data obtained;

[0021] If the determined voice quality meets the preset voice quality conditions, the determined data quality meets the preset data quality conditions, the first similarity is greater than the voiceprint fault tolerance threshold, and the second similarity is greater than the identity representation threshold, the voiceprint features of the registered user are updated using the extracted voiceprint features.

[0022] In one embodiment of this application, determining the speech quality of the obtained speech data includes:

[0023] Based on the obtained voice data, determine the voiceprint environment quality, which reflects the noise level of the user's speaking environment; and / or

[0024] Semantic recognition is performed on the obtained speech data, and the voiceprint content quality reflecting the speech content of the user to be verified is determined based on the recognition results.

[0025] In one embodiment of this application, updating the registered user's voiceprint features using the extracted voiceprint features when the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold includes:

[0026] If the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold, the extracted voiceprint features are stored as target features for updating the voiceprint features of the registered user.

[0027] When the number of stored target features reaches a preset number, the voiceprint features of the registered user are updated using the stored target features.

[0028] In one embodiment of this application, storing the extracted voiceprint features as target features for updating the voiceprint features of the registered user when the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold includes:

[0029] If the first similarity is greater than the voiceprint fault tolerance threshold, the second similarity is greater than the identity representation threshold, and the time interval exceeds a preset duration, the extracted voiceprint features are stored as target features for updating the voiceprint features of the registered user. The time interval is the interval between the last time the target feature was stored and the current time.

[0030] In one embodiment of this application, the voiceprint characteristics of the registered user are obtained in the following manner:

[0031] Obtain the voice data of the registered user and determine the voice quality of the obtained voice data;

[0032] If the determined voice quality meets the preset voice quality conditions, the voiceprint features of the registered user are extracted based on the obtained voice data.

[0033] In one embodiment of this application, other biometric characteristics of the registered user are obtained in the following manner:

[0034] Obtain other biological data of the registered user and determine the data quality of the obtained other biological data;

[0035] If the determined data quality meets the preset data quality conditions, other biological characteristics of the registered user are extracted based on the other biological data obtained.

[0036] Secondly, embodiments of this application provide a voiceprint feature updating device, the device comprising:

[0037] The voiceprint feature acquisition module is used to acquire the voice data of the user to be verified and extract the voiceprint features of the user to be verified based on the acquired voice data.

[0038] The other feature acquisition module is used to acquire other biometric data of the user to be verified, and extract other biometric features of the user to be verified based on the acquired other biometric data, wherein the other biometric data is: biometric data other than voice data that supports identity verification;

[0039] The first similarity calculation module is used to calculate the first similarity between the extracted voiceprint features and the voiceprint features of the registered user, and to calculate the second similarity between the extracted other biometric features and the other biometric features of the registered user.

[0040] The feature update module is used to update the voiceprint features of the registered user using the extracted voiceprint features when the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold.

[0041] In one embodiment of this application, the voiceprint fault tolerance threshold is less than the voiceprint verification threshold, the identity representation threshold is greater than the other feature verification threshold, and the voiceprint verification threshold and the other feature verification threshold are: thresholds used for identity verification based on voiceprint features and other biometric features;

[0042] The device further includes an authentication module for:

[0043] If the first similarity is greater than the voiceprint verification threshold and the second similarity is greater than the other feature verification thresholds, it is determined that the user to be verified and the registered user are the same user; and / or

[0044] If the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold, the user to be verified and the registered user are determined to be the same user.

[0045] In one embodiment of this application, the feature update module is specifically used for:

[0046] If the first similarity is greater than the voiceprint fault tolerance threshold and less than the voiceprint verification threshold, and the second similarity is greater than the identity representation threshold, the voiceprint features of the registered user are updated using the extracted voiceprint features.

[0047] In one embodiment of this application, the feature update module includes:

[0048] The speech quality determination unit is used to determine the speech quality of the acquired speech data;

[0049] The first feature update unit is used to update the voiceprint features of the registered user using the extracted voiceprint features, provided that the determined voice quality meets the preset voice quality conditions, the first similarity is greater than the voiceprint fault tolerance threshold, and the second similarity is greater than the identity representation threshold.

[0050] In one embodiment of this application, the feature update unit is specifically used for:

[0051] Determine the data quality of other biological data obtained;

[0052] If the determined voice quality meets the preset voice quality conditions, the determined data quality meets the preset data quality conditions, the first similarity is greater than the voiceprint fault tolerance threshold, and the second similarity is greater than the identity representation threshold, the voiceprint features of the registered user are updated using the extracted voiceprint features.

[0053] In one embodiment of this application, the voice quality determination unit is specifically used for:

[0054] Based on the obtained voice data, determine the voiceprint environment quality, which reflects the noise level of the user's speaking environment; and / or

[0055] Semantic recognition is performed on the obtained speech data, and the voiceprint content quality reflecting the speech content of the user to be verified is determined based on the recognition results.

[0056] In one embodiment of this application, the feature update module includes:

[0057] The feature storage unit is used to store the extracted voiceprint features as target features for updating the voiceprint features of the registered user when the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold.

[0058] The second feature update unit is used to update the voiceprint features of the registered user using the stored target features when the number of stored target features reaches a preset number.

[0059] In one embodiment of this application, the feature storage unit is specifically used for:

[0060] If the first similarity is greater than the voiceprint fault tolerance threshold, the second similarity is greater than the identity representation threshold, and the time interval exceeds a preset duration, the extracted voiceprint features are stored as target features for updating the voiceprint features of the registered user. The time interval is the interval between the last time the target feature was stored and the current time.

[0061] In one embodiment of this application, the device further includes a first registration module, configured to obtain the voiceprint characteristics of the registered user in the following manner:

[0062] Obtain the voice data of the registered user and determine the voice quality of the obtained voice data;

[0063] If the determined voice quality meets the preset voice quality conditions, the voiceprint features of the registered user are extracted based on the obtained voice data.

[0064] In one embodiment of this application, the device further includes a second registration module for obtaining other biometric characteristics of the registered user in the following manner:

[0065] Obtain other biological data of the registered user and determine the data quality of the obtained other biological data;

[0066] If the determined data quality meets the preset data quality conditions, other biological characteristics of the registered user are extracted based on the other biological data obtained.

[0067] Thirdly, embodiments of this application provide an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0068] Memory, used to store computer programs;

[0069] When a processor executes a program stored in memory, it implements any of the steps described in the first aspect.

[0070] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of any of the methods described in the first aspect.

[0071] This application also provides a computer program product containing instructions that, when run on a computer, cause the computer to execute any of the voiceprint feature update methods described above.

[0072] Beneficial effects of the embodiments in this application:

[0073] In the voiceprint feature update scheme provided in this application embodiment, voice data of the user to be verified can be obtained, voiceprint features of the user to be verified can be extracted based on the obtained voice data, and other biometric data of the user to be verified can be obtained. Other biometric features of the user to be verified can be extracted based on the obtained other biometric data, wherein the other biometric data is: biometric data other than voice data that supports identity verification; a first similarity between the extracted voiceprint features and the voiceprint features of a registered user is calculated, and a second similarity between the extracted other biometric features and the other biometric features of the registered user is calculated; if the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold, the voiceprint features of the registered user are updated using the extracted voiceprint features. In this scenario, if the similarity between the other biometric features of the user to be verified and the other biometric features of the registered user is greater than the identity representation threshold, and the similarity between the voiceprint features of the user to be verified and the voiceprint features of the registered user is greater than the voiceprint fault tolerance threshold, it is considered that the user to be verified and the registered user may be the same user. However, because the registered user's voice has changed, the voiceprint features obtained during the current identity verification differ from those obtained during registration. Therefore, the currently obtained voiceprint features can be used to update the registered user's voiceprint features, ensuring that the registered user's voiceprint features match the current voice, facilitating subsequent identity verification based on the updated voiceprint features. Thus, it is evident that applying the voiceprint feature updating scheme provided in this application can improve the reliability of identity verification. Attached Figure Description

[0074] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other embodiments can be obtained based on these drawings.

[0075] Figure 1 A flowchart illustrating a voiceprint feature updating method provided in an embodiment of this application;

[0076] Figure 2 A flowchart illustrating another voiceprint feature updating method provided in an embodiment of this application;

[0077] Figure 3A schematic diagram illustrating a voiceprint feature update process provided in an embodiment of this application;

[0078] Figure 4 This application provides a schematic diagram illustrating a process for identity verification using voiceprint features.

[0079] Figure 5 This is a schematic diagram of the structure of a voiceprint feature updating device provided in an embodiment of this application;

[0080] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0081] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art based on this application are within the scope of protection of this application.

[0082] In scenarios employing multimodal features for identity verification, voiceprint features are often used as one modality due to the convenience and security of voice interaction. However, as people age, subtle changes occur in human organs and vocalization patterns, leading to time-varying characteristics in voiceprint features. Therefore, in practical applications, voiceprint-based identity verification may experience performance degradation over time, resulting in lower reliability of such schemes.

[0083] To improve the reliability of identity verification, this application provides a voiceprint feature updating method, apparatus, electronic device, and storage medium, which will be described in detail below.

[0084] In one embodiment of this application, a voiceprint feature update method is provided. This method can be applied to electronic devices such as mobile phones, access control devices, computers, and servers, and is suitable for scenarios that require authentication using multimodal features. The method includes:

[0085] Obtain the voice data of the user to be verified, extract the voiceprint features of the user to be verified based on the obtained voice data, and obtain other biometric data of the user to be verified. Extract other biometric features of the user to be verified based on the obtained other biometric data. The other biometric data refers to biometric data other than voice data that supports identity verification.

[0086] Calculate the first similarity between the extracted voiceprint features and the voiceprint features of registered users, and calculate the second similarity between the extracted other biometric features and the other biometric features of registered users;

[0087] If the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold, the voiceprint features of the registered user are updated using the extracted voiceprint features.

[0088] In this scenario, if the similarity between the other biometric features of the user to be verified and the other biometric features of the registered user is greater than the identity representation threshold, and the similarity between the voiceprint features of the user to be verified and the voiceprint features of the registered user is greater than the voiceprint fault tolerance threshold, it is considered that the user to be verified and the registered user may be the same user. However, because the registered user's voice has changed, the voiceprint features obtained during the current identity verification differ from those obtained during registration. Therefore, the currently obtained voiceprint features can be used to update the registered user's voiceprint features, ensuring that the registered user's voiceprint features match the current voice, facilitating subsequent identity verification based on the updated voiceprint features. Thus, it is evident that applying the voiceprint feature update scheme provided in the above embodiments can improve the reliability of identity verification.

[0089] The above-mentioned voiceprint feature update method will be described in detail below.

[0090] See Figure 1 , Figure 1 This is a flowchart illustrating a voiceprint feature update method provided in an embodiment of this application. The method includes the following steps S101-S103:

[0091] S101, obtain the voice data of the user to be verified, extract the voiceprint features of the user to be verified based on the obtained voice data, and obtain other biological data of the user to be verified, and extract other biological features of the user to be verified based on the obtained other biological data.

[0092] Among them, the users to be verified are those who need to be authenticated.

[0093] The above-mentioned voice data is collected voice data. For example, it can be obtained by collecting voice segments spoken randomly by the user to be verified, or by collecting voice segments of the user to be verified reading preset text. The preset text can be "Xiaoyi Xiaoyi", "Hey, Siri", "74629854", etc.

[0094] The aforementioned voiceprint features can include fundamental frequency, formant, and other characteristics.

[0095] The aforementioned other biometric data refers to biometric data that supports identity verification, excluding voice data. This can be data in other modalities besides voice, such as a user's facial image data, fingerprint data, iris data, etc.

[0096] The other biometric features mentioned above are: biometric features other than voiceprint features that support identity verification, such as facial features, fingerprint features, iris features, etc.

[0097] Specifically, the system can obtain the voice data of the user to be verified, extract features from the voice data to obtain the voiceprint features of the user to be verified, and obtain other biometric data of the user to be verified, extract features from the other biometric data to obtain other biometric features of the user to be verified.

[0098] In one embodiment of this application, the voice segments spoken by the user to be verified can be directly collected to obtain the aforementioned voice data; alternatively, the voice data can be obtained by receiving voice segments spoken by the user to be verified collected by a voice acquisition device, such as a microphone, pickup, or mobile phone.

[0099] Correspondingly, it can directly collect other biological data of the user to be verified; it can also receive other biological data collected by other biological data collection devices, such as cameras, fingerprint collectors, etc.

[0100] In one embodiment of this application, a voiceprint feature extraction model can be used to extract features from the obtained speech data to obtain voiceprint features; alternatively, the obtained speech data can be converted into audio spectrum data, and the audio spectrum data can be analyzed to obtain voiceprint features.

[0101] Correspondingly, other feature extraction models can be used to extract features from the obtained biological data to obtain other biological features; alternatively, preset algorithms can be used to extract features from the obtained biological data to obtain other biological features. These algorithms can be edge extraction algorithms, image segmentation algorithms, etc.

[0102] S102, calculate the first similarity between the extracted voiceprint features and the voiceprint features of the registered users, and calculate the second similarity between the extracted other biometric features and the other biometric features of the registered users.

[0103] The aforementioned registered users are those who have pre-registered.

[0104] Specifically, during the registration phase, voice data and other biometric data of the registered user can be obtained. Based on this data, the registered user's voiceprint characteristics and other biometric features are derived and stored for subsequent identity verification. In addition, the registered user's identity identifier can be obtained to facilitate the subsequent identification of verified users. This identity identifier can be the user's name, ID number, employee ID, etc.

[0105] During the verification phase, the similarity between the voiceprint features of the user to be verified and the voiceprint features of the registered users stored can be calculated as the first similarity. The higher the first similarity, the greater the possibility that the user to be verified and the registered user are the same user.

[0106] It also calculates the similarity between other biometric features of the user to be verified and other biometric features of the registered users stored in the database, and uses this as a second similarity score. The higher the second similarity score, the greater the likelihood that the user to be verified and the registered user are the same user.

[0107] In one embodiment of this application, when calculating the similarity between features, Euclidean distance, cosine similarity, Manhattan distance, etc., can be calculated as the similarity between features.

[0108] S103, if the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold, update the voiceprint features of the registered user using the extracted voiceprint features.

[0109] Among them, the aforementioned voiceprint error tolerance threshold and identity representation threshold are both preset thresholds.

[0110] Specifically, if the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold, it indicates that the user to be verified and the registered user may be the same user. Therefore, the voiceprint features of the currently obtained user to be verified can be used to update the stored voiceprint features of the registered user to ensure that the updated voiceprint features match the voiceprint features of the current registered user better, which will facilitate subsequent identity verification using the updated voiceprint features.

[0111] In one embodiment of this application, when updating the voiceprint features of a registered user using the extracted voiceprint features, the voiceprint features of the registered user can be directly updated to the extracted voiceprint features, or a weighted average between the extracted voiceprint features and the stored voiceprint features can be calculated as the updated voiceprint features, etc.

[0112] Specifically, the updated voiceprint feature e can be calculated using the following formula. new :

[0113] e new=α*e old +β*e update

[0114] Where α represents the first weight, and e represents the second weight. old This represents the voiceprint features of currently stored registered users, where β represents the second weight, and e update The extracted voiceprint features are represented by the sum of the first weight and the second weight, which can be 1.

[0115] In this scenario, if the similarity between the other biometric features of the user to be verified and the other biometric features of the registered user is greater than the identity representation threshold, and the similarity between the voiceprint features of the user to be verified and the voiceprint features of the registered user is greater than the voiceprint fault tolerance threshold, it is considered that the user to be verified and the registered user may be the same user. However, because the registered user's voice has changed, the voiceprint features obtained during the current identity verification differ from the voiceprint features obtained during registration. Therefore, the voiceprint features of the registered user can be updated using the currently obtained voiceprint features to ensure that the registered user's voiceprint features match the current voice, facilitating subsequent identity verification based on the updated voiceprint features. Thus, it can be seen that applying the voiceprint feature update scheme provided in the above embodiments can improve the reliability of identity verification.

[0116] In one embodiment of this application, the voiceprint fault tolerance threshold is less than the voiceprint verification threshold, and the identity representation threshold is greater than the other feature verification threshold. The voiceprint verification threshold and the other feature verification threshold are thresholds used for identity verification based on voiceprint features and other biometric features.

[0117] Based on the above scheme, there are several possible scenarios during the identity verification phase, which will be described below.

[0118] Scenario 1: If the first similarity is greater than the voiceprint verification threshold and the second similarity is greater than the other feature verification thresholds, the user to be verified and the registered user are determined to be the same user.

[0119] Specifically, if the second similarity between the other biometric features of the user to be verified and the other biometric features of the registered user is greater than the other feature verification threshold, and the similarity between the voiceprint features of the user to be verified and the voiceprint features of the registered user is greater than the voiceprint verification threshold, it is considered that the voiceprint features of the user to be verified and the registered user are similar, and other biometric features are similar, so it can be directly determined that the user to be verified and the registered user may be the same user.

[0120] In one embodiment of this application, under the above circumstances, the extracted voiceprint features can be used to update the voiceprint features of the registered user, thereby achieving adaptive updating of the voiceprint features.

[0121] Alternatively, the voiceprint features of registered users can be left unupdated. Since the extracted voiceprint features are quite similar to the currently stored voiceprint features, it indicates that the voice of the registered user has not changed significantly. Therefore, there is no need to update the voiceprint features of the registered user, thus saving computing resources.

[0122] Scenario 2: If the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold, the user to be verified and the registered user are determined to be the same user.

[0123] Specifically, if the second similarity between the other biometric features of the user to be verified and the other biometric features of the registered user is greater than the identity representation threshold, and the first similarity between the voiceprint features of the user to be verified and the voiceprint features of the registered user is greater than the voiceprint fault tolerance threshold, then since the other biometric features of the user to be verified and the registered user are highly similar, and the voiceprint features of the user to be verified and the registered user are relatively similar, regardless of whether the first similarity is higher than the voiceprint verification threshold, it can be considered that the user to be verified and the registered user may be the same user.

[0124] In one embodiment of this application, when the first similarity is greater than the voiceprint fault tolerance threshold and less than the voiceprint verification threshold, and the second similarity is greater than the identity representation threshold, the voiceprint features of the registered user are updated using the extracted voiceprint features.

[0125] Specifically, under the condition that the following formula is satisfied, the extracted voiceprint features can be used to update the voiceprint features of the registered user:

[0126] threshold voice_low <score voice <threshold voice_normal

[0127] threshold face_normal <threshold face_high <score face

[0128] Among them, the above threshold voice_low The score represents the voiceprint tolerance threshold. voice The threshold represents the first similarity between the voiceprint features of the user to be verified and the voiceprint features of the registered user. voice_normal Indicates the voiceprint verification threshold;

[0129] threshold face_normal The threshold represents the validation threshold for other features. face_high The score represents the threshold for identity representation. faceThis represents the second similarity between the other biometric characteristics of the user to be verified and the other biometric characteristics of the registered user.

[0130] In the above scheme, if the second similarity is greater than the identity representation threshold and the first similarity is greater than the voiceprint fault tolerance threshold, the user to be verified and the registered user can be considered to be the same user. Since the first similarity is less than the voiceprint verification threshold, it indicates that the registered user's voice has changed significantly. Therefore, it is necessary to use the extracted voiceprint features to update the registered user's voiceprint features to achieve adaptive updating of voiceprint features, which facilitates subsequent identity verification based on the updated voiceprint features.

[0131] In one embodiment of this application, when updating the voiceprint features in step S103 above, the following can be done:

[0132] If the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold, the extracted voiceprint features are stored as target features for updating the voiceprint features of registered users; if the number of stored target features reaches a preset number, the stored target features are used to update the voiceprint features of registered users.

[0133] The preset quantity can be 5, 10, 20, etc.

[0134] Specifically, for the current user to be verified, when the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold, the voiceprint features of the user to be verified can be stored as target features for later use. When the number of stored target features reaches the preset number, it can be determined that the registered user's voice has indeed changed significantly. Therefore, the stored target features can be used to update the user's voiceprint features. This can prevent the user's voice from changing accidentally and causing the registered user's voiceprint features to be updated, thus improving the reliability of voiceprint feature updates.

[0135] In one embodiment of this application, a weighted average value between a preset number of target features and the voiceprint features of a registered user can be calculated as the updated voiceprint features;

[0136] Alternatively, the arithmetic mean of the above-mentioned preset number of target features can be calculated to obtain the average feature. Then, the weighted average between the average feature and the voiceprint feature of the registered user can be calculated as the updated voiceprint feature.

[0137] Alternatively, the above-mentioned average features can be directly used as the updated voiceprint features, and this application does not limit this.

[0138] In one embodiment of this application, if the first similarity is greater than the voiceprint fault tolerance threshold, the second similarity is greater than the identity representation threshold, and the time interval exceeds a preset duration, the extracted voiceprint features are stored as target features for updating the voiceprint features of registered users.

[0139] The time interval is the interval between the last time the target feature was stored and the current time.

[0140] The preset duration can be 5 hours, 1 day, 7 days, etc., and this application embodiment does not limit it.

[0141] Specifically, the voiceprint features of the user to be verified can be obtained once at least every preset time interval, where the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold. The obtained voiceprint features are then used to update the voiceprint features of the registered user. This avoids frequent updates to the stored voiceprint features of the registered user and prevents attacks on the voiceprint features used for identity verification.

[0142] In one embodiment of this application, the voice quality of the obtained voice data can be determined. If the determined voice quality meets the preset voice quality conditions, the first similarity is greater than the voiceprint fault tolerance threshold, and the second similarity is greater than the identity representation threshold, the voiceprint features of the registered user are updated using the extracted voiceprint features.

[0143] The aforementioned speech quality can reflect the quality of speech data, and further reflect the quality of voiceprint features obtained based on speech data.

[0144] Specifically, if the voice quality meets the preset voice quality conditions, it means that the quality of the extracted voiceprint features is high. Therefore, if the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold, the extracted voiceprint features can be used to update the voiceprint features of the registered user.

[0145] If the voice quality does not meet the requirements, it indicates that the quality of the extracted voiceprint features is low, so the voiceprint features of registered users do not need to be updated.

[0146] In one embodiment of this application, when determining the voice quality of the obtained voice data, the voiceprint environment quality, which reflects the noise level of the speaking environment of the user to be verified, can be determined based on the obtained voice data.

[0147] Specifically, the speech separation degree and signal-to-noise ratio of the speech data can be calculated to obtain the voiceprint environment quality. This quality reflects the noise level of the surrounding environment when the user being verified speaks. The higher the voiceprint environment quality, the quieter the environment, and the higher the quality of the voiceprint features obtained based on the speech data; the lower the voiceprint environment quality, the noisier the environment, and the lower the quality of the voiceprint features obtained based on the speech data.

[0148] In addition, semantic recognition can be performed on the obtained voice data, and the voiceprint content quality reflecting the content spoken by the user to be verified can be determined based on the recognition results.

[0149] Specifically, semantic recognition can be performed on the aforementioned speech data to obtain recognition results. These results reflect the content of the user's speech and determine the quality of the voiceprint content based on the recognition results. When the content of the user's speech has actual meaning, such as "the weather is nice," the quality of the voiceprint content is high, and the quality of the voiceprint features obtained based on the speech data is also high. Conversely, when the content of the user's speech does not have actual meaning, such as "ah ah ah ah ah," the quality of the voiceprint content is low, and the quality of the voiceprint features obtained based on the speech data is also low.

[0150] In one embodiment of this application, the data quality of other obtained biological data can also be determined; if the determined voice quality meets the preset voice quality conditions, the determined data quality meets the preset data quality conditions, the first similarity is greater than the voiceprint fault tolerance threshold, and the second similarity is greater than the identity representation threshold, the voiceprint features of the registered user are updated using the extracted voiceprint features.

[0151] The aforementioned data quality can reflect the quality of other biological data obtained, and thus the quality of other biological characteristics obtained based on other biological data.

[0152] For example, assuming the other biological data mentioned above is a face image, the data quality can reflect the image quality. When the signal-to-noise ratio of the face image is high and the distortion is small, the data quality of the face image is high, and the quality of the facial features obtained based on the face image is high.

[0153] Specifically, if the determined voice quality meets the preset voice quality conditions and the determined data quality meets the preset data quality conditions, it indicates that the extracted voiceprint features and other biometric features are of high quality. Therefore, if the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold, the extracted voiceprint features can be used to update the voiceprint features of the registered user.

[0154] If the voice quality or data quality does not meet the requirements, it indicates that the quality of the extracted voiceprint features or other biometric features is low. Therefore, the voiceprint features of registered users do not need to be updated.

[0155] See Figure 2 , Figure 2 This is a flowchart illustrating another voiceprint feature updating method provided in an embodiment of this application. The method includes the following steps S201-S206:

[0156] S201, obtain the voice data of the user to be verified, extract the voiceprint features of the user to be verified based on the obtained voice data, and obtain other biological data of the user to be verified, and extract other biological features of the user to be verified based on the obtained other biological data.

[0157] S202, calculate the first similarity between the extracted voiceprint features and the voiceprint features of the registered users, and calculate the second similarity between the extracted other biometric features and the other biometric features of the registered users.

[0158] S203, if the first similarity is greater than the voiceprint verification threshold and the second similarity is greater than the other feature verification thresholds, determine that the user to be verified and the registered user are the same user, and consider the user to be verified to have passed the verification; otherwise, proceed to step S204.

[0159] S204, if the first similarity is not greater than the voiceprint verification threshold or the second similarity is not greater than other feature verification thresholds, determine the speech quality of the obtained speech data and the data quality of the other obtained biological data.

[0160] S205, if the determined voice quality meets the preset voice quality conditions and the determined data quality meets the preset data quality conditions, if the first similarity is greater than the voiceprint fault tolerance threshold and less than the voiceprint verification threshold, the second similarity is greater than the identity representation threshold, and the time interval exceeds the preset duration, the extracted voiceprint features are stored as target features for updating the voiceprint features of registered users.

[0161] The time interval is the interval between the last time the target feature was stored and the current time.

[0162] S206, when the number of stored target features reaches a preset number, update the voiceprint features of the registered user using the stored target features.

[0163] In one embodiment of this application, the voiceprint characteristics of a registered user can be obtained in the following way:

[0164] Obtain the voice data of registered users and determine the voice quality of the obtained voice data; if the determined voice quality meets the preset voice quality conditions, extract the voiceprint features of registered users based on the obtained voice data.

[0165] Specifically, voice segments of registered users can be obtained to acquire their voice data. The voice quality of the voice data can then be determined. If the voice quality meets the requirements, it indicates that the environment in which the registered user is speaking is relatively quiet, and the quality of the voiceprint features obtained from the voice data is high. Therefore, the voiceprint features of the registered user can be extracted from the voice data.

[0166] Correspondingly, other biometric characteristics of registered users can be obtained through the following methods:

[0167] Obtain other biological data of registered users and determine the data quality of the obtained other biological data; if the determined data quality meets the preset data quality conditions, extract other biological characteristics of registered users based on the obtained other biological data.

[0168] Specifically, other biological data of registered users can be obtained, and the data quality of the aforementioned other biological data can be determined. If the data quality meets the data quality conditions, it indicates that the quality of other biological characteristics obtained based on the other biological data is high. Therefore, other biological characteristics of registered users can be extracted based on the other biological data.

[0169] See Figure 3 , Figure 3 This is a schematic diagram illustrating a voiceprint feature update process provided in an embodiment of this application. Figure 3 As shown, the voice data and face image of the user to be verified can be collected first. The voice data is then used to extract features to obtain the voiceprint features of the user to be verified, and the face image is then used to extract features to obtain the face features of the user to be verified.

[0170] The voiceprint and facial features of the user to be verified are compared with those of the registered user. Specifically, the first similarity between the voiceprint features of the user to be verified and the voiceprint features of the registered user can be calculated, and the second similarity between the facial features of the user to be verified and the facial features of the registered user can be calculated.

[0171] Based on the first similarity and the second similarity mentioned above, the user to be verified is identified. If the first similarity is greater than the voiceprint verification threshold and the second similarity is greater than the face feature verification threshold, it is determined that the user to be verified and the registered user are the same user, and the user to be verified is considered to have passed the verification.

[0172] Based on the above voice data and face image analysis, the current environmental quality can be determined. Specifically, the voice quality of the obtained voice data and the image quality of the obtained face image can be determined.

[0173] Using the aforementioned environmental quality, first similarity, and second similarity, it is determined whether the voiceprint feature can be used for updating. If the determined voice quality meets the preset voice quality conditions and the determined image quality meets the preset image quality conditions, and if the first similarity is greater than the voiceprint fault tolerance threshold and less than the voiceprint verification threshold, the second similarity is greater than the identity representation threshold, and the time interval exceeds the preset duration, the extracted voiceprint feature is stored as the target feature for updating the voiceprint feature of the registered user.

[0174] The time interval is the interval between the last time the target feature was stored and the current time.

[0175] When the number of stored target features reaches a preset number, the voiceprint features of the registered user are updated using the stored target features.

[0176] To describe this solution more clearly, the following section introduces voiceprints.

[0177] Voiceprint recognition technology, also known as speaker identification technology, is a technique that identifies a speaker through their voice. Intuitively, while voiceprints aren't as readily apparent as the individual differences in faces or fingerprints, each person's vocal organs, such as the vocal tract, mouth, and nasal cavity, exhibit unique characteristics, resulting in variations in their voice. The most direct example is when we call home; a simple "Hello?" accurately identifies whether it's our parents or siblings answering the phone. This uniqueness of the speaker's identity information carried in their voice makes voiceprints a powerful force in biometrics, similar to faces and fingerprints, assisting or even replacing traditional digital passwords, playing a crucial role in security and personal information encryption.

[0178] See Figure 4 , Figure 4 This is a schematic diagram of a process for identity verification using voiceprint features, which can be divided into two parts: voiceprint registration and voiceprint verification.

[0179] In the voiceprint registration section, the speech data of the target speaker can be collected, and VAD (Voice Activity Detection) can be performed on this speech data to remove silence parts, enhance the speech data, detect the quality of the speech data, and achieve effective speech extraction. The extracted effective speech is then input into the voiceprint feature extraction model, and the extraction algorithm is used to extract the voiceprint features. The voiceprint features are then saved to the speech database as the voiceprint features of the target speaker.

[0180] In the voiceprint verification stage, the speaker's voiceprint features can be extracted through the same process as in the voiceprint registration stage. The voiceprint features are then compared with the voiceprint features in the speech database using a comparison algorithm. Based on the comparison results, identity verification is performed, and thus the identity verification result is obtained.

[0181] In the voiceprint feature update scheme provided in the above embodiments, voice data of the user to be verified can be obtained, voiceprint features can be extracted based on the obtained voice data, and other biometric data of the user to be verified can be obtained. Other biometric features can be extracted based on the obtained other biometric data, wherein the other biometric data are: biometric data other than voice data that supports identity verification; a first similarity between the extracted voiceprint features and the voiceprint features of a registered user can be calculated, and a second similarity between the extracted other biometric features and the other biometric features of the registered user can be calculated; if the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold, the voiceprint features of the registered user can be updated using the extracted voiceprint features. In this scenario, if the similarity between the other biometric features of the user to be verified and the other biometric features of the registered user is greater than the identity representation threshold, and the similarity between the voiceprint features of the user to be verified and the voiceprint features of the registered user is greater than the voiceprint fault tolerance threshold, it is considered that the user to be verified and the registered user may be the same user. However, because the registered user's voice has changed, the voiceprint features obtained during the current identity verification differ from those obtained during registration. Therefore, the currently obtained voiceprint features can be used to update the registered user's voiceprint features, ensuring that the registered user's voiceprint features match the current voice, facilitating subsequent identity verification based on the updated voiceprint features. Thus, it is evident that applying the voiceprint feature update scheme provided in the above embodiments can improve the reliability of identity verification.

[0182] Corresponding to the above-described voiceprint feature updating method, this application also provides a voiceprint feature updating device, which will be described in detail below.

[0183] See Figure 5 , Figure 5 This is a schematic diagram of a voiceprint feature updating device provided in an embodiment of this application. The device includes:

[0184] The voiceprint feature acquisition module 501 is used to acquire the voice data of the user to be verified and extract the voiceprint features of the user to be verified based on the acquired voice data.

[0185] Other feature acquisition module 502 is used to acquire other biometric data of the user to be verified, and extract other biometric features of the user to be verified based on the acquired other biometric data, wherein the other biometric data is: biometric data other than voice data that supports identity verification;

[0186] The similarity calculation module 503 is used to calculate the first similarity between the extracted voiceprint features and the voiceprint features of the registered user, and to calculate the second similarity between the extracted other biometric features and the other biometric features of the registered user.

[0187] The feature update module 504 is used to update the voiceprint features of the registered user using the extracted voiceprint features when the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold.

[0188] In one embodiment of this application, the voiceprint fault tolerance threshold is less than the voiceprint verification threshold, the identity representation threshold is greater than the other feature verification threshold, and the voiceprint verification threshold and the other feature verification threshold are: thresholds used for identity verification based on voiceprint features and other biometric features;

[0189] The device further includes an authentication module for:

[0190] If the first similarity is greater than the voiceprint verification threshold and the second similarity is greater than the other feature verification thresholds, it is determined that the user to be verified and the registered user are the same user; and / or

[0191] If the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold, the user to be verified and the registered user are determined to be the same user.

[0192] In one embodiment of this application, the feature update module 504 is specifically used for:

[0193] If the first similarity is greater than the voiceprint fault tolerance threshold and less than the voiceprint verification threshold, and the second similarity is greater than the identity representation threshold, the voiceprint features of the registered user are updated using the extracted voiceprint features.

[0194] In one embodiment of this application, the feature update module 504 includes:

[0195] The speech quality determination unit is used to determine the speech quality of the acquired speech data;

[0196] The first feature update unit is used to update the voiceprint features of the registered user using the extracted voiceprint features, provided that the determined voice quality meets the preset voice quality conditions, the first similarity is greater than the voiceprint fault tolerance threshold, and the second similarity is greater than the identity representation threshold.

[0197] In one embodiment of this application, the feature update unit is specifically used for:

[0198] Determine the data quality of other biological data obtained;

[0199] If the determined voice quality meets the preset voice quality conditions, the determined data quality meets the preset data quality conditions, the first similarity is greater than the voiceprint fault tolerance threshold, and the second similarity is greater than the identity representation threshold, the voiceprint features of the registered user are updated using the extracted voiceprint features.

[0200] In one embodiment of this application, the voice quality determination unit is specifically used for:

[0201] Based on the obtained voice data, determine the voiceprint environment quality, which reflects the noise level of the user's speaking environment; and / or

[0202] Semantic recognition is performed on the obtained speech data, and the voiceprint content quality reflecting the speech content of the user to be verified is determined based on the recognition results.

[0203] In one embodiment of this application, the feature update module 504 includes:

[0204] The feature storage unit is used to store the extracted voiceprint features as target features for updating the voiceprint features of the registered user when the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold.

[0205] The second feature update unit is used to update the voiceprint features of the registered user using the stored target features when the number of stored target features reaches a preset number.

[0206] In one embodiment of this application, the feature storage unit is specifically used for:

[0207] If the first similarity is greater than the voiceprint fault tolerance threshold, the second similarity is greater than the identity representation threshold, and the time interval exceeds a preset duration, the extracted voiceprint features are stored as target features for updating the voiceprint features of the registered user. The time interval is the interval between the last time the target feature was stored and the current time.

[0208] In one embodiment of this application, the device further includes a first registration module, configured to obtain the voiceprint characteristics of the registered user in the following manner:

[0209] Obtain the voice data of the registered user and determine the voice quality of the obtained voice data;

[0210] If the determined voice quality meets the preset voice quality conditions, the voiceprint features of the registered user are extracted based on the obtained voice data.

[0211] In one embodiment of this application, the device further includes a second registration module for obtaining other biometric characteristics of the registered user in the following manner:

[0212] Obtain other biological data of the registered user and determine the data quality of the obtained other biological data;

[0213] If the determined data quality meets the preset data quality conditions, other biological characteristics of the registered user are extracted based on the other biological data obtained.

[0214] In the voiceprint feature update scheme provided in the above embodiments, voice data of the user to be verified can be obtained, voiceprint features can be extracted based on the obtained voice data, and other biometric data of the user to be verified can be obtained. Other biometric features can be extracted based on the obtained other biometric data, wherein the other biometric data are: biometric data other than voice data that supports identity verification; a first similarity between the extracted voiceprint features and the voiceprint features of a registered user can be calculated, and a second similarity between the extracted other biometric features and the other biometric features of the registered user can be calculated; if the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold, the voiceprint features of the registered user can be updated using the extracted voiceprint features. In this scenario, if the similarity between the other biometric features of the user to be verified and the other biometric features of the registered user is greater than the identity representation threshold, and the similarity between the voiceprint features of the user to be verified and the voiceprint features of the registered user is greater than the voiceprint fault tolerance threshold, it is considered that the user to be verified and the registered user may be the same user. However, because the registered user's voice has changed, the voiceprint features obtained during the current identity verification differ from those obtained during registration. Therefore, the currently obtained voiceprint features can be used to update the registered user's voiceprint features, ensuring that the registered user's voiceprint features match the current voice, facilitating subsequent identity verification based on the updated voiceprint features. Thus, it is evident that applying the voiceprint feature update scheme provided in the above embodiments can improve the reliability of identity verification.

[0215] This application also provides an electronic device, such as... Figure 6As shown, it includes a processor 601, a communication interface 602, a memory 603, and a communication bus 604, wherein the processor 601, the communication interface 602, and the memory 603 communicate with each other through the communication bus 604.

[0216] Memory 603 is used to store computer programs;

[0217] The processor 601 is used to implement a voiceprint feature update method when executing a program stored in the memory 603.

[0218] The communication bus mentioned in the above electronic devices can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the diagram, but this does not mean that there is only one bus or one type of bus.

[0219] The communication interface is used for communication between the aforementioned electronic devices and other devices.

[0220] The memory may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0221] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0222] In another embodiment provided in this application, a computer-readable storage medium is also provided, which stores a computer program that, when executed by a processor, implements the steps of any of the above-described voiceprint feature update methods.

[0223] In another embodiment provided in this application, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute any of the voiceprint feature update methods described above.

[0224] In the voiceprint feature update scheme provided in the above embodiments, voice data of the user to be verified can be obtained, voiceprint features can be extracted based on the obtained voice data, and other biometric data of the user to be verified can be obtained. Other biometric features can be extracted based on the obtained other biometric data, wherein the other biometric data are: biometric data other than voice data that supports identity verification; a first similarity between the extracted voiceprint features and the voiceprint features of a registered user can be calculated, and a second similarity between the extracted other biometric features and the other biometric features of the registered user can be calculated; if the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold, the voiceprint features of the registered user can be updated using the extracted voiceprint features. In this scenario, if the similarity between the other biometric features of the user to be verified and the other biometric features of the registered user is greater than the identity representation threshold, and the similarity between the voiceprint features of the user to be verified and the voiceprint features of the registered user is greater than the voiceprint fault tolerance threshold, it is considered that the user to be verified and the registered user may be the same user. However, because the registered user's voice has changed, the voiceprint features obtained during the current identity verification differ from those obtained during registration. Therefore, the currently obtained voiceprint features can be used to update the registered user's voiceprint features, ensuring that the registered user's voiceprint features match the current voice, facilitating subsequent identity verification based on the updated voiceprint features. Thus, it is evident that applying the voiceprint feature update scheme provided in the above embodiments can improve the reliability of identity verification.

[0225] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid state disk (SSD)).

[0226] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, 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. Without further limitations, 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 said element.

[0227] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the device embodiments, electronic device embodiments, computer-readable storage medium embodiments, and computer program product embodiments are basically similar to the method embodiments, so the descriptions are relatively simple, and relevant parts can be referred to the descriptions of the method embodiments.

[0228] The above description is merely a preferred embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application are included within the scope of protection of this application.

Claims

1. A method for updating voiceprint features, characterized in that, The method includes: The process involves obtaining the voice data of the user to be verified, extracting the voiceprint features of the user based on the obtained voice data, obtaining other biometric data of the user to be verified, and extracting other biometric features of the user based on the obtained other biometric data. The other biometric data refers to biometric data other than voice data that supports identity verification. Calculate the first similarity between the extracted voiceprint features and the voiceprint features of a registered user, and calculate the second similarity between the extracted other biometric features and the other biometric features of the registered user; If the first similarity is greater than the voiceprint fault tolerance threshold and less than the voiceprint verification threshold, and the second similarity is greater than the identity representation threshold, the voiceprint features of the registered user are updated using the extracted voiceprint features. The voiceprint fault tolerance threshold is less than the voiceprint verification threshold, and the identity representation threshold is greater than other feature verification thresholds. The voiceprint verification threshold and the other feature verification thresholds are thresholds used for identity verification based on voiceprint features and other biometric features. If the first similarity is greater than the voiceprint verification threshold and the second similarity is greater than the other feature verification thresholds, it is determined that the user to be verified and the registered user are the same user; and / or If the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold, the user to be verified and the registered user are determined to be the same user.

2. The method according to claim 1, characterized in that, The step of updating the registered user's voiceprint features using the extracted voiceprint features when the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold includes: Determine the speech quality of the acquired speech data; If the determined voice quality meets the preset voice quality conditions, the first similarity is greater than the voiceprint fault tolerance threshold, and the second similarity is greater than the identity representation threshold, the voiceprint features of the registered user are updated using the extracted voiceprint features.

3. The method according to claim 2, characterized in that, The step of updating the registered user's voiceprint features using the extracted voiceprint features, under the condition that the determined voice quality meets the preset voice quality conditions, the first similarity is greater than the voiceprint fault tolerance threshold, and the second similarity is greater than the identity representation threshold, includes: Determine the data quality of other biological data obtained; If the determined voice quality meets the preset voice quality conditions, the determined data quality meets the preset data quality conditions, the first similarity is greater than the voiceprint fault tolerance threshold, and the second similarity is greater than the identity representation threshold, the voiceprint features of the registered user are updated using the extracted voiceprint features.

4. The method according to claim 2, characterized in that, Determining the speech quality of the obtained speech data includes: Based on the obtained voice data, determine the voiceprint environment quality, which reflects the noise level of the user's speaking environment; and / or Semantic recognition is performed on the obtained speech data, and the voiceprint content quality reflecting the speech content of the user to be verified is determined based on the recognition results.

5. The method according to claim 1, characterized in that, The step of updating the registered user's voiceprint features using the extracted voiceprint features when the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold includes: If the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold, the extracted voiceprint features are stored as target features for updating the voiceprint features of the registered user. When the number of stored target features reaches a preset number, the voiceprint features of the registered user are updated using the stored target features.

6. The method according to claim 5, characterized in that, The step of storing the extracted voiceprint features as target features for updating the voiceprint features of the registered user when the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold includes: If the first similarity is greater than the voiceprint fault tolerance threshold, the second similarity is greater than the identity representation threshold, and the time interval exceeds a preset duration, the extracted voiceprint features are stored as target features for updating the voiceprint features of the registered user. The time interval is the interval between the last time the target feature was stored and the current time.

7. A voiceprint feature updating device, characterized in that, The device includes: The voiceprint feature acquisition module is used to acquire the voice data of the user to be verified and extract the voiceprint features of the user to be verified based on the acquired voice data. The other feature acquisition module is used to acquire other biometric data of the user to be verified, and extract other biometric features of the user to be verified based on the acquired other biometric data, wherein the other biometric data is: biometric data other than voice data that supports identity verification; The first similarity calculation module is used to calculate the first similarity between the extracted voiceprint features and the voiceprint features of the registered user, and to calculate the second similarity between the extracted other biometric features and the other biometric features of the registered user. The feature update module is used to update the voiceprint features of the registered user using the extracted voiceprint features when the first similarity is greater than the voiceprint fault tolerance threshold and less than the voiceprint verification threshold, and the second similarity is greater than the identity representation threshold. The voiceprint fault tolerance threshold is less than the voiceprint verification threshold, and the identity representation threshold is greater than other feature verification thresholds. The voiceprint verification threshold and the other feature verification thresholds are thresholds used for identity verification based on voiceprint features and other biometric features. The device further includes an identity verification module, configured to determine that the user to be verified and the registered user are the same user when the first similarity is greater than the voiceprint verification threshold and the second similarity is greater than the other feature verification threshold; and / or to determine that the user to be verified and the registered user are the same user when the first similarity is greater than the voiceprint fault tolerance threshold and the second similarity is greater than the identity representation threshold.

8. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, when executing a program stored in memory, implements the steps of the method described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the method described in any one of claims 1-6.

Images