CN112599137A - Method and device for verifying voiceprint model recognition effect and computer equipment - Google Patents
Method and device for verifying voiceprint model recognition effect and computer equipment Download PDFInfo
- Publication number
- CN112599137A CN112599137A CN202011485999.5A CN202011485999A CN112599137A CN 112599137 A CN112599137 A CN 112599137A CN 202011485999 A CN202011485999 A CN 202011485999A CN 112599137 A CN112599137 A CN 112599137A
- Authority
- CN
- China
- Prior art keywords
- verification
- success rate
- test
- audio
- recognition
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000000694 effects Effects 0.000 title claims abstract description 82
- 238000000034 method Methods 0.000 title claims abstract description 42
- 238000012795 verification Methods 0.000 claims abstract description 305
- 238000012360 testing method Methods 0.000 claims abstract description 206
- 238000004590 computer program Methods 0.000 claims description 18
- 230000005236 sound signal Effects 0.000 claims description 15
- 238000011156 evaluation Methods 0.000 claims description 11
- 238000010200 validation analysis Methods 0.000 claims description 3
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 6
- 230000002159 abnormal effect Effects 0.000 description 5
- 238000010183 spectrum analysis Methods 0.000 description 3
- 238000001228 spectrum Methods 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L17/00—Speaker identification or verification techniques
- G10L17/04—Training, enrolment or model building
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/48—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
- G10L25/51—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Computational Linguistics (AREA)
- Signal Processing (AREA)
- Telephonic Communication Services (AREA)
Abstract
The application relates to artificial intelligence, and particularly discloses a method, a device, computer equipment and a storage medium for verifying voiceprint model identification effects. On one hand, the data source used for verifying the voiceprint model can reflect the crowd distribution of a business scene and can truly reflect the use condition of an actual business scene, and on the other hand, login verification is carried out at least twice during login testing to meet the use will of a user, so that the verification result of the voiceprint model recognition effect obtained by the method can truly reflect the use effect of the actual business scene, the model effect deviation caused by non-model factors is eliminated, and the verification accuracy is improved.
Description
Technical Field
The present application relates to the field of artificial intelligence technologies, and in particular, to a method, an apparatus, a computer device, and a storage medium for verifying a voiceprint model recognition effect.
Background
The voiceprint has the advantages of convenient acquisition, low identification cost and the like, so that the application of voiceprint identification is wider and wider. Voiceprint recognition is the conversion of acoustic signals into electrical signals, which are then recognized by a computer. Specifically, a voiceprint model is trained by using a neural network, and voice is recognized. The recognition effect of the voiceprint model has a decisive influence on the application of the voiceprint.
After training, the voiceprint model is typically validated to assess whether it can be applied online. In the traditional scheme, a pre-collected audio file is used, 3 audios in each group are used as voiceprint registration audios, 1 audio is used as verification audio to verify the false alarm rate of voiceprints (the audios in the same group are recorded by the same person, the user is judged to be the user when the voiceprint is identified, the false alarm rate is counted if the user is judged not to be the user), other audios not in the group are used to verify the false alarm rate (the audios in the different groups are recorded by different persons, the user is judged to be the user when the voiceprint is identified, the false alarm rate is counted if the user is judged to be the user), and the accuracy and the reliability of a voiceprint model are verified based on the two indexes.
When the method is used for verification, only the influence of the model on the voiceprint verification accuracy is considered, and other influence factors exist in practical application, for example, the crowd distribution in different application scenes is different, so that the verification method does not consider the influence of a non-model on the voiceprint recognition effect, and the verification result is inaccurate.
Disclosure of Invention
In view of the above, it is necessary to provide a method, an apparatus, a computer device and a storage medium for verifying the recognition effect of a voiceprint model, which can improve the verification accuracy.
A method of verifying the recognition effect of a voiceprint model, the method comprising:
acquiring a verification data set, wherein the population distribution of test users in the verification data set is related to a service scene, the verification data set comprises registration audios and verification audios of the test users, and each test user at least comprises two groups of verification audios;
calling a voiceprint model, and respectively carrying out at least two times of voiceprint verification on each test user according to the registration audio and the verification audio of each test user in the verification data set to obtain the success rate of test identification;
and obtaining a verification result of the voiceprint model recognition effect according to the test recognition success rate.
In one embodiment, the invoking the voiceprint model, and performing at least two voiceprint verifications on each test user according to the registration audio and the verification audio of each test user in the verification data set to obtain the test identification success rate includes:
calling a voiceprint model, and respectively carrying out at least two times of voiceprint verification on each test user according to the registration audio frequency and the verification audio frequency of each test user in the verification data set to obtain a primary recognition success rate, a secondary recognition success rate and a recognition success rate;
carrying out weighted summation on the primary identification success rate, the secondary identification success rate and the identification success rate to obtain a test identification success rate; wherein the weight of the primary recognition success rate is greater than the weight of the secondary recognition success rate and the weight of the recognition success rate.
In one embodiment, the obtaining the verification data set includes:
acquiring the crowd distribution of a service scene;
determining the number of test users of each gender and age stage according to the crowd distribution to obtain target test users;
and acquiring at least three groups of registration audios and at least two groups of verification audios of each target test user to obtain a verification data set.
In one embodiment, the method further comprises:
according to the number of successfully identified users in the users with normal audio frequency, the success rate of service identification under the condition of normal audio frequency is obtained;
the obtaining of the verification result of the voiceprint model recognition effect according to the test recognition success rate includes: and obtaining a verification result of the voiceprint model recognition effect according to the test recognition success rate and the service recognition success rate.
In one embodiment, verifying the number of successfully identified users among users with normal audio according to an actual service scenario to obtain a service identification success rate under a normal audio condition includes:
acquiring verification audio data of a user in actual service, analyzing the verification audio data, and acquiring the user with normal verification audio;
acquiring the number of successfully identified users in the users with normal verification audio in the actual service;
and calculating the success rate of service identification under the normal audio frequency condition according to the number of the users with normal audio frequency verification and the number of the users successfully identified in the users with normal audio frequency verification.
In one embodiment, acquiring verification audio data of a user in an actual service, analyzing the verification audio data, and acquiring a user with normal verification audio, includes:
detecting and verifying the signal-to-noise ratio of the audio data to obtain high-quality sound data;
performing frequency-spectrum-normal analysis and voice recognition on the high-quality sound signal to obtain an effective sound signal;
and eliminating the voice signals of the logged and registered non-identical person in the effective voice signals to obtain the user with normal verified audio.
In one embodiment, obtaining a verification result of the voiceprint model recognition effect according to the test recognition success rate and the service recognition success rate includes:
determining the weight of the test recognition success rate and the service recognition success rate according to the online time of the application program and the number of users with normal verification audio;
weighting the test recognition success rate and the service recognition success rate to obtain a score of a voiceprint model;
and obtaining a verification result of the recognition effect of the voiceprint model according to the score.
An apparatus for verifying the recognition effect of a voiceprint model, the apparatus comprising:
the verification data acquisition module is used for acquiring a verification data set, the population distribution of test users in the verification data set is related to a service scene, the verification data set comprises registration audios and verification audios of the test users, and each test user at least has two groups of verification audios;
the identification analysis module is used for calling a voiceprint model, and performing voiceprint verification on each test user at least twice according to the registration audio frequency and the verification audio frequency of each test user in the verification data set to obtain the success rate of test identification;
and the evaluation module is used for obtaining a verification result of the voiceprint model recognition effect according to the test recognition success rate.
A computer device comprising a memory and a processor, the memory storing a computer program, the processor implementing the following steps when executing the computer program:
acquiring a verification data set, wherein the population distribution of test users in the verification data set is related to a service scene, the verification data set comprises registration audios and verification audios of the test users, and each test user at least comprises two groups of verification audios;
calling a voiceprint model, and respectively carrying out at least two times of voiceprint verification on each test user according to the registration audio and the verification audio of each test user in the verification data set to obtain the success rate of test identification;
and obtaining a verification result of the voiceprint model recognition effect according to the test recognition success rate.
A computer storage medium having a computer program stored thereon, the computer program when executed by a processor implementing the steps of:
acquiring a verification data set, wherein the population distribution of test users in the verification data set is related to a service scene, the verification data set comprises registration audios and verification audios of the test users, and each test user at least comprises two groups of verification audios;
calling a voiceprint model, and respectively carrying out at least two times of voiceprint verification on each test user according to the registration audio and the verification audio of each test user in the verification data set to obtain the success rate of test identification;
and obtaining a verification result of the voiceprint model recognition effect according to the test recognition success rate.
According to the method, the device, the computer equipment and the storage medium for verifying the voiceprint model recognition effect, the crowd distribution of the test users in the verification data set is related to the service scene, the voiceprint model is called, and at least two times of voiceprint verification are respectively carried out on each test user according to the registration audio frequency and the verification audio frequency of each test user in the verification data set, so that the test recognition success rate is obtained. On one hand, the data source used for verifying the voiceprint model can reflect the crowd distribution of a business scene and can truly reflect the use condition of an actual business scene, and on the other hand, login verification is carried out at least twice during login testing to meet the use will of a user, so that the verification result of the voiceprint model recognition effect obtained by the method can truly reflect the use effect of the actual business scene, the model effect deviation caused by non-model factors is eliminated, and the verification accuracy is improved.
Drawings
FIG. 1 is a diagram of an application scenario of a method for verifying a voiceprint model recognition effect in one embodiment;
FIG. 2 is a flow diagram illustrating a method for verifying a voiceprint model recognition effect in one embodiment;
FIG. 3 is a flowchart illustrating a method for verifying the recognition effect of a voiceprint model in another embodiment;
FIG. 4 is a diagram illustrating analysis of verification audio in one embodiment;
FIG. 5 is a flowchart illustrating steps for obtaining a service identification success rate in one embodiment;
FIG. 6 is a block diagram of an apparatus for verifying the recognition effect of a voiceprint model in one embodiment;
FIG. 7 is a diagram illustrating an internal structure of a computer device according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
The method, the device, the computer equipment and the storage medium for verifying the voiceprint model recognition effect can be applied to the application environment shown in fig. 1. Wherein the terminal 102 communicates with the server 104 via a network. The application program of the terminal 102 supports voiceprint verification, and acquires a voice signal of a user through the terminal, and sends the voice signal to the server, and the server calls a voiceprint model to perform identification and authentication. The server also verifies the recognition effect of the voiceprint model. Specifically, a server acquires a verification data set, wherein the population distribution of test users in the verification data set is related to a service scene, the verification data set comprises registration audios and verification audios of the test users, and each test user at least has two groups of verification audios; calling a voiceprint model, and respectively carrying out at least two times of voiceprint verification on each test user according to the registration audio and the verification audio of each test user in the verification data set to obtain the success rate of test identification; and obtaining a verification result of the voiceprint model recognition effect according to the test recognition success rate.
The terminal 102 may be, but not limited to, various personal computers, notebook computers, smart phones, tablet computers, and portable wearable devices, and the server 104 may be implemented by an independent server or a server cluster formed by a plurality of servers.
In one embodiment, as shown in fig. 2, a method for verifying the recognition effect of a voiceprint model is provided, which is described by taking the method as an example applied to the server in fig. 1, and includes the following steps:
step 202, obtaining a verification data set, wherein the population distribution of test users in the verification data set is related to a service scene, the verification data set comprises registration audios and verification audios of the test users, and each test user at least comprises two groups of verification audios.
Many times, voiceprint recognition can be influenced by external factors and personal factors comprehensively to cause model effect deviation, and the influence of the part needs to be eliminated as much as possible during evaluation, so that for selection of test crowds, scene classification needs to be introduced to perfect an evaluated test data source. In the application, the crowd distribution of the test users in the verification data set is related to the service scene. Specifically, the crowd distribution of the test user can be determined according to the service scene of the application program applied by the voiceprint model, so that the test user in the verification data set for the test conforms to the crowd distribution of the actual service scene, and the using effect of the voiceprint model in the service scene is realized closely. For example, in a learning application where the user population is primarily students, the main population is distributed among the teenager population. The test user increases the proportion of the adolescent population. And if the user group is mainly young and strong years, the test user increases the proportion of the young and strong years.
Wherein the verification data set includes a registration audio and a verification audio for each test user. The registration audio may be at least three groups and the verification audio may be at least two groups. The registered audio simulation is audio data when the user registers, and is a reference for verification. The verification audio simulates audio data submitted by a user in a user scene identified by a voiceprint, such as audio data submitted during login verification. The audio contents of the registration audio and the verification audio may be determined according to an actual service.
Specifically, the step of obtaining a verification data set comprises: acquiring the crowd distribution of a service scene; determining the number of test users of each gender and age stage according to the crowd distribution to obtain target test users; and acquiring at least three groups of registration audios and at least two groups of verification audios of each target test user to obtain a verification data set.
Specifically, original audio data required by voiceprint verification is acquired in a targeted mode according to the division of gender and age groups, male and female are divided according to the gender, young, adult and old are divided according to the age group, and finally audio files combined by the gender and the age are obtained through combination to obtain a verification data set.
In one application scenario, the main group of users of a financial application is young and middle-aged people 25-54 years old, and as shown in table 1, the number of test users in the age stage of 25-54 among the test users in the verification data set is the greatest in proportion.
Table 1 verification of test user profiles for data sets
The scene of the application program reflects the target audience distribution condition of the application program, and the testing crowd of the application program is determined according to the scene of the application program, so that the verification data set for testing can reflect the crowd distribution condition in a service scene, and the reliability of the verification effect is improved.
And step 204, calling a voiceprint model, and respectively carrying out at least two times of voiceprint verification on each test user according to the registration audio and the verification audio of each test user in the verification data set to obtain the test identification success rate.
Specifically, the trained voiceprint model is called, and login verification is performed twice for each test user. The test identification success rate is the ratio of successful user identification in the verification data set obtained by verifying the voiceprint model by the verification data set.
The calling of the voiceprint model, at least twice voiceprint verification of each test user according to the registration audio and the verification audio of each test user in the verification data set, and obtaining the success rate of test identification, comprises the following steps: calling a voiceprint model, and respectively carrying out at least two times of voiceprint verification on each test user according to the registration audio frequency and the verification audio frequency of each test user in the verification data set to obtain a primary recognition success rate, a secondary recognition success rate and a recognition success rate; carrying out weighted summation on the primary identification success rate, the secondary identification success rate and the identification success rate to obtain a test identification success rate; wherein the weight of the primary recognition success rate is greater than the weight of the secondary recognition success rate and the weight of the recognition success rate.
And each test user is specifically provided with N groups of verification audios, wherein N is more than or equal to 2. The success rate of the first time is the number of users/the number of test users who have all succeeded in the voiceprint recognition for N times, the success rate of the second time is the number of users/the number of test users who have failed in the voiceprint recognition for N times only for 1 time, and the success rate of the login is the number of users/the number of test users who have succeeded for one time or more in the voiceprint recognition for N times.
And (3) each test user in the verification data set carries out voiceprint recognition verification at least twice respectively, and a primary recognition success rate, a secondary recognition success rate and a recognition success rate are respectively obtained according to the number of all successful users identified by the voiceprints for N times, the number of users failed by the voiceprint identification for N times for 1 time only and the number of users successfully identified by the voiceprints for N times for one time or more.
The first identification success rate, the second identification success rate and the login success rate respectively have different weights, wherein the weight of the first identification success rate is the highest, the weight of the second identification success rate is the next to the second, the weight of the identification success rate is the lowest, and the sum of the first identification success rate, the second identification success rate and the login success rate is 1.
Testing and identifying success rate is A primary identifying success rate + B secondary identifying success rate + C logging success rate
And step 206, obtaining a verification result of the voiceprint model recognition effect according to the test recognition success rate.
In the use process of a practical user, after one-time voiceprint recognition fails, the user usually tries the voiceprint verification for the second time, but the user experience is influenced when the voiceprint is failed for many times, and the intention of the user to use the voiceprint verification product is greatly discounted. In the embodiment of the application, the voiceprint verification is performed on the test user at least twice, the recognition effect is comprehensively evaluated according to the primary recognition success rate and the secondary recognition success rate, the use habit of the user is reflected, whether the user wants to use a voiceprint product or not is reflected, and therefore the index verification mode belongs to an index closer to a service and has a strong guiding significance.
According to the method for verifying the voiceprint model recognition effect, the crowd distribution of the test users in the verification data set is related to the service scene, the voiceprint model is called, and voiceprint verification is performed on each test user at least twice respectively according to the registration audio frequency and the verification audio frequency of each test user in the verification data set, so that the test recognition success rate is obtained. On one hand, the data source used for verifying the voiceprint model can reflect the crowd distribution of a business scene and can truly reflect the use condition of an actual business scene, and on the other hand, login verification is carried out at least twice during login testing to meet the use will of a user, so that the verification result of the voiceprint model recognition effect obtained by the method can truly reflect the use effect of the actual business scene, the model effect deviation caused by non-model factors is eliminated, and the verification accuracy is improved. In another embodiment, as shown in fig. 3, the method further comprises:
step 302, obtaining a verification data set, where the population distribution of a test user in the verification data set is related to a business scenario, where the verification data set includes a registration audio and a verification audio of the test user, and each test user has at least two groups of verification audios.
And step 304, calling a voiceprint model, and respectively carrying out at least two times of voiceprint verification on each test user according to the registration audio and the verification audio of each test user in the verification data set to obtain the test identification success rate.
And step 306, verifying the number of successfully identified users in the users with normal audio frequency according to the actual service scene to obtain the service identification success rate under the normal audio frequency condition.
The verification that the audio is normal refers to the condition that the verification audio submitted by voiceprint recognition verification is not abnormal in an actual service scene. Specifically, the verification audio may be analyzed to evaluate whether the verification audio is abnormal. The abnormal conditions include poor audio quality, invalid data, login and attention of a non-identical person, and the like.
In an actual business scenario, voiceprint authentication failures may be the cause of the sound itself, for example, the audio quality is poor, login and registration are not one, etc. These are validation failures caused by voice problems, not voiceprint model effect problems. As shown in fig. 4, in an actual service scenario, the verification audio data is analyzed, where the verification audio accounts for only 54% of the normal percentage, and the quality problem of other abnormal sounds also causes login failure, and only if the login of the user with normal login sound succeeds, the recognition effect of the voiceprint model can be reflected really. Therefore, when evaluating the voiceprint model in the actual service scene, users with problems in sound quality should be removed, and only data of users with normal verification audio frequency are considered.
Specifically, as shown in fig. 5, the step of obtaining the service identification success rate includes:
step 502, obtaining verification audio data of users in actual service, analyzing the verification audio data, and obtaining users with normal verification audio.
Specifically, detecting and verifying the signal-to-noise ratio of audio data to obtain high-quality sound data; performing frequency-spectrum-normal analysis and voice recognition on the high-quality sound signal to obtain an effective sound signal; and eliminating the voice signals of the logged and registered non-identical person in the effective voice signals to obtain the user with normal verified audio.
The snr refers to a ratio of signal to noise in an electronic device or electronic system. The signal refers to an electronic signal from the outside of the device to be processed by the device, the noise refers to an irregular extra signal (or information) which does not exist in the original signal generated after passing through the device, and the signal does not change along with the change of the original signal. The larger the signal-to-noise ratio, the smaller the noise mixed in the signal, the higher the sound quality of the sound playback, otherwise the opposite. Through signal-to-noise ratio analysis, the sound signals with large noise can be eliminated.
Spectral analysis is a technique that decomposes complex signals into simpler signals. The spectral analysis is performed to find information (e.g., amplitude, power, strength, or phase) of a signal at different frequencies. The frequency spectrum is a representation of a signal in the time domain in the frequency domain, and can be obtained by performing fourier transform on the signal. The frequency spectrum can indicate that a signal is composed of sine waves of which frequencies, and information such as the size and the phase of each frequency sine wave can also be seen. By frequency spectrum analysis of the sound, sound signals with blank audio and small sound in the verified audio data can be eliminated.
When login verification is carried out, the recorded content is usually determined, and audio signals with invalid audio content or dialect can be eliminated by identifying the verification audio data.
For valid voice signals, there may be portions of the voice signals that are registered as being non-identical, which may be manually identified.
Step 504, acquiring the number of successfully identified users in the users with normal verification audio in the actual service.
For example, 1600 users in 1700 users with normal verification audio log in successfully.
Specifically, the service identification success rate is the number of users who successfully identified/the number of users who normally identified among the users who normally identified, and in the above example, the service identification success rate is 1600/1700. The number of users with normal audio is verified, the audio data of the users with poor audio quality, invalid audio and non-identical login and registration in the actual login service are eliminated, and the evaluation of the voiceprint model identification effect caused by the non-model can be eliminated.
After step 306, further comprising:
and 308, obtaining a verification result of the voiceprint model recognition effect according to the test recognition success rate and the service recognition success rate.
For the voiceprint model which is not on-line, the test recognition success rate can be only adopted as the evaluation standard of the voiceprint model recognition effect.
For the voiceprint model which is on-line, the test recognition success rate and the service recognition success rate can be combined to be used as the evaluation standard of the voiceprint model recognition effect.
Specifically, determining the weight of the test recognition success rate and the service recognition success rate according to the online time of the application program and the number of users with normal verification audio; and weighting the test recognition success rate and the service recognition success rate to obtain the score of the voiceprint model, and obtaining the verification result of the voiceprint model recognition effect according to the score.
Specifically, for a voiceprint model with a short online time, because actual service data are less, test data can be taken as a main data, and a higher weight is set for the success rate of test identification. For the voiceprint model with longer online time, because actual service data are more, the actual service data can be used as a main evaluation reference, and the service identification success rate is set to be higher weight. And weighting the test recognition success rate and the service recognition success rate to obtain the score of the voiceprint model, wherein the higher the score is, the better the recognition effect of the voiceprint model is.
According to the method for verifying the voiceprint model recognition effect, during testing, the evaluated test data source is perfected according to scene classification, model effect deviation caused by the comprehensive influence of external factors and personal factors is eliminated, and for an online model, the online voiceprint model effect is evaluated by analyzing sound, eliminating abnormal sound and eliminating factors of non-model problems.
It should be understood that although the various steps in the flowcharts of fig. 2-3, and 5 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 2-3, and fig. 5 may include multiple sub-steps or multiple stages that are not necessarily performed at the same time, but may be performed at different times, and the order of performing the sub-steps or stages is not necessarily sequential, but may be performed alternately or alternatingly with other steps or at least some of the sub-steps or stages of other steps.
In one embodiment, as shown in fig. 6, there is provided an apparatus for verifying the recognition effect of a voiceprint model, including:
a verification data obtaining module 602, configured to obtain a verification data set, where population distribution of test users in the verification data set is related to a service scenario, the verification data set includes a registration audio and a verification audio of the test users, and each test user has at least two groups of verification audios.
And the recognition analysis module 604 is configured to invoke a voiceprint model, and perform at least two times of voiceprint verification on each test user according to the registration audio and the verification audio of each test user in the verification data set, so as to obtain a test recognition success rate.
And the evaluation module 606 is configured to obtain a verification result of the voiceprint model recognition effect according to the test recognition success rate.
According to the device for verifying the voiceprint model identification effect, the crowd distribution of the test users in the verification data set is related to the service scene, the voiceprint model is called, and voiceprint verification is performed on each test user at least twice respectively according to the registration audio frequency and the verification audio frequency of each test user in the verification data set, so that the test identification success rate is obtained. On one hand, the data source used for verifying the voiceprint model can reflect the crowd distribution of a business scene and can truly reflect the use condition of an actual business scene, and on the other hand, login verification is carried out at least twice during login testing to meet the use will of a user, so that the verification result of the voiceprint model recognition effect obtained by the method can truly reflect the use effect of the actual business scene, the model effect deviation caused by non-model factors is eliminated, and the verification accuracy is improved.
In another embodiment, a recognition analysis module includes:
and the identification module is used for calling the voiceprint model, and performing at least twice voiceprint verification on each test user according to the registration audio and the verification audio of each test user in the verification data set to obtain a primary recognition success rate, a secondary recognition success rate and a recognition success rate.
The analysis module is used for weighting and summing the primary recognition success rate, the secondary recognition success rate and the recognition success rate to obtain a test recognition success rate; wherein the weight of the primary recognition success rate is greater than the weight of the secondary recognition success rate and the weight of the recognition success rate.
In another embodiment, the verification data acquisition module is used for acquiring the crowd distribution of the business scene; determining the number of test users of each gender and age stage according to the crowd distribution to obtain target test users; and acquiring at least three groups of registration audios and at least two groups of verification audios of each target test user to obtain a verification data set.
In another embodiment, the apparatus further comprises:
and the verification analysis module is used for verifying the number of successfully identified users in the users with normal audio frequency according to the actual service scene to obtain the service identification success rate under the normal audio frequency condition.
And the evaluation module is used for obtaining a verification result of the voiceprint model recognition effect according to the test recognition success rate and the service recognition success rate.
In another embodiment, the validation analysis includes, including:
and the audio analysis module is used for acquiring verification audio data of the user in the actual service, analyzing the verification audio data and acquiring the user with normal verification audio.
The verification identification module is used for acquiring the number of successfully identified users in the users with normal verification audio in the actual service;
and the service analysis module is used for calculating the service identification success rate under the normal audio frequency condition according to the number of the users with normal verification audio frequency and the number of the users successfully identified in the users with normal verification audio frequency.
In another embodiment, the audio analysis module is used for detecting and verifying the signal-to-noise ratio of the audio data to obtain high-quality sound data; performing frequency-spectrum-normal analysis and voice recognition on the high-quality sound signal to obtain an effective sound signal; and eliminating the voice signals of the logged and registered non-identical person in the effective voice signals to obtain the user with normal verified audio.
In another embodiment, an evaluation module comprises:
and the weight determining module is used for determining the weight of the test recognition success rate and the service recognition success rate according to the online time of the application program and the number of the users with normal verification audio.
And the weighting processing module is used for weighting the test recognition success rate and the service recognition success rate to obtain the score of the voiceprint model.
And the evaluation analysis module is used for obtaining a verification result of the voiceprint model recognition effect according to the score.
For the specific limitation of the apparatus for verifying the recognition effect of the voiceprint model, reference may be made to the above limitation on the method for verifying the recognition effect of the voiceprint model, and details are not described herein again. The modules in the device for verifying the voiceprint model recognition effect A can be wholly or partially realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, which may be a server, the internal structure of which may be as shown in fig. 7. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used for storing audio data. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a method of verifying the recognition effect of a voiceprint model.
Those skilled in the art will appreciate that the architecture shown in fig. 7 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, there is provided a computer device comprising a memory storing a computer program and a processor implementing the following steps when the processor executes the computer program:
acquiring a verification data set, wherein the population distribution of test users in the verification data set is related to a service scene, the verification data set comprises registration audios and verification audios of the test users, and each test user at least comprises two groups of verification audios;
calling a voiceprint model, and respectively carrying out at least two times of voiceprint verification on each test user according to the registration audio and the verification audio of each test user in the verification data set to obtain the success rate of test identification;
and obtaining a verification result of the voiceprint model recognition effect according to the test recognition success rate.
In one embodiment, the invoking the voiceprint model, and performing at least two voiceprint verifications on each test user according to the registration audio and the verification audio of each test user in the verification data set to obtain the test identification success rate includes:
calling a voiceprint model, and respectively carrying out at least two times of voiceprint verification on each test user according to the registration audio frequency and the verification audio frequency of each test user in the verification data set to obtain a primary recognition success rate, a secondary recognition success rate and a recognition success rate;
carrying out weighted summation on the primary identification success rate, the secondary identification success rate and the identification success rate to obtain a test identification success rate; wherein the weight of the primary recognition success rate is greater than the weight of the secondary recognition success rate and the weight of the recognition success rate.
In one embodiment, the obtaining the verification data set includes:
acquiring the crowd distribution of a service scene;
determining the number of test users of each gender and age stage according to the crowd distribution to obtain target test users;
and acquiring at least three groups of registration audios and at least two groups of verification audios of each target test user to obtain a verification data set.
In one embodiment, the processor, when executing the computer program, performs the steps of:
according to the number of successfully identified users in the users with normal audio frequency, the success rate of service identification under the condition of normal audio frequency is obtained;
the obtaining of the verification result of the voiceprint model recognition effect according to the test recognition success rate includes: and obtaining a verification result of the voiceprint model recognition effect according to the test recognition success rate and the service recognition success rate.
In one embodiment, verifying the number of successfully identified users among users with normal audio according to an actual service scenario to obtain a service identification success rate under a normal audio condition includes:
acquiring verification audio data of a user in actual service, analyzing the verification audio data, and acquiring the user with normal verification audio;
acquiring the number of successfully identified users in the users with normal verification audio in the actual service;
and calculating the success rate of service identification under the normal audio frequency condition according to the number of the users with normal audio frequency verification and the number of the users successfully identified in the users with normal audio frequency verification.
In one embodiment, acquiring verification audio data of a user in an actual service, analyzing the verification audio data, and acquiring a user with normal verification audio, includes:
detecting and verifying the signal-to-noise ratio of the audio data to obtain high-quality sound data;
performing frequency-spectrum-normal analysis and voice recognition on the high-quality sound signal to obtain an effective sound signal;
and eliminating the voice signals of the logged and registered non-identical person in the effective voice signals to obtain the user with normal verified audio.
In one embodiment, obtaining a verification result of the voiceprint model recognition effect according to the test recognition success rate and the service recognition success rate includes:
determining the weight of the test recognition success rate and the service recognition success rate according to the online time of the application program and the number of users with normal verification audio;
weighting the test recognition success rate and the service recognition success rate to obtain a score of a voiceprint model;
and obtaining a verification result of the recognition effect of the voiceprint model according to the score.
In one embodiment, a computer storage medium is provided, having a computer program stored thereon, the computer program, when executed by a processor, implementing the steps of:
acquiring a verification data set, wherein the population distribution of test users in the verification data set is related to a service scene, the verification data set comprises registration audios and verification audios of the test users, and each test user at least comprises two groups of verification audios;
calling a voiceprint model, and respectively carrying out at least two times of voiceprint verification on each test user according to the registration audio and the verification audio of each test user in the verification data set to obtain the success rate of test identification;
and obtaining a verification result of the voiceprint model recognition effect according to the test recognition success rate.
In one embodiment, the invoking the voiceprint model, and performing at least two voiceprint verifications on each test user according to the registration audio and the verification audio of each test user in the verification data set to obtain the test identification success rate includes:
calling a voiceprint model, and respectively carrying out at least two times of voiceprint verification on each test user according to the registration audio frequency and the verification audio frequency of each test user in the verification data set to obtain a primary recognition success rate, a secondary recognition success rate and a recognition success rate;
carrying out weighted summation on the primary identification success rate, the secondary identification success rate and the identification success rate to obtain a test identification success rate; wherein the weight of the primary recognition success rate is greater than the weight of the secondary recognition success rate and the weight of the recognition success rate.
In one embodiment, the obtaining the verification data set includes:
acquiring the crowd distribution of a service scene;
determining the number of test users of each gender and age stage according to the crowd distribution to obtain target test users;
and acquiring at least three groups of registration audios and at least two groups of verification audios of each target test user to obtain a verification data set.
In one embodiment, the processor, when executing the computer program, performs the steps of:
according to the number of successfully identified users in the users with normal audio frequency, the success rate of service identification under the condition of normal audio frequency is obtained;
the obtaining of the verification result of the voiceprint model recognition effect according to the test recognition success rate includes: and obtaining a verification result of the voiceprint model recognition effect according to the test recognition success rate and the service recognition success rate.
In one embodiment, verifying the number of successfully identified users among users with normal audio according to an actual service scenario to obtain a service identification success rate under a normal audio condition includes:
acquiring verification audio data of a user in actual service, analyzing the verification audio data, and acquiring the user with normal verification audio;
acquiring the number of successfully identified users in the users with normal verification audio in the actual service;
and calculating the success rate of service identification under the normal audio frequency condition according to the number of the users with normal audio frequency verification and the number of the users successfully identified in the users with normal audio frequency verification.
In one embodiment, acquiring verification audio data of a user in an actual service, analyzing the verification audio data, and acquiring a user with normal verification audio, includes:
detecting and verifying the signal-to-noise ratio of the audio data to obtain high-quality sound data;
performing frequency-spectrum-normal analysis and voice recognition on the high-quality sound signal to obtain an effective sound signal;
and eliminating the voice signals of the logged and registered non-identical person in the effective voice signals to obtain the user with normal verified audio.
In one embodiment, obtaining a verification result of the voiceprint model recognition effect according to the test recognition success rate and the service recognition success rate includes:
determining the weight of the test recognition success rate and the service recognition success rate according to the online time of the application program and the number of users with normal verification audio;
weighting the test recognition success rate and the service recognition success rate to obtain a score of a voiceprint model;
and obtaining a verification result of the recognition effect of the voiceprint model according to the score.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A method of verifying the recognition effect of a voiceprint model, the method comprising:
acquiring a verification data set, wherein the population distribution of test users in the verification data set is related to a service scene, the verification data set comprises registration audios and verification audios of the test users, and each test user at least comprises two groups of verification audios;
calling a voiceprint model, and respectively carrying out at least two times of voiceprint verification on each test user according to the registration audio and the verification audio of each test user in the verification data set to obtain the success rate of test identification;
and obtaining a verification result of the voiceprint model recognition effect according to the test recognition success rate.
2. The method of claim 1, wherein the invoking the voiceprint model and performing at least two voiceprint verifications on each test user according to the registration audio and the verification audio of each test user in the verification data set to obtain a test identification success rate comprises:
calling a voiceprint model, and respectively carrying out at least two times of voiceprint verification on each test user according to the registration audio frequency and the verification audio frequency of each test user in the verification data set to obtain a primary recognition success rate, a secondary recognition success rate and a recognition success rate;
carrying out weighted summation on the primary identification success rate, the secondary identification success rate and the identification success rate to obtain a test identification success rate; wherein the weight of the primary recognition success rate is greater than the weight of the secondary recognition success rate and the weight of the recognition success rate.
3. The method of claim 1, wherein the obtaining the validation data set comprises:
acquiring the crowd distribution of a service scene;
determining the number of test users of each gender and age stage according to the crowd distribution to obtain target test users;
and acquiring at least three groups of registration audios and at least two groups of verification audios of each target test user to obtain a verification data set.
4. The method of claim 1, further comprising:
according to the number of successfully identified users in the users with normal audio frequency, the success rate of service identification under the condition of normal audio frequency is obtained;
the obtaining of the verification result of the voiceprint model recognition effect according to the test recognition success rate includes: and obtaining a verification result of the voiceprint model recognition effect according to the test recognition success rate and the service recognition success rate.
5. The method of claim 4, wherein obtaining the success rate of service identification under the normal audio frequency condition according to the number of successfully identified users in the users with normal audio frequency verification under the actual service scenario comprises:
acquiring verification audio data of a user in actual service, analyzing the verification audio data, and acquiring the user with normal verification audio;
acquiring the number of successfully identified users in the users with normal verification audio in the actual service;
and calculating the success rate of service identification under the normal audio frequency condition according to the number of the users with normal audio frequency verification and the number of the users successfully identified in the users with normal audio frequency verification.
6. The method of claim 5, wherein obtaining the verification audio data of the user in the actual service, analyzing the verification audio data, and obtaining the user with normal verification audio comprises:
detecting and verifying the signal-to-noise ratio of the audio data to obtain high-quality sound data;
performing frequency-spectrum-normal analysis and voice recognition on the high-quality sound signal to obtain an effective sound signal;
and eliminating the voice signals of the logged and registered non-identical person in the effective voice signals to obtain the user with normal verified audio.
7. The method of claim 4, wherein obtaining a verification result of the voiceprint model recognition effect according to the test recognition success rate and the service recognition success rate comprises:
determining the weight of the test recognition success rate and the service recognition success rate according to the online time of the application program and the number of users with normal verification audio;
weighting the test recognition success rate and the service recognition success rate to obtain a score of a voiceprint model;
and obtaining a verification result of the recognition effect of the voiceprint model according to the score.
8. An apparatus for verifying the recognition effect of a voiceprint model, the apparatus comprising:
the verification data acquisition module is used for acquiring a verification data set, the population distribution of test users in the verification data set is related to a service scene, the verification data set comprises registration audios and verification audios of the test users, and each test user at least has two groups of verification audios;
the identification analysis module is used for calling a voiceprint model, and performing voiceprint verification on each test user at least twice according to the registration audio frequency and the verification audio frequency of each test user in the verification data set to obtain the success rate of test identification;
and the evaluation module is used for obtaining a verification result of the voiceprint model recognition effect according to the test recognition success rate.
9. A computer device comprising a memory and a processor, the memory storing a computer program, wherein the processor implements the steps of the method of any one of claims 1 to 7 when executing the computer program.
10. A computer storage medium on which a computer program is stored, characterized in that the computer program, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011485999.5A CN112599137A (en) | 2020-12-16 | 2020-12-16 | Method and device for verifying voiceprint model recognition effect and computer equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011485999.5A CN112599137A (en) | 2020-12-16 | 2020-12-16 | Method and device for verifying voiceprint model recognition effect and computer equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112599137A true CN112599137A (en) | 2021-04-02 |
Family
ID=75196309
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011485999.5A Pending CN112599137A (en) | 2020-12-16 | 2020-12-16 | Method and device for verifying voiceprint model recognition effect and computer equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112599137A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113611328A (en) * | 2021-06-30 | 2021-11-05 | 公安部第一研究所 | Voiceprint recognition voice evaluation method and device |
| CN114240101A (en) * | 2021-12-02 | 2022-03-25 | 支付宝(杭州)信息技术有限公司 | Risk identification model verification method, device and equipment |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6253179B1 (en) * | 1999-01-29 | 2001-06-26 | International Business Machines Corporation | Method and apparatus for multi-environment speaker verification |
| CN103229233A (en) * | 2010-12-10 | 2013-07-31 | 松下电器产业株式会社 | Modeling device and method for speaker recognition, and speaker recognition system |
| EP2797078A1 (en) * | 2013-04-26 | 2014-10-29 | Agnitio S.L. | Estimation of reliability in speaker recognition |
| CN107610709A (en) * | 2017-08-01 | 2018-01-19 | 百度在线网络技术（北京）有限公司 | A kind of method and system for training Application on Voiceprint Recognition model |
| CN107680600A (en) * | 2017-09-11 | 2018-02-09 | 平安科技（深圳）有限公司 | Sound-groove model training method, audio recognition method, device, equipment and medium |
| CN108305633A (en) * | 2018-01-16 | 2018-07-20 | 平安科技（深圳）有限公司 | Speech verification method, apparatus, computer equipment and computer readable storage medium |
| CN109447563A (en) * | 2018-09-05 | 2019-03-08 | 平安科技（深圳）有限公司 | Voice Work attendance method, device, computer equipment and storage medium |
| CN110415681A (en) * | 2019-09-11 | 2019-11-05 | 北京声智科技有限公司 | A kind of speech recognition effect testing method and system |
| CN111341326A (en) * | 2020-02-18 | 2020-06-26 | RealMe重庆移动通信有限公司 | Voice processing method and related product |
| CN111653283A (en) * | 2020-06-28 | 2020-09-11 | 讯飞智元信息科技有限公司 | Cross-scene voiceprint comparison method, device, equipment and storage medium |
| CN111835522A (en) * | 2020-05-19 | 2020-10-27 | 北京捷通华声科技股份有限公司 | Audio processing method and device |
-
2020
- 2020-12-16 CN CN202011485999.5A patent/CN112599137A/en active Pending
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6253179B1 (en) * | 1999-01-29 | 2001-06-26 | International Business Machines Corporation | Method and apparatus for multi-environment speaker verification |
| CN103229233A (en) * | 2010-12-10 | 2013-07-31 | 松下电器产业株式会社 | Modeling device and method for speaker recognition, and speaker recognition system |
| EP2797078A1 (en) * | 2013-04-26 | 2014-10-29 | Agnitio S.L. | Estimation of reliability in speaker recognition |
| CN107610709A (en) * | 2017-08-01 | 2018-01-19 | 百度在线网络技术（北京）有限公司 | A kind of method and system for training Application on Voiceprint Recognition model |
| CN107680600A (en) * | 2017-09-11 | 2018-02-09 | 平安科技（深圳）有限公司 | Sound-groove model training method, audio recognition method, device, equipment and medium |
| CN108305633A (en) * | 2018-01-16 | 2018-07-20 | 平安科技（深圳）有限公司 | Speech verification method, apparatus, computer equipment and computer readable storage medium |
| CN109447563A (en) * | 2018-09-05 | 2019-03-08 | 平安科技（深圳）有限公司 | Voice Work attendance method, device, computer equipment and storage medium |
| CN110415681A (en) * | 2019-09-11 | 2019-11-05 | 北京声智科技有限公司 | A kind of speech recognition effect testing method and system |
| CN111341326A (en) * | 2020-02-18 | 2020-06-26 | RealMe重庆移动通信有限公司 | Voice processing method and related product |
| CN111835522A (en) * | 2020-05-19 | 2020-10-27 | 北京捷通华声科技股份有限公司 | Audio processing method and device |
| CN111653283A (en) * | 2020-06-28 | 2020-09-11 | 讯飞智元信息科技有限公司 | Cross-scene voiceprint comparison method, device, equipment and storage medium |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113611328A (en) * | 2021-06-30 | 2021-11-05 | 公安部第一研究所 | Voiceprint recognition voice evaluation method and device |
| CN114240101A (en) * | 2021-12-02 | 2022-03-25 | 支付宝(杭州)信息技术有限公司 | Risk identification model verification method, device and equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11842740B2 (en) | Seamless authentication and enrollment | |
| Balamurali et al. | Toward robust audio spoofing detection: A detailed comparison of traditional and learned features | |
| US9099085B2 (en) | Voice authentication systems and methods | |
| WO2020177380A1 (en) | Voiceprint detection method, apparatus and device based on short text, and storage medium | |
| CA2736133C (en) | Voice authentication system and methods | |
| Dey et al. | Speech biometric based attendance system | |
| WO2018166187A1 (en) | Server, identity verification method and system, and a computer-readable storage medium | |
| US9697836B1 (en) | Authentication of users of self service channels | |
| US12010108B2 (en) | Techniques to provide sensitive information over a voice connection | |
| KR20190053917A (en) | Identity recognition method and apparatus | |
| Chen et al. | Towards understanding and mitigating audio adversarial examples for speaker recognition | |
| WO2021042537A1 (en) | Voice recognition authentication method and system | |
| CN111858892A (en) | Voice interaction method, device, equipment and medium based on knowledge graph | |
| CN112599137A (en) | Method and device for verifying voiceprint model recognition effect and computer equipment | |
| US20210366489A1 (en) | Voice authentication system and method | |
| Chen et al. | AS2T: Arbitrary source-to-target adversarial attack on speaker recognition systems | |
| Poh et al. | Incorporating model-specific score distribution in speaker verification systems | |
| US20220246153A1 (en) | System and method for detecting fraudsters | |
| US11841932B2 (en) | System and method for updating biometric evaluation systems | |
| CN110138743A (en) | Acquisition methods, device, computer equipment and the storage medium of resource bid file | |
| CN111063359B (en) | Telephone return visit validity judging method, device, computer equipment and medium | |
| CN111199742A (en) | Identity verification method and device and computing equipment | |
| TWM622203U (en) | Voiceprint identification device for financial transaction system | |
| TWI817897B (en) | Low-noise voiceprint identification device for financial transaction system and method thereof | |
| US20220375461A1 (en) | System and method for voice biometrics authentication |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |