Voiceprint comparison method and device for multi-source data and readable medium

Abstract

The invention discloses a voiceprint comparison method and device for multi-source data and a readable medium, and the method comprises the steps: obtaining voice data from different sources, extracting the voiceprint features of the voice data, and obtaining a voiceprint feature library of each source; obtaining a to-be-recognized voice and a registered voice, and respectively extracting corresponding voiceprint features; the similarity between the two sources and each voiceprint feature in the voiceprint feature library of each source is calculated, the mean value of the similarity is solved, and the corresponding assumed source is determined; calculating the similarity between the voiceprint feature of the voice to be recognized and each voiceprint feature in the voiceprint feature library of the assumed source of the registered voice, and calculating the mean value of the similarity to obtain a first mean value; the similarity between the voiceprint feature of the registered voice and each voiceprint feature in the voiceprint feature library of the assumed source of the voice to be recognized is calculated, the mean value of the similarity is solved, a second mean value is obtained, and whether the voice to be recognized and the registered voice belong to the same person or not is judged based on the first mean value and the second mean value; and the voiceprint scoring result distribution of unknown sources has consistency.

Description

Technical Field

[0001] The present invention relates to the field of voiceprint recognition, and in particular to a method, device and readable medium for voiceprint comparison of multi-source data. Background Art

[0002] Voiceprint comparison refers to extracting the voiceprint features of two voices and scoring their similarity, and judging the possibility that the speakers of the two voices are the same based on the scoring results. In actual scenarios, the sources of voices are diverse. Under the same voiceprint recognition model, voice data from different sources will affect the scoring results. Assuming that under normal circumstances, the similarity is above 80, it is the same person. For example, when comparing voiceprints in scenario A (telephone), the scoring result may be high, and it is believed that the similarity is above 90, and the voice is considered to be the same person when the similarity is above 70. When comparing voiceprints in scenario B (WeChat), the scoring result may be low, and it is believed that the similarity is above 70, it is the same person. In actual applications, the source of the voice may not be available or difficult to obtain, so it is important to solve the problem of floating thresholds caused by voices from different sources when comparing voiceprints. Summary of the invention

[0003] The purpose of this application is to propose a voiceprint comparison method, device and readable medium for multi-source data in response to the above-mentioned technical problems.

[0004] In a first aspect, the present invention provides a voiceprint comparison method for multi-source data, comprising the following steps:

[0005] Acquire voice data transmitted from different sources, extract voiceprint features of each voice data, and store them in separate libraries according to the source of the voice data to obtain a voiceprint feature library corresponding to each source;

[0006] Acquire the speech to be recognized and the registered speech of unknown sources, and extract the voiceprint features of the speech to be recognized and the voiceprint features of the registered speech respectively;

[0007] Calculate the similarity between the voiceprint feature of the speech to be recognized and each voiceprint feature in the voiceprint feature library of each source and obtain the average of the similarities; calculate the similarity between the voiceprint feature of the registered speech and each voiceprint feature in the voiceprint feature library of each source and obtain the average of the similarities; select the source corresponding to the maximum average of the similarities as the presumed source of the speech to be recognized and the presumed source of the registered speech;

[0008] Calculate the similarity between the voiceprint feature of the speech to be recognized and each voiceprint feature in the voiceprint feature library of the assumed source of the registered speech, and take the average of the similarities to obtain a first mean; calculate the similarity between the voiceprint feature of the registered speech and each voiceprint feature in the voiceprint feature library of the assumed source of the speech to be recognized and take the average of the similarities to obtain a second mean; obtain a comparison result between the speech to be recognized and the registered speech based on the first mean and the second mean; and determine whether the speech to be recognized and the registered speech belong to the same person according to the comparison result.

[0009] Preferably, the different sources include a telephone and at least one software having a calling function.

[0010] Preferably, the comparison result of the speech to be recognized and the registered speech is obtained based on the first mean and the second mean, specifically including:

[0011] First, calculate the similarity between the voiceprint features of the speech to be recognized and the voiceprint features of the registered speech;

[0012] The comparison result between the speech to be recognized and the registered speech is calculated based on the similarity, the first mean and the second mean, as shown in the following formula:

[0013] score=llr(v a ,v b )-(m a +m b ) / 2;

[0014] Among them, score represents the comparison result, llr(v a ,v b ) represents the voiceprint feature v of the speech to be recognized a And the voiceprint feature v of the registered voice b similarity, llr() represents the calculation function of similarity, m a represents the first mean, m b Represents the second mean.

[0015] Preferably, the similarity calculation function includes a cosine scoring function or a plda scoring function.

[0016] Preferably, judging whether the voice to be recognized and the registered voice belong to the same person according to the comparison result specifically includes:

[0017] The comparison result is compared with a fixed threshold. If the comparison result is greater than or equal to the fixed threshold, it is determined that the voice to be recognized and the registered voice belong to the same person.

[0018] If the comparison result is less than a fixed threshold, it is determined that the voice to be recognized and the registered voice do not belong to the same person.

[0019] Preferably, the voiceprint features are extracted using the same voiceprint recognition model or voiceprint feature extraction model.

[0020] In a second aspect, the present invention provides a voiceprint comparison device for multi-source data, comprising:

[0021] A voiceprint feature library building module is configured to obtain voice data transmitted from different sources, extract the voiceprint features of each voice data, and store them in separate libraries according to the sources of the voice data to obtain a voiceprint feature library corresponding to each source;

[0022] A voiceprint feature extraction module is configured to obtain a to-be-recognized voice and a registered voice of unknown source, and extract the voiceprint features of the to-be-recognized voice and the voiceprint features of the registered voice respectively;

[0023] The presumed source determination module is configured to calculate the similarity between the voiceprint feature of the speech to be recognized and each voiceprint feature in the voiceprint feature library of each source and obtain the average of the similarities, calculate the similarity between the voiceprint feature of the registration speech and each voiceprint feature in the voiceprint feature library of each source and obtain the average of the similarities, and select the source corresponding to the maximum average of the similarities as the presumed source of the speech to be recognized and the presumed source of the registration speech;

[0024] The comparison module is configured to calculate the similarity between the voiceprint feature of the voice to be recognized and each voiceprint feature in the voiceprint feature library of the assumed source of the registered voice and to obtain the average of the similarities to obtain a first average, calculate the similarity between the voiceprint feature of the registered voice and each voiceprint feature in the voiceprint feature library of the assumed source of the voice to be recognized and to obtain the average of the similarities to obtain a second average, obtain the comparison result of the voice to be recognized and the registered voice based on the first average and the second average, and determine whether the voice to be recognized and the registered voice belong to the same person according to the comparison result.

[0025] In a third aspect, the present invention provides an electronic device comprising one or more processors; a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any implementation manner in the first aspect.

[0026] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in any implementation manner in the first aspect.

[0027] In a fifth aspect, the present invention provides a computer program product, comprising a computer program, which, when executed by a processor, implements the method described in any implementation manner in the first aspect.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] (1) The voiceprint comparison method for multi-source data proposed in the present invention establishes voiceprint feature libraries from different sources, so that the speech to be recognized and the registered speech are compared and scored with each voiceprint feature in the voiceprint feature libraries from different sources, and the average of all scoring results is calculated, thereby eliminating the differences in comparison results from different sources and making the distribution of voiceprint comparison results of speech from different sources consistent. In this way, the thresholds used subsequently are consistent, that is, a fixed threshold can be used, and there is no need to adjust the threshold size according to different sources.

[0030] (2) The voiceprint comparison method for multi-source data proposed in the present invention uses the high average similarity of voiceprint feature scores of the same source data to make source assumptions, and corrects the final comparison results to improve the usability of actual scenarios.

[0031] (3) The multi-source data voiceprint comparison method proposed in the present invention can accurately identify voiceprints from different sources, and this method has no restrictions on the voiceprint recognition model and is highly applicable. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0033] Figure 1 A schematic diagram of the flow of a voiceprint comparison method for multi-source data according to an embodiment of the present application;

[0034] Figure 2 A schematic diagram of a voiceprint comparison device for multi-source data according to an embodiment of the present application;

[0035] Figure 3 A schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0036] In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0037] Figure 1A voiceprint comparison method for multi-source data provided by an embodiment of the present application is shown, comprising the following steps:

[0038] S1, obtain voice data transmitted from different sources, extract the voiceprint features of each voice data respectively, and store them in separate libraries according to the sources of the voice data to obtain a voiceprint feature library corresponding to each source.

[0039] In a specific embodiment, the different sources include a telephone and at least one software having a calling function.

[0040] Specifically, voice data from different sources are collected, and one voice data from each person is collected from each source, for a total of 1,000 pieces. Different sources include telephones or software with call functions such as WeChat and QQ. When collecting, try to ensure that other conditions remain unchanged and good. The voice data collected from different sources are processed, and their voiceprint features are extracted using a voiceprint recognition model or a voiceprint feature extraction model. The voiceprint feature libraries are divided according to the source to obtain voiceprint feature libraries from different sources. For example, the voiceprint feature library constructed from the voice data collected from the phone is the phone voiceprint feature library A. 电话 The voiceprint feature library constructed from the voice data collected from the WeChat software is the WeChat voiceprint feature library A. 微信 , and so on, to build a voiceprint feature library for each source.

[0041] S2, obtaining the speech to be recognized and the registered speech of unknown sources, and extracting the voiceprint features of the speech to be recognized and the voiceprint features of the registered speech respectively.

[0042] In a specific embodiment, the voiceprint features are extracted using the same voiceprint recognition model or voiceprint feature extraction model.

[0043] Specifically, for the speech to be recognized whose source is unknown, the corresponding voiceprint feature is extracted to obtain the voiceprint feature v of the speech to be recognized a In the embodiment of the present application, the source of the registered voice may be known or unknown, which is not limited here. The speaker corresponding to the registered voice is a known speaker, that is, the identity information of the speaker corresponding to the registered voice is known. Similarly, the voiceprint feature corresponding to the registered voice is extracted to obtain the voiceprint feature v of the registered voice. b In the embodiment of the present application, the same voiceprint recognition model or voiceprint feature extraction model as that in step S1 is used as much as possible to extract voiceprint features from the to-be-recognized voice and the registered voice, so as to avoid differences in the extracted voiceprint features due to differences in the voiceprint recognition model or voiceprint feature extraction model. The voiceprint recognition model or voiceprint feature extraction model can adopt an existing network model or an improved network model, and its structure is not limited.

[0044] S3, calculate the similarity between the voiceprint features of the speech to be recognized and each voiceprint feature in the voiceprint feature library of each source and obtain the average of the similarities, calculate the similarity between the voiceprint features of the registered speech and each voiceprint feature in the voiceprint feature library of each source and obtain the average of the similarities, and select the source corresponding to the largest average of the similarities as the presumed source of the speech to be recognized and the presumed source of the registered speech.

[0045] Specifically, for the voiceprint features of the voice to be recognized and the voiceprint features of the registered voice, the voiceprint similarity is calculated in different source libraries and the average is obtained. 电话 In the example, the voiceprint feature v of the speech to be recognized is a and telephone voiceprint feature library A 电话 The 1000 voiceprint features in are compared and scored, and then the mean value m of the scoring results is calculated. 电话 , and so on, we can get the voiceprint feature v of the speech to be recognized a The similarity between each voiceprint feature in the voiceprint feature library from other sources is calculated and its mean is calculated; since the mean calculated for voiceprint features from the same source will be higher, the source corresponding to the largest mean is considered to be the assumed source of the voiceprint feature of the speech to be recognized. a , and also register the voiceprint feature v b The similarity between each voiceprint feature in the voiceprint feature library of each source is calculated and the mean is calculated. The source corresponding to the largest mean is selected as the assumed source of the registered voice. b .s a 、s b ∈{phone, WeChat, QQ, …}.

[0046] S4, calculate the similarity between the voiceprint feature of the voice to be recognized and each voiceprint feature in the voiceprint feature library of the assumed source of the registered voice, and take the average of the similarities to obtain a first mean, calculate the similarity between the voiceprint feature of the registered voice and each voiceprint feature in the voiceprint feature library of the assumed source of the voice to be recognized and take the average of the similarities to obtain a second mean, calculate the comparison result of the voice to be recognized and the registered voice based on the first mean and the second mean, and judge whether the voice to be recognized and the registered voice belong to the same person according to the comparison result.

[0047] In a specific embodiment, the comparison result of the speech to be recognized and the registered speech is obtained based on the first mean and the second mean, specifically including:

[0048] First, calculate the similarity between the voiceprint features of the speech to be recognized and the voiceprint features of the registered speech;

[0049] The comparison result between the speech to be recognized and the registered speech is calculated based on the similarity, the first mean and the second mean, as shown in the following formula:

[0050] score=llr(v a ,v b )-(m a +m b ) / 2;

[0051] Among them, score represents the comparison result, llr(v a ,v b ) represents the voiceprint feature v of the speech to be recognized a And the voiceprint feature v of the registered voice b similarity, llr() represents the calculation function of similarity, m a represents the first mean, m b Represents the second mean.

[0052] In a specific embodiment, the similarity calculation function includes a cosine scoring function or a plda scoring function.

[0053] In a specific embodiment, judging whether the voice to be recognized and the registered voice belong to the same person according to the comparison result specifically includes:

[0054] The comparison result is compared with a fixed threshold. If the comparison result is greater than or equal to the fixed threshold, it is determined that the voice to be recognized and the registered voice belong to the same person.

[0055] If the comparison result is less than a fixed threshold, it is determined that the voice to be recognized and the registered voice do not belong to the same person.

[0056] Specifically, the voiceprint feature v of the speech to be recognized a with the supposed source of the registered speech b Compare and score each voiceprint feature in the voiceprint feature library, and calculate the mean of the scoring results to obtain the first mean m a , the voiceprint feature v of the registered speech b The assumed source of the speech to be recognized a Compare and score each voiceprint feature in the voiceprint feature library, and calculate the mean of the scoring results to obtain the first mean m b According to the first mean m a and the first mean m b The voiceprint feature v of the speech to be recognized a The voiceprint feature v of the registered voice b The similarity calculation formula is corrected. In the voiceprint feature v of the speech to be recognized a The voiceprint feature v of the registered voice b Based on the scoring results obtained by comparison, subtract the first mean m a and the first mean m bThe mean of is used to eliminate the differences in scoring results from different sources, make the scoring results of voiceprint features from different sources consistent, and make the thresholds used subsequently consistent.

[0057] Further references Figure 2 As an implementation of the methods shown in the above figures, the present application provides an embodiment of a voiceprint comparison device for multi-source data. Figure 1 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0058] The present application provides a voiceprint comparison device for multi-source data, including:

[0059] The voiceprint feature library building module 1 is configured to obtain voice data transmitted from different sources, extract the voiceprint features of each voice data, and store them in separate libraries according to the source of the voice data to obtain a voiceprint feature library corresponding to each source;

[0060] The voiceprint feature extraction module 2 is configured to obtain the to-be-recognized voice and the registered voice of unknown sources, and respectively extract the voiceprint features of the to-be-recognized voice and the voiceprint features of the registered voice;

[0061] The presumed source determination module 3 is configured to calculate the similarity between the voiceprint feature of the speech to be recognized and each voiceprint feature in the voiceprint feature library of each source and obtain the average of the similarities, calculate the similarity between the voiceprint feature of the registration speech and each voiceprint feature in the voiceprint feature library of each source and obtain the average of the similarities, and select the source corresponding to the maximum average of the similarities as the presumed source of the speech to be recognized and the presumed source of the registration speech;

[0062] The comparison module 4 is configured to calculate the similarity between the voiceprint feature of the voice to be recognized and each voiceprint feature in the voiceprint feature library of the assumed source of the registered voice and to obtain the average of the similarities to obtain a first average, calculate the similarity between the voiceprint feature of the registered voice and each voiceprint feature in the voiceprint feature library of the assumed source of the voice to be recognized and to obtain the average of the similarities to obtain a second average, obtain the comparison result of the voice to be recognized and the registered voice based on the first average and the second average, and determine whether the voice to be recognized and the registered voice belong to the same person according to the comparison result.

[0063] Figure 3 Schematic diagram of the hardware structure of the electronic device provided by the embodiment of the present invention. Figure 3As shown, the electronic device of this embodiment includes: a processor 301 and a memory 302; wherein the memory 302 is used to store computer-executable instructions; the processor 301 is used to execute the computer-executable instructions stored in the memory to implement the various steps performed by the electronic device in the above embodiment. For details, please refer to the relevant description in the above method embodiment.

[0064] Optionally, the memory 302 may be independent or integrated with the processor 301 .

[0065] When the memory 302 is independently provided, the electronic device further includes a bus 303 for connecting the memory 302 and the processor 301 .

[0066] The embodiment of the present invention further provides a computer storage medium, in which computer execution instructions are stored. When the processor 301 executes the computer execution instructions, the above method is implemented.

[0067] The embodiment of the present invention further provides a computer program product, including a computer program. When the computer program is executed by the processor 301, the above method is implemented.

[0068] In the embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of modules is only a logical function division, and there may be other division methods in actual implementation, such as multiple modules can be combined or integrated into another system, or some features can be ignored or not executed. Another point, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or modules, which can be electrical, mechanical or other forms.

[0069] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to implement the solution of this embodiment.

[0070] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, or each module may exist physically separately, or two or more modules may be integrated into one unit. The above-mentioned module-composed unit may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0071] The above-mentioned integrated module implemented in the form of a software function module can be stored in a computer-readable storage medium. The above-mentioned software function module is stored in a storage medium, including a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor 301 to perform some steps of the methods of various embodiments of the present application.

[0072] It should be understood that the processor 301 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. A general-purpose processor may be a microprocessor or the processor 301 may be any conventional processor 301, etc. The steps of the method disclosed in the invention may be directly embodied in the execution of the hardware processor 301, or may be executed by a combination of hardware and software modules in the processor 301.

[0073] The memory 302 may include a high-speed RAM memory, and may also include a non-volatile storage NVM, such as at least one disk memory, and may also be a USB flash drive, a mobile hard disk, a read-only memory, a magnetic disk, or an optical disk.

[0074] The bus 303 may be an Industry Standard Architecture (ISA), a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus 303 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, the bus 303 in the drawings of the present application is not limited to only one bus 303 or one type of bus 303.

[0075] The above storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The storage medium can be any available medium that can be accessed by a general or special purpose computer.

[0076] An exemplary storage medium is coupled to the processor 301, so that the processor 301 can read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be a component of the processor 301. The processor 301 and the storage medium can be located in an application specific integrated circuit (Application Specific Integrated Circuits, referred to as ASIC). Of course, the processor 301 and the storage medium can also exist as discrete components in an electronic device or a main control device.

[0077] Those skilled in the art can understand that all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, disk or optical disk and other media that can store program codes.

[0078] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A voiceprint comparison method for multi-source data, characterized in that: The following steps are involved: Acquire voice data transmitted from different sources, extract voiceprint features of each voice data respectively, and store them in separate libraries according to the sources of the voice data to obtain a voiceprint feature library corresponding to each source; Acquire a speech to be recognized and a registered speech of unknown source, and extract voiceprint features of the speech to be recognized and voiceprint features of the registered speech respectively; Calculating the similarity between the voiceprint feature of the speech to be recognized and each voiceprint feature in the voiceprint feature library of each source and obtaining the average of the similarities, calculating the similarity between the voiceprint feature of the registration speech and each voiceprint feature in the voiceprint feature library of each source and obtaining the average of the similarities, and selecting the source corresponding to the maximum average of the similarities as the presumed source of the speech to be recognized and the presumed source of the registration speech; Calculate the similarity between the voiceprint feature of the speech to be recognized and each voiceprint feature in the voiceprint feature library of the assumed source of the registered speech, and take the average of the similarities to obtain a first mean; calculate the similarity between the voiceprint feature of the registered speech and each voiceprint feature in the voiceprint feature library of the assumed source of the speech to be recognized and take the average of the similarities to obtain a second mean; obtain a comparison result of the speech to be recognized and the registered speech based on the first mean and the second mean, and determine whether the speech to be recognized and the registered speech belong to the same person according to the comparison result.

2. The voiceprint comparison method of multi-source data according to claim 1 is characterized in that: The different sources include a telephone and at least one software having a calling function.

3. The voiceprint comparison method of multi-source data according to claim 1 is characterized in that: The comparison result between the to-be-recognized voice and the registered voice is obtained by calculating based on the first mean and the second mean, specifically including: First, calculating the similarity between the voiceprint features of the to-be-recognized speech and the voiceprint features of the registered speech; The comparison result of the speech to be recognized and the registered speech is calculated according to the similarity, the first mean and the second mean, as shown in the following formula: score=llr(v a ,v b )-(m a +m b ) / 2; Among them, score represents the comparison result, llr(v a ,v b ) represents the voiceprint feature v of the speech to be recognized a And the voiceprint feature v of the registered voice b similarity, llr() represents the calculation function of similarity, m a represents the first mean, m b Represents the second mean.

4. The voiceprint comparison method of multi-source data according to claim 3 is characterized in that: The similarity calculation function includes a cosine scoring function or a plda scoring function.

5. The voiceprint comparison method of multi-source data according to claim 1, characterized in that: Determining whether the voice to be recognized and the registered voice belong to the same person according to the comparison result specifically includes: Comparing the comparison result with a fixed threshold, if the comparison result is greater than or equal to the fixed threshold, determining that the to-be-recognized voice and the registered voice belong to the same person; If the comparison result is less than a fixed threshold, it is determined that the voice to be recognized and the registered voice do not belong to the same person.

6. The voiceprint comparison method of multi-source data according to claim 1, characterized in that: The voiceprint features are extracted using the same voiceprint recognition model or voiceprint feature extraction model.

7. A voiceprint comparison device for multi-source data, characterized in that: include: A voiceprint feature library building module is configured to obtain voice data transmitted from different sources, extract the voiceprint features of each voice data, and store them in separate libraries according to the sources of the voice data to obtain a voiceprint feature library corresponding to each source; A voiceprint feature extraction module is configured to obtain a to-be-recognized voice and a registered voice of unknown source, and extract the voiceprint features of the to-be-recognized voice and the voiceprint features of the registered voice respectively; The presumed source determination module is configured to calculate the similarity between the voiceprint feature of the speech to be recognized and each voiceprint feature in the voiceprint feature library of each source and obtain the average of the similarities, calculate the similarity between the voiceprint feature of the registration speech and each voiceprint feature in the voiceprint feature library of each source and obtain the average of the similarities, and select the source corresponding to the maximum average of the similarities as the presumed source of the speech to be recognized and the presumed source of the registration speech; The comparison module is configured to calculate the similarity between the voiceprint feature of the voice to be recognized and each voiceprint feature in the voiceprint feature library of the assumed source of the registered voice and to obtain the average of the similarities to obtain a first average, calculate the similarity between the voiceprint feature of the registered voice and each voiceprint feature in the voiceprint feature library of the assumed source of the voice to be recognized and to obtain the average of the similarities to obtain a second average, obtain the comparison result of the voice to be recognized and the registered voice based on the first average and the second average, and determine whether the voice to be recognized and the registered voice belong to the same person according to the comparison result.

8. An electronic device comprising: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images