Voiceprint data processing method, voiceprint recognition method, equipment and storage medium
By extending and eliminating features during the voiceprint data recording phase, target voiceprint data that meets length requirements and avoids mixing in invalid fragments is generated, thus solving the problem of low recognition accuracy in existing voiceprint recognition technology and achieving higher recognition accuracy.
Patent Information
- Application Number
- CN202510915160.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-03
- Publication Date
- 2025-09-23
AI Technical Summary
Existing voiceprint recognition technologies generally face the problem of low recognition accuracy in practical applications.
By performing data extension and feature removal processing during the voiceprint data recording stage, target voiceprint data that meets the length requirements and avoids mixing in invalid fragments is generated, including extension processing and removal of non-voiceprint features, ensuring data integrity and purity.
The accuracy of voiceprint recognition is improved. By generating target voiceprint data that meets the length requirements, the interference of invalid fragments is avoided, and the accuracy of recognition is improved.
Smart Images
Figure CN120690209A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of voiceprint information processing technology, and in particular to a voiceprint data processing method, a voiceprint recognition method, a device, and a storage medium. Background Art
[0002] In today's digital age, voiceprint recognition technology, as a crucial component of biometrics, is widely used in numerous fields, including security authentication, smart home control, and financial transactions. It extracts and analyzes the characteristics of human voice to identify individuals. However, current practical applications of voiceprint recognition technology generally face the technical challenge of low recognition accuracy. Summary of the Invention
[0003] The purpose of this application is to provide a voiceprint data processing method, a voiceprint recognition method, a device and a storage medium, aiming to improve the accuracy of voiceprint recognition.
[0004] The present invention provides a method for processing voiceprint data, including: Obtain the voiceprint data to be recorded; Determining whether the duration of the voiceprint data to be recorded reaches a preset threshold duration; If not, performing extension processing on the voiceprint data to be recorded to obtain first extended voiceprint data; If so, the non-voiceprint features in the first voiceprint data to be processed are eliminated to obtain the target voiceprint data; when the duration of the voiceprint data to be recorded does not reach the preset threshold duration, the first voiceprint data to be processed is the first extended voiceprint data, otherwise the first voiceprint data to be processed is the voiceprint data to be recorded.
[0005] In some embodiments, the extending processing of the voiceprint data to be recorded includes: Copy several copies of the voiceprint data to be recorded; The voiceprint data to be recorded are spliced into voiceprint data whose duration reaches the preset threshold duration to obtain the first extended voiceprint data.
[0006] In some embodiments, the extending processing of the voiceprint data to be recorded includes: Extracting a voiceprint data segment containing a voiceprint feature from the voiceprint data to be recorded; The voiceprint data to be recorded and several voiceprint data segments are spliced together to form voiceprint data with a duration reaching the preset threshold duration, to obtain the first extended voiceprint data.
[0007] In some embodiments, removing non-voiceprint features from the first voiceprint data to be processed includes: The first interference segment and the second interference segment in the first voiceprint data to be processed are trimmed, and the remaining data segments after trimming are spliced to obtain the target voiceprint data; the first interference segment is a data segment whose sound intensity value does not reach the first sound intensity threshold, and the second interference segment is a data segment whose sound intensity value of the non-voiceprint feature reaches the second sound intensity threshold.
[0008] In some embodiments, the voiceprint data processing method further includes: Determining whether a sound intensity value of a target data segment reaches a third sound intensity threshold; the target data segment is a data segment located after a midpoint audio frame in the target voiceprint data; If not, the sound intensity value of the target data segment is increased.
[0009] In some embodiments, the voiceprint data processing method further includes: Inserting noise features into the target voiceprint data and adjusting the frequency and phase of the voiceprint features in the target voiceprint data to obtain enhanced voiceprint data; The target voiceprint data and the enhanced voiceprint data are stored in a preset storage space.
[0010] In some embodiments, before obtaining the voiceprint data to be recorded, the method further includes: In response to the received voiceprint recording operation, preset recording guidance data is broadcast; the recording guidance data includes corpus guidance information for prompting the corpus required for voiceprint recording and environment guidance information for prompting the environmental conditions required for the voiceprint recording site.
[0011] The present application also provides a voiceprint recognition method, including: Obtain the voiceprint data to be identified; Determining whether the duration of the voiceprint data to be identified reaches a preset threshold duration; If not, performing extension processing on the voiceprint data to be identified to obtain second extended voiceprint data; If so, the second voiceprint data to be processed and the target voiceprint data are compared, and the voiceprint recognition result of the voiceprint data to be identified is determined according to the data comparison result; when the duration of the voiceprint data to be identified does not reach the preset threshold duration, the second voiceprint data to be processed is the second extended voiceprint data, otherwise the second voiceprint data to be processed is the voiceprint data to be identified, and the target voiceprint data is constructed by the above-mentioned voiceprint data processing method.
[0012] An embodiment of the present application further provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the above method when executing the computer program.
[0013] An embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program implements the above method when executed by a processor.
[0014] The beneficial effects of the present application are as follows: During the voiceprint recording stage, when the length of the voiceprint data to be recorded is insufficient, the target voiceprint data that meets the requirements is generated through dual processing of data extension and feature elimination, so that the target voiceprint data used for voiceprint recognition and comparison meets the length requirements and avoids the mixing of invalid fragments. When the target voiceprint data is used for voiceprint recognition, the accuracy of voiceprint recognition can be improved. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] Figure 1 This is a diagram of the application environment of the voiceprint data processing method provided in an embodiment of the present application.
[0016] Figure 2 This is a flowchart of the voiceprint data processing method provided in an embodiment of the present application.
[0017] Figure 3 This is a flow chart of the voiceprint recognition method provided in an embodiment of the present application.
[0018] Figure 4 This is a schematic diagram of the hardware structure of the electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0019] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.
[0020] It should be noted that although the device schematics illustrate functional module divisions and the flowcharts illustrate logical sequences, in certain circumstances, the steps illustrated may be performed in a sequence that differs from the module divisions in the device or the sequence in the flowcharts. Terms such as "first" and "second" in the specification, claims, and drawings are used to distinguish similar items and are not intended to describe a specific sequence or precedence.
[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application pertains. The terms used herein are for the purpose of describing the embodiments of this application only and are not intended to limit this application.
[0022] The voiceprint data processing method provided in the embodiments of this application can be executed by a computer device, which can be a terminal device or a server. Terminal devices include, but are not limited to, mobile phones, computers, smart home appliances, vehicle-mounted terminals, aircraft, etc. The server can be a standalone physical server, a server cluster composed of multiple physical servers, a distributed system, or a cloud server.
[0023] In addition, the information, data, and signals involved in the embodiments of this application are authorized by the relevant objects or fully authorized by all parties, and the collection, use, and processing of relevant data comply with the relevant laws, regulations, and standards of relevant countries and regions.
[0024] Figure 1 This is a diagram of the application environment of the voiceprint data processing method provided in the embodiment of this application. Figure 1 The voiceprint data processing method is applied to a voiceprint data processing system. The voiceprint data processing system includes a terminal 110 and a server 120. The terminal 110 and the server 120 are connected via a network. The terminal 110 can be a desktop terminal or a mobile terminal. The mobile terminal can be at least one of a mobile phone, a tablet computer, a laptop computer, etc. The server 120 can be implemented as a standalone server or a server cluster consisting of multiple servers. The terminal 110 is used to send the voiceprint data to be recorded to the server 120. The server 120 is used to obtain the voiceprint data to be recorded and determine whether the duration of the voiceprint data to be recorded reaches a preset threshold duration. If not, the voiceprint data to be recorded is extended to obtain first extended voiceprint data. Non-voiceprint features in the first voiceprint data to be processed are removed to obtain target voiceprint data. If the duration of the voiceprint data to be recorded does not reach the preset threshold duration, the first voiceprint data to be processed is the first extended voiceprint data. Otherwise, the first voiceprint data to be processed is the voiceprint data to be recorded.
[0025] It should be understood that Figure 1 The application scenarios shown are only examples. In actual applications, the voiceprint data processing method provided by the embodiments of the present application can also be applied to other scenarios. For example, the above-mentioned voiceprint data processing method can be directly applied to the terminal 110, which is used to obtain the voiceprint data to be recorded, determine whether the duration of the voiceprint data to be recorded reaches a preset threshold duration, and if not, perform extension processing on the voiceprint data to be recorded to obtain first extended voiceprint data, eliminate non-voiceprint features in the first voiceprint data to be processed, and obtain target voiceprint data. Among them, when the duration of the voiceprint data to be recorded does not reach the preset threshold duration, the first voiceprint data to be processed is the first extended voiceprint data; otherwise, the first voiceprint data to be processed is the voiceprint data to be recorded.
[0026] To facilitate understanding of the voiceprint data processing method provided in the embodiments of the present application, the application scenario of the voiceprint data processing method is exemplarily introduced below, taking the execution subject as the terminal 110 as an example.
[0027] Figure 2 This is a flow chart of a voiceprint data processing method provided by an embodiment of the present application. Figure 2 In some embodiments, the method includes but is not limited to steps S201 to S204.
[0028] Step S201: Acquire the voiceprint data to be recorded.
[0029] The voiceprint data to be recorded refers to the raw voice data stream acquired through an audio acquisition device. Specifically, this can be achieved by recording with a microphone or parsing an audio file at the recording site. The execution entity obtains the corresponding voiceprint data to be recorded by parsing the audio file or acquiring the audio information recorded by the microphone.
[0030] At the recording site, a calibrated professional microphone can be used as the recording device. The microphone sampling frequency can be fixed at 16kHz and the quantization bit is set to 16 bits to ensure that the sound signal can be accurately captured. The distance between the microphone and the user should be maintained between 20-30 cm and at a 45-degree angle to obtain the best sound collection effect. Users are required to record their voiceprints in a specially designed low-noise environment. This environment can be treated with sound-absorbing materials to ensure that the background noise does not exceed 30 decibels. At the same time, clear signs should be set up around the recording site to remind people to keep quiet.
[0031] Step S202: determine whether the duration of the voiceprint data to be recorded reaches a preset threshold duration.
[0032] The preset threshold duration refers to the minimum time required for model training or recognition. For example, it is set to 2 seconds or 3 seconds, and the optimal value is determined through experiments.
[0033] If not, execute step S203; if so, execute step S204.
[0034] Step S203: performing extension processing on the input voiceprint data to obtain first extended voiceprint data.
[0035] Extension processing refers to extending short-term data to reach a preset duration. It can be achieved by copying original data segments or extracting valid segments and splicing them together.
[0036] Step S204: remove non-voiceprint features from the first voiceprint data to be processed to obtain target voiceprint data.
[0037] Non-voiceprint features refer to background noise, device noise, or other non-human voice components in voice data, which can be identified through sound intensity threshold judgment or spectrum analysis.
[0038] When the duration of the voiceprint data to be recorded does not reach the preset threshold duration, the first voiceprint data to be processed is the first extended voiceprint data; otherwise, the first voiceprint data to be processed is the voiceprint data to be recorded.
[0039] During voiceprint recording, the execution subject obtains the voiceprint data to be recorded in real time. When the duration of the voiceprint data to be recorded does not reach the preset threshold duration, the first extended voiceprint data that meets the duration requirement is generated by copying the original data or extracting the valid voiceprint segment. When the duration of the voiceprint data to be recorded does not reach the preset threshold duration, the first extended voiceprint data is analyzed. When the duration of the voiceprint data to be recorded reaches the preset threshold duration, the voiceprint data to be recorded is analyzed to identify and remove interference segments that do not meet the voiceprint characteristics, such as silence segments below the intensity threshold or segments containing burst noise, and finally obtain pure target voiceprint data. Therefore, through the dual processing of extension and feature removal, it is ensured that the length of the voiceprint data meets the requirements and invalid segments are avoided. For example, traditional silence filling will introduce featureless data segments, while the embodiment of the present application maintains the continuity of the voiceprint by splicing valid segments.
[0040] In some embodiments, extending the voiceprint data to be recorded includes: copying a plurality of voiceprint data to be recorded; and concatenating the voiceprint data to be recorded into voiceprint data whose duration reaches a preset threshold duration to obtain first extended voiceprint data.
[0041] Copying multiple copies of the voiceprint data to be recorded means making multiple complete copies of the voiceprint data to be recorded. This can be achieved by using an audio data replication algorithm to generate multiple copies of the same voiceprint data by repeating the original data. This operation can avoid the loss of voiceprint features due to data truncation or silence padding, ensuring that the extended data retains the complete original features.
[0042] Splicing voiceprint data that reaches a preset threshold length involves sequentially concatenating multiple copies of voiceprint data. This can be achieved using audio frame overlay or timeline expansion techniques, adjusting the total duration of the spliced data to meet the preset requirements. This operation can address the problem of the original voiceprint data being too short, preventing the recognition model from effectively extracting features, and provides sufficient valid data for subsequent processing.
[0043] When the duration of the voiceprint data to be recorded does not reach the preset threshold, a copy algorithm is first used to generate multiple copies that are exactly the same as the original data. For example, if the original data duration is 1 second and the preset threshold is 3 seconds, two copies of the original data are copied to obtain a splicing result with a total duration of 3 seconds. Subsequently, the copied data are connected in chronological order to form a continuous and complete voiceprint data stream. In this way, the extended data not only meets the duration requirement, but also fully retains the original voiceprint features, avoiding interference introduced by silent filling or random noise. Therefore, by copying the original data for splicing, the integrity of the voiceprint features is guaranteed, and additional interference is avoided, thereby improving the accuracy of subsequent voiceprint recognition. In addition, this method does not need to rely on complex data generation models, reducing computing resource consumption.
[0044] In some embodiments, the extended processing of the voiceprint data to be recorded includes: extracting voiceprint data segments containing voiceprint features from the voiceprint data to be recorded; splicing the voiceprint data to be recorded and several voiceprint data segments into voiceprint data with a duration reaching a preset threshold, to obtain first extended voiceprint data.
[0045] A voiceprint data segment refers to a continuous audio segment containing valid voiceprint features that is screened out from the voiceprint data to be recorded. It can be implemented by using an endpoint detection algorithm or a silence segmentation algorithm, and the starting and ending points of the valid voice segment are identified by setting an energy threshold or a spectral feature threshold.
[0046] During voiceprint data extension, a voiceprint feature analysis algorithm is first used to locate valid speech segments within the incoming voiceprint data. For example, short-time energy analysis or Mel-frequency cepstral coefficients are used to extract active voiceprint regions. The incoming voiceprint data and its associated voiceprint data segments are then replicated multiple times, and the number of segments required for splicing is calculated based on the preset duration requirement. Finally, linear interpolation is used to eliminate signal jumps at the splicing locations, resulting in a continuous first extended voiceprint data set that meets the duration requirement. This method prioritizes the reuse of valid voiceprint segments during data extension, avoiding the introduction of blank or noisy areas. Thus, by selectively reused valid voiceprint segments, the integrity of the voiceprint features is maintained while ensuring the duration of the extended data, effectively reducing the interference of invalid data on subsequent processing steps.
[0047] In some embodiments, removing non-voiceprint features from the first voiceprint data to be processed includes: trimming a first interference segment and a second interference segment from the first voiceprint data to be processed, and splicing the remaining data segments after trimming to obtain the target voiceprint data. The first interference segment is a data segment whose sound intensity value does not reach a first sound intensity threshold, and the second interference segment is a data segment whose sound intensity value of the non-voiceprint feature reaches a second sound intensity threshold.
[0048] The first interference segment refers to a segment in the first voiceprint data to be processed that cannot effectively represent the voiceprint features due to insufficient sound intensity. This can be achieved by setting a dynamic threshold to detect and locate the audio interval below the preset intensity value, for example, using a sliding window to calculate the short-time energy and comparing it with the threshold. The second interference segment refers to a segment in the first voiceprint data to be processed where the noise signal intensity of non-human voices exceeds an acceptable range. This can be achieved by using spectrum analysis combined with a voiceprint feature matching algorithm to identify non-human voice components, for example, using Mel-frequency cepstral coefficients to distinguish between human voices and background noise. The cropping operation is achieved by intercepting the boundary points of the interference interval marked on the timeline and deleting the corresponding data. The splicing operation uses linear interpolation or overlap retention method to connect valid segments.
[0049] After determining the first voiceprint data to be processed, a dual-threshold mechanism is used to separately process low-intensity invalid segments and high-intensity noise segments. The first voiceprint data to be processed is first framed, and the sound intensity value is calculated frame by frame. Frame sequences that are continuously below the first sound intensity threshold are marked as first interference segments. Simultaneously, non-voiceprint features are detected in real time. When the intensity of the non-voiceprint component exceeds the second sound intensity threshold, the corresponding frame sequence is marked as a second interference segment. After marking, both types of interference segments are batch-cropped, while valid voiceprint segments are retained and spliced according to the original time sequence to ensure the temporal coherence of the target voiceprint data. Thus, by separately processing the two types of interference using a dual-threshold mechanism, both feature extraction errors caused by retaining low-intensity invalid data and the interference of high-intensity non-voiceprint noise on the recognition model are avoided. Frame-level precise cropping, rather than global deletion, maximizes the integrity of valid voiceprint segments.
[0050] In some embodiments, the voiceprint data processing method further includes: determining whether the sound intensity value of the target data segment reaches a third sound intensity threshold; if not, increasing the sound intensity value of the target data segment. The target data segment is a data segment located after the midpoint audio frame in the target voiceprint data.
[0051] The target data segment refers to the portion of audio data located after the midpoint of the timeline in the target voiceprint data obtained after extension and cropping. This can be achieved by dividing the audio frame sequence and intercepting the second half of the continuous segment. This is used to detect whether the signal strength of the second half of the voiceprint data meets the recognition requirements. The third sound intensity threshold refers to the pre-set lower limit of intensity used to determine the effectiveness of the sound. It can be determined by experimentally analyzing the average intensity value of normal speech and is used to identify voiceprint feature attenuation caused by environmental interference or unstable pronunciation.
[0052] After the target voiceprint data is generated, the sound intensity of the data segment after the midpoint of the audio time is tested. If the sound intensity of this segment falls below the third sound intensity threshold (for example, 80 decibels), it indicates that the voiceprint feature is weakening in the second half of the utterance. Signal processing algorithms are then used to enhance the intensity of this segment, for example, by using dynamic gain adjustment to raise the sound intensity above the threshold while avoiding signal overload. This process ensures a uniform intensity distribution across the entire voiceprint data and eliminates localized feature loss caused by intermittent speech or environmental interference. By locating specific segments after the midpoint for targeted enhancement, this effectively addresses the weakening of voiceprint features caused by prolonged speech fatigue or accumulated environmental noise, thereby improving data integrity.
[0053] In some embodiments, the voiceprint data processing method further includes: inserting noise features into the target voiceprint data and adjusting the frequency and phase of the voiceprint features in the target voiceprint data to obtain enhanced voiceprint data; and storing the target voiceprint data and the enhanced voiceprint data in a preset storage space.
[0054] Inserting noise features refers to superimposing characteristic signals similar to the ambient background noise into the target voiceprint data. This can be achieved using a white noise generation algorithm or an ambient noise sample library matching algorithm. This is used to simulate noise interference in real-world scenarios and improve the voiceprint recognition model's adaptability to noisy environments. Adjusting frequency and phase refers to offsetting the fundamental frequency, harmonic distribution, and time series of the voiceprint features through digital signal processing technology. This can be achieved using a Fourier transform combined with a phase random perturbation algorithm to generate data samples with spectral differences.
[0055] Once the target voiceprint data is generated, a noise signal with a preset signal-to-noise ratio is first superimposed on the data using a noise feature insertion module. A frequency adjustment algorithm is used to randomly offset the voiceprint's fundamental frequency within a ±5% range, and a phase perturbation algorithm is used to randomly adjust the signal phase within a range of 0-2π. The processed enhanced voiceprint data is input into a storage module along with the original target data. Indexes are established based on voiceprint feature type, noise level, adjustment parameters, and other dimensions, and stored in a database. During the model training phase, the system dynamically calls voiceprint data combinations with different noise levels or spectral characteristics based on current training requirements. Thus, by actively introducing controllable noise and spectral perturbations, an enhanced dataset covering multiple scenarios is generated, giving the training data more comprehensive environmental adaptability and individual differences, effectively addressing the problem of insufficient model generalization due to data uniformity.
[0056] In some embodiments, before obtaining the voiceprint data to be recorded, it also includes: in response to the received voiceprint recording operation, broadcasting preset recording guidance data; the recording guidance data includes corpus guidance information for prompting the corpus required for voiceprint recording and environmental guidance information for prompting the environmental conditions required for the voiceprint recording site.
[0057] Recording guidance data refers to the guidance information of standardized recording requirements delivered to users through a preset interactive method. Specifically, it can be achieved by using a voice broadcast module to output pre-recorded audio instructions in real time, and is used to standardize the user's behavior during the voiceprint collection process. Corpus guidance information refers to the standardized requirements for the content of user voice utterances. Specifically, it can be achieved by using a pre-designed standard sentence library to randomly select aloud text fragments to ensure that the voiceprint data contains sufficient effective voice features. Environmental guidance information refers to the control requirements for the physical conditions of the recording scene. Specifically, it can be achieved by using acoustic sensors to monitor the environmental noise level in real time and trigger alarm prompts to reduce the impact of background interference on the quality of voiceprint data.
[0058] When the voiceprint collection device detects a user triggering a recording operation, it automatically activates the voice interaction system to play pre-stored guidance content. This guidance uses standardized sentence templates to guide users to speak according to a specified rhythm and content, for example, prompting them to repeat a specific number combination in a steady voice. Environmental guidance analyzes the noise spectrum characteristics of the microphone input signal in real time and triggers secondary prompts when it detects environmental interference exceeding a threshold, such as prompting the user to change the recording location when persistent low-frequency noise is detected. This process constrains user behavior through a structured guidance mechanism, ensuring that the collected voiceprint data meets preset standards for content integrity and environmental purity. For example, when a user says "Help me record my voiceprint," the device responds to the recording operation with the message "Please read the following text once: Cedar is a smart home assistant that can help you solve various problems in life," guiding the user to read the specific text aloud while displaying environmental guidance information: "Keep the microphone between 20-30 cm and the user at a 45-degree angle." Therefore, through the preset standardized guidance process, the standardization and consistency of voiceprint samples are controlled from the source of data collection, effectively avoiding the quality problems of original data caused by improper user operation or environmental interference.
[0059] Figure 3 This is a flow chart of the voiceprint recognition method provided by the embodiment of the present application. Figure 3 In some embodiments, the method includes but is not limited to steps S301 to S304.
[0060] Step S301: Acquire voiceprint data to be recognized.
[0061] Step S302: determine whether the duration of the voiceprint data to be identified reaches a preset threshold duration.
[0062] If not, execute step S303; if so, execute step S304.
[0063] Step S303: performing extension processing on the voiceprint data to be recognized to obtain second extended voiceprint data.
[0064] Step S304: performing a data comparison between the second voiceprint data to be processed and the target voiceprint data, and determining a voiceprint recognition result of the voiceprint data to be recognized according to the data comparison result.
[0065] When the duration of the voiceprint data to be identified does not reach the preset threshold duration, the second voiceprint data to be processed is the second extended voiceprint data; otherwise, the second voiceprint data to be processed is the voiceprint data to be identified.
[0066] The target voiceprint data is constructed by the voiceprint data processing method provided in the above embodiment.
[0067] During voiceprint recognition, the execution entity obtains the voiceprint data to be identified in real time. When the duration of the voiceprint data to be identified does not reach the preset threshold duration, a second extended voiceprint data that meets the preset duration requirement is generated by copying the original data fragments or extracting valid voiceprint fragments for splicing. This extended data is input into the voiceprint matching model together with the standardized target voiceprint data. By comparing the distribution differences of feature parameters such as Mel-frequency cepstral coefficients, the voiceprint similarity score is finally output as the recognition result. In this process, the extension processing ensures the integrity of the input data, and the target voiceprint data has a unified feature expression benchmark after preprocessing. Therefore, the extension processing generates data that meets the standard duration, allowing the feature extraction process to cover the complete fundamental frequency trajectory and resonance peak characteristics. At the same time, the target voiceprint data is processed by non-voiceprint feature elimination and intensity adjustment, effectively reducing the impact of device acquisition differences on the recognition results. Figure 4 It is a block diagram of an electronic device according to an exemplary embodiment.
[0068] Refer to the following Figure 4 4 to describe the electronic device 400 according to this embodiment of the present disclosure. Figure 4 The electronic device 400 shown is merely an example and should not limit the functionality and scope of use of the embodiments of the present disclosure.
[0069] like Figure 4 As shown, electronic device 400 is implemented as a general-purpose computing device. Components of electronic device 400 may include, but are not limited to, at least one processing unit 410, at least one storage unit 420, a bus 430 connecting various system components (including storage unit 420 and processing unit 410), a display unit 440, and the like.
[0070] The storage unit stores program codes, which can be executed by the processing unit 410 , so that the processing unit 410 executes the steps described in the above method part of this specification according to various exemplary embodiments of the present disclosure.
[0071] The storage unit 420 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 4201 and / or a cache memory unit 4202 , and may further include a read-only memory unit (ROM) 4203 .
[0072] The storage unit 420 may also include a program / utility 4204 having a set (at least one) of program modules 4205, such program modules 4205 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.
[0073] Bus 430 may represent one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.
[0074] The electronic device 400 may also communicate with one or more external devices 400′ (e.g., a keyboard, a pointing device, a Bluetooth device, etc.), one or more devices that enable a user to interact with the electronic device 400, and / or any device that enables the electronic device 400 to communicate with one or more other computing devices (e.g., a router, a modem, etc.). Such communication may occur via an input / output (I / O) interface 450. Furthermore, the electronic device 400 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network such as the Internet) via a network adapter 460. The network adapter 460 may communicate with other modules of the electronic device 400 via the bus 430. It should be understood that, although not shown in the figures, other hardware and / or software modules may be used in conjunction with the electronic device 400, including but not limited to microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.
[0075] An embodiment of the present application further provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above method is implemented.
[0076] The voiceprint data processing method, voiceprint recognition method, device and storage medium provided in the embodiments of the present application, during the voiceprint recording stage, when the length of the voiceprint data to be recorded is insufficient, generate target voiceprint data that meets the requirements through dual processing of data extension and feature elimination, so that the target voiceprint data used for voiceprint recognition and comparison meets the length requirements and avoids the mixing of invalid fragments. When the target voiceprint data is used for voiceprint recognition, the accuracy of voiceprint recognition can be improved.
[0077] Through the description of the above embodiments, it is easy for those skilled in the art to understand that the example embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solution according to the embodiments of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, or a network device, etc.) to execute the above-mentioned method according to the embodiments of the present disclosure.
[0078] The program product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or component, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.
[0079] Computer-readable storage media may include a data signal propagated in baseband or as part of a carrier wave, which carries readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The readable storage medium may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the readable storage medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination thereof.
[0080] Those skilled in the art will appreciate that the modules described above can be distributed in the device according to the description of the embodiment, or can be modified accordingly to be used in one or more devices that are different from the embodiment. The modules of the above embodiment can be combined into one module or further divided into multiple submodules.
[0081] While the exemplary embodiments of the present disclosure have been specifically illustrated and described above, it should be understood that the present disclosure is not limited to the detailed structures, configurations, or implementations described herein; rather, the present disclosure is intended to encompass various modifications and equivalent configurations within the spirit and scope of the appended claims.
Claims
1. A voiceprint data processing method, characterized in that: include: Obtain the voiceprint data to be recorded; Determining whether the duration of the voiceprint data to be recorded reaches a preset threshold duration; If not, performing extension processing on the voiceprint data to be recorded to obtain first extended voiceprint data; If so, the non-voiceprint features in the first voiceprint data to be processed are eliminated to obtain the target voiceprint data; when the duration of the voiceprint data to be recorded does not reach the preset threshold duration, the first voiceprint data to be processed is the first extended voiceprint data, otherwise the first voiceprint data to be processed is the voiceprint data to be recorded.
2. The voiceprint data processing method according to claim 1, characterized in that: The extending process of the voiceprint data to be recorded includes: Copy several copies of the voiceprint data to be recorded; The voiceprint data to be recorded are spliced into voiceprint data whose duration reaches the preset threshold duration to obtain the first extended voiceprint data.
3. The voiceprint data processing method according to claim 1, characterized in that: The extending process of the voiceprint data to be recorded includes: Extracting a voiceprint data segment containing a voiceprint feature from the voiceprint data to be recorded; The voiceprint data to be recorded and several voiceprint data segments are spliced together to form voiceprint data with a duration reaching the preset threshold duration, to obtain the first extended voiceprint data.
4. The voiceprint data processing method according to claim 1, characterized in that: The removing of non-voiceprint features from the first voiceprint data to be processed includes: The first interference segment and the second interference segment in the first voiceprint data to be processed are trimmed, and the remaining data segments after trimming are spliced to obtain the target voiceprint data; the first interference segment is a data segment whose sound intensity value does not reach the first sound intensity threshold, and the second interference segment is a data segment whose sound intensity value of the non-voiceprint feature reaches the second sound intensity threshold.
5. The voiceprint data processing method according to claim 1, characterized in that: Also includes: Determining whether a sound intensity value of a target data segment reaches a third sound intensity threshold; the target data segment is a data segment located after a midpoint audio frame in the target voiceprint data; If not, the sound intensity value of the target data segment is increased.
6. The voiceprint data processing method according to claim 1, characterized in that: Also includes: Inserting noise features into the target voiceprint data and adjusting the frequency and phase of the voiceprint features in the target voiceprint data to obtain enhanced voiceprint data; The target voiceprint data and the enhanced voiceprint data are stored in a preset storage space.
7. The voiceprint data processing method according to claim 1, characterized in that: Before obtaining the voiceprint data to be recorded, the method further includes: In response to the received voiceprint recording operation, preset recording guidance data is broadcast; the recording guidance data includes corpus guidance information for prompting the corpus required for voiceprint recording and environment guidance information for prompting the environmental conditions required for the voiceprint recording site.
8. A voiceprint recognition method, characterized in that: include: Obtain the voiceprint data to be identified; Determining whether the duration of the voiceprint data to be identified reaches a preset threshold duration; If not, performing extension processing on the voiceprint data to be identified to obtain second extended voiceprint data; If so, the second voiceprint data to be processed and the target voiceprint data are compared, and the voiceprint recognition result of the voiceprint data to be identified is determined according to the data comparison result; when the duration of the voiceprint data to be identified does not reach the preset threshold duration, the second voiceprint data to be processed is the second extended voiceprint data; otherwise, the second voiceprint data to be processed is the voiceprint data to be identified, and the target voiceprint data is constructed by the voiceprint data processing method described in any one of claims 1 to 7.
9. An electronic device, characterized in that: The electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the method according to any one of claims 1 to 8 when executing the computer program.
10. A computer-readable storage medium storing 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 8 is implemented.