CN116206611A - Method and device for automatically constructing case-related voiceprint library - Google Patents

Abstract

The invention provides a method for automatically constructing a case-related voiceprint library, which comprises the following steps: extracting voice, namely acquiring and extracting a sound file stored in electronic equipment of related personnel; cutting the voice, namely cutting the extracted voice file into voice fragments by means of a voice recognition engine ASR; calculating a voice characteristic, namely calculating a voice frame characteristic vector by using a Mel Frequency Cepstrum Coefficient (MFCC) as an acoustic characteristic, and quantizing the voice print vector; performing voice clustering, performing PCA conversion to perform principal component analysis, selecting a K-means algorithm to perform intelligent clustering on voices, and extracting voice characteristics of all related persons; and constructing a voiceprint library, extracting voiceprint characteristics of the clustered voice files according to the voice clustering result, and establishing a normalized standard application library. By adding voiceprint information of involved persons into the person information database, criminal suspects can be quickly locked through a voiceprint automatic identification technology in the follow-up case detection process, the detection range is reduced to be extremely small, and the case detection efficiency is greatly improved.

Description

A method and device for automatically constructing a voiceprint database involved in a case

technical field

The invention belongs to the technical field of voiceprint technology, and in particular relates to a method and device for automatically constructing a voiceprint library involved in the case.

Background technique

With the increasing trend of Internet-related cases in recent years, the value of voiceprint features in the field of public security has become increasingly prominent. The voiceprint feature is one of the important biological features of the human body, and the voiceprint feature can uniquely determine the identity of the user. After the voiceprint library is completed, it can directly serve multiple types of police, which will effectively improve the efficiency and ability of relevant agencies to investigate and solve crimes. At present, the hit rate of the top ten in the million-level comparison of the voiceprint database of relevant agencies across the country has reached 90%, and the hit rate of the first place has reached 70%, and the comparison accuracy is still improving year by year. Therefore, electronic data exploration and evidence collection technology extracts voiceprint data from mobile phones, computers and other carriers, and uses voiceprint identification technology to identify individuals and groups involved in the voices involved in related cases, which can accurately identify the identity of specific personnel, and provide services for investigation and case litigation. Stronger evidence support.

However, domestic electronic data forensics products, foreign Encase forensic, Recon imager and other products with high market share have not yet had the function of extracting voice files from the data stored in electronic devices to automatically build a voiceprint library.

Moreover, the current voiceprint library relies more on manual collection when collecting voiceprints. After collection, it is necessary to manually fill in the ID number, name, gender and other information of the collected person, which makes batch collection of voiceprint data cumbersome and difficult. Time-consuming; in the actual voiceprint collection process, the quality of the voiceprint database is often uneven, the voiceprint information extraction process is not thoroughly recorded and archived, the extracted voiceprint data is chaotic, and the corresponding relationship cannot be established, resulting in a high recognition accuracy rate. The reduction cannot provide a data basis for the further application of the voiceprint library.

In view of this, it is very meaningful to propose a method and device for automatically constructing the voiceprint database involved in the case.

Contents of the invention

In order to solve the existing problems of cumbersome and time-consuming batch collection of voiceprints and the low quality of the voiceprint library, the present invention provides a method and device for automatically constructing the voiceprint library involved in the case to solve the above-mentioned technical defects.

In the first aspect, the present invention proposes a method for automatically constructing the voiceprint library involved in the case, which includes the following steps:

Voice extraction, obtain and extract the sound files stored in the electronic equipment of the relevant personnel, scan out the audio format files that meet the requirements in the equipment and save them as the voice data source of the voiceprint database of the relevant personnel;

Voice cutting, by means of the voice recognition engine ASR, the voice file extracted is divided into voice segments;

Calculate the speech features, use the Mel frequency cepstrum coefficient MFCC as the acoustic feature, calculate the speech frame feature vector and quantize the voiceprint vector;

Speech clustering, according to the calculated results of voiceprint vector quantization, PCA conversion is performed for principal component analysis, K-means algorithm is selected for intelligent clustering of voices, voice features of all relevant personnel are extracted, and voice fragments with the best quality are found and assembled into Voice files of specified length are used as clustering results;

Build a voiceprint library, extract the voiceprint features of the clustered voice files according to the results of voice clustering, and establish a standardized standard application library.

Preferably, in the voice extraction step, if there is an instant chat software, the chat voice is encrypted and saved in the chat file, and the voice data needs to be extracted separately after the chat file is decrypted first.

Further preferably, extracting the voice data separately after decrypting the chat file specifically includes:

The behavior of the software process is tracked through the dynamic taint analysis method, and the taint propagation path for obtaining the key behavior is constructed. The system automatically records the network data packets and protocol status of each link, and completes the protocol simulation environment by analyzing the data packets and protocols. build;

Analyze the memory structure of the software process, and obtain the current login token and other environment variables from the process space;

Set the token and variables in the protocol simulation environment, and start to simulate the process of obtaining the key from the server. After the key is successfully extracted, the local chat record file is decrypted and analyzed, and the voice data storage path is found;

Scan all files in the device, find out voice files to extract and record.

Further preferably, the step of calculating speech features specifically includes:

Preprocess the speech clips to reduce noise;

Add Hamming window to reduce Jibbs effect;

After the discrete Fourier transform is performed, the sequence triangular filter is used for filtering;

Carry out discrete cosine transform DCT to obtain the feature vector of the speech frame, and carry out vector quantization.

Further preferably, in the speech clustering step, the clustered speech may be intelligently segmented multiple times as required, and then re-clustered again until the desired result is achieved.

Further preferably, the main information of the voiceprint library includes: voiceprint unique ID identification value, storage path of voice clustering files, file format, voice duration, voice collection terminal information, voice source information, speaker information;

Among them, voice collection terminal information you look at mobile phone or computer, model, system information; voice source information includes telephone voice or instant messaging voice, specific application, phone number or IP, voice generation time; speaker information includes speaker UUID, ID card number, phone number.

In the second aspect, the embodiment of the present invention also provides a device for automatically constructing the voiceprint library involved in the case, which specifically includes:

The voice extraction module is used to obtain and extract the sound files stored in the electronic equipment of the relevant personnel, scan out the audio format files that meet the requirements in the equipment, and save them as the voice data source of the voiceprint database of the relevant personnel;

Voice cutting module, for cutting the extracted described voice file into voice segments by means of voice recognition engine ASR;

Calculating the voice feature module, for using the Mel frequency cepstral coefficient MFCC as the acoustic feature, calculating the voice frame feature vector and quantizing the voiceprint vector;

The voice clustering module is used to perform PCA conversion for principal component analysis based on the calculated voiceprint vector quantization results, select the K-means algorithm to intelligently cluster the voices, extract the voice features of all relevant personnel, and find the voice with the best quality The fragments are assembled into a voice file of a specified length as the clustering result;

Construct the voiceprint library module, which is used to extract the voiceprint features of the clustered voice files according to the results of voice clustering, and establish a standardized standard application library.

In a third aspect, an embodiment of the present invention provides an electronic device, including: one or more processors; a storage device for storing one or more programs, when one or more programs are executed by one or more processors , so that one or more processors implement the method described in any implementation manner of the first aspect.

In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the method described in any implementation manner in the first aspect is implemented.

Compared with the prior art, the beneficial results of the present invention are:

(1) By adding the voiceprint information of the persons involved in the case into the personnel information database, the criminal suspect can be quickly identified through the voiceprint automatic recognition technology in the follow-up case detection process, the scope of investigation is reduced to a very small size, and the detection efficiency of the case is greatly improved. The method has been applied in the Meiya Pico Excalibur reconnaissance and reconnaissance integration project.

(2) The method of the present invention adopts one-button operation to automatically and quickly extract the voice files in the device (such as confirming fraudulent calls, voice recordings of people suspected of fraudulent calls, etc.), and supports instant messaging software (comprising QQ, WeChat, etc.) , DingTalk, Line, FeiQ, WhatApps, etc.) to decrypt chat voices and extract source voice data, intelligently segment and cluster, and extract voiceprint features to build a voiceprint library, which is practical and easy to operate, and meets voiceprint recognition Work and the needs of building a database. The disadvantage of this method is that as users pay more and more attention to data security and the types of software such as instant messaging become more and more abundant, it will also require more efforts to extract encrypted chat voice files.

Description of drawings

The accompanying drawings are included to provide a further understanding of the embodiments and are incorporated in and constitute a part of this specification. The drawings illustrate the embodiments and together with the description serve to explain principles of the invention. Other embodiments and many intended advantages of the embodiments will readily be appreciated as they become better understood by reference to the following detailed description. The elements of the drawings are not necessarily to scale relative to each other. Like reference numerals designate corresponding similar parts.

Fig. 1 is an exemplary device architecture diagram in which an embodiment of the present invention can be applied;

Fig. 2 is a schematic flow chart of the method for automatically constructing the voiceprint library involved in the case according to an embodiment of the present invention;

Fig. 3 is a schematic flowchart of the construction of the voiceprint database in the method for automatically constructing the voiceprint database involved in the case according to the embodiment of the present invention;

Fig. 4 is a schematic flowchart of speech extraction in the method for automatically constructing the voiceprint library involved in the case according to an embodiment of the present invention;

5 is a schematic diagram of the MFCC calculation process in the method for automatically constructing the voiceprint library involved in the case according to an embodiment of the present invention;

Fig. 6 is a schematic flow diagram of the device for automatically constructing the voiceprint library involved in the case according to an embodiment of the present invention;

FIG. 7 is a schematic structural diagram of a computer device suitable for realizing the electronic equipment of the embodiment of the present invention.

Detailed ways

In the following detailed description, reference is made to the accompanying drawings, which form a part hereof, and which show by way of illustration specific embodiments in which the invention may be practiced. In this regard, directional terms such as "top", "bottom", "left", "right", "upper", "lower", etc. are used with reference to the orientation of the figures being described. Because components of an embodiment may be positioned in several different orientations, directional terminology is used for purposes of illustration and is by no means limiting. It is to be understood that other embodiments may be utilized or logic changes may be made without departing from the scope of the present invention. The following detailed description should therefore not be taken in a limiting sense, and the scope of the invention is defined by the appended claims.

It should be understood that the numbers of terminal devices, networks and servers in Fig. 1 are only illustrative. According to the implementation needs, there can be any number of terminal devices, networks and servers.

Fig. 1 shows an exemplary system architecture 100 to which the method for processing information or the device for processing information according to the embodiment of the present invention can be applied.

As shown in FIG. 1 , a system architecture 100 may include terminal devices 101 , 102 , 103 , a network 104 and a server 105 . The network 104 is used as a medium for providing communication links between the terminal devices 101 , 102 , 103 and the server 105 . Network 104 may include various connection types, such as wires, wireless communication links, or fiber optic cables, among others.

Users can use terminal devices 101 , 102 , 103 to interact with server 105 via network 104 to receive or send messages and the like. Various communication client applications can be installed on the terminal devices 101, 102, 103, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, and the like.

The terminal devices 101, 102, and 103 may be various electronic devices with communication functions, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, and the like.

The server 105 may be a server that provides various services, such as a background information processing server that processes verification request information sent by the terminal devices 101 , 102 , and 103 . The background information processing server can analyze and process the received verification request information, and obtain processing results (for example, verification success information used to indicate that the verification request is a legitimate request).

It should be noted that the method for processing information provided by the embodiment of the present invention is generally executed by the server 105 , and correspondingly, the device for processing information is generally disposed in the server 105 . In addition, the method for sending information provided by the embodiment of the present invention is generally executed by the terminal equipment 101 , 102 , 103 , and correspondingly, the apparatus for sending information is generally set in the terminal equipment 101 , 102 , 103 .

It should be noted that the server may be hardware or software. When the server is hardware, it can be implemented as a distributed server cluster composed of multiple servers, or as a single server. When the server is software, it can be implemented as multiple software or software modules (for example, to provide distributed services), or can be implemented as a single software or multiple software modules, which are not specifically limited here.

The current solution for building a voiceprint library has the following deficiencies:

(1) Batch collection of voiceprints is cumbersome and time-consuming: Public security organs have high-quality requirements for collected samples, and the input voiceprints must be effective and targeted to facilitate subsequent rapid voiceprint recognition. At present, when the voiceprint library collects voiceprints, it relies more on manual collection. After collection, it is necessary to manually fill in the ID number, name, gender and other information of the collected person, which makes batch collection of voiceprint data cumbersome and time-consuming. .

(2) The quality of the voiceprint database is not high: in the actual voiceprint collection process, the quality of the voiceprint database is often uneven, and the voiceprint information extraction process is not recorded and archived in detail, and the extracted voiceprint data is chaotic and cannot be constructed Correspondence, resulting in a reduction in recognition accuracy, unable to provide a data basis for the further application of the voiceprint library.

In order to effectively solve the above problems, we propose a method and device for automatically constructing the voiceprint database involved in the case. The specific improvements are as follows:

(1) One-click extraction of call recording files, communication voice files, etc. stored in the electronic equipment involved in the case, and automatically integrate all voice data as the source of voiceprint database construction. In addition, it can automatically decrypt the encrypted and stored local chat record files in the instant messaging software, and extract the voice data in the chat, which is more targeted.

(2) Automatically calculate the characteristic value of the voiceprint, establish a standard voiceprint library that meets the specifications, facilitate subsequent data connection with the back-end platform, and realize information sharing of criminal suspects and identity tracing of other participants in the incident. For example, the system can search, inquire, and return all call records that the fraudster participated in through the voiceprint database engine, reversely trace all the call behaviors related to the fraudulent call path, and help the public security organs find the real fraud mastermind.

This solution mainly provides functions such as one-key extraction of voice, intelligent voice clustering, and automatic construction of voiceprint database, etc., which are used to realize one-key extraction of voice data of persons involved in the case and automatically build voiceprint database. This method mainly realizes the process of automatically constructing the voiceprint library for the original data of the electronic evidence involved in the case.

Fig. 2 shows that the embodiment of the present invention discloses a method for automatically constructing the voiceprint library involved in the case, as shown in Fig. 2 and Fig. 3, the method includes the following steps:

S1, voice extraction, obtain and extract the sound files stored in the electronic equipment of the relevant personnel, scan out the audio format files that meet the requirements in the equipment and save them as the voice data source of the voiceprint database of the relevant personnel;

In a specific embodiment, one-key voice extraction: relying on the data extraction capability of electronic forensics, the voices stored in the electronic devices (including Android, IOS operating system smartphones, tablet devices, notebook computers, and desktop computers) of the persons involved in the case are automatically and comprehensively extracted. The files are effectively extracted, and the required audio format files (such as wav, mp3, amr, silk, m4a, acc, etc.) in the device are scanned and saved as the voice data source of the voiceprint database of the people involved in the case.

Furthermore, if there is an instant chat software, the chat voice may be encrypted and stored in the chat file. In this case, it is necessary to decrypt the chat file first and then extract the voice data separately. The specific process is shown in Figure 4.

Specifically, the voice data of instant messaging software exists as an attachment. As application manufacturers continue to strengthen data protection, storage paths are becoming more difficult to find, and many local data are encrypted. The software also needs to obtain the login key to decrypt and analyze the local chat records, and extract the voice files in the chat.

①Track the behavior of the software process through the dynamic taint analysis method, and construct the taint propagation path of the behavior of obtaining the key. The system automatically records the network data packets and protocol status of each link, and completes the protocol simulation environment through the analysis of the data packets and protocols construction;

② Analyze the memory structure of the software process, and obtain the current login token and other environment variables from the process space;

③Set the token and variables into the protocol simulation environment, and start to simulate the process of obtaining the key from the server. After the key is successfully extracted, the local chat record file is decrypted and analyzed, and the voice data storage path is found;

④Scan all the files in the device, find out the voice files, extract and record them.

S2, voice cutting, cutting the extracted voice file into voice segments by means of the voice recognition engine ASR;

Specifically, voice intelligent cutting: the extracted voice file is divided into voice segments by means of a voice recognition engine (ASR). After the ASR engine processes the audio file, it will get the speech recognition text and the corresponding start and end time of the text. According to this time information, a section of audio file can be easily cut. What needs to be emphasized here is that we do not use the text results of speech recognition, but only need to know the boundary information of each recognized text. Even if there is an error in the recognition, it will not have any impact on the construction of the voiceprint library, because we always use It is all about the acoustic information of the sound, and does not care about the text information corresponding to this sound.

S3. Calculating speech features, using the Mel frequency cepstral coefficient MFCC as the acoustic feature, calculating the speech frame feature vector and quantizing the voiceprint vector;

Specifically, as shown in Fig. 5, we choose the Mel-frequency cepstral coefficient (MFCC) as the acoustic feature. MFCC is a cepstrum parameter extracted in the Mel scale frequency domain, which can be used to characterize the acoustic characteristics of speaker information, and can effectively identify the speaker of a certain audio.

S4. Speech clustering. According to the calculated voiceprint vector quantization results, perform PCA conversion for principal component analysis, select the K-means algorithm to intelligently cluster the voices, extract the voice features of all relevant personnel, and find the best quality voice clips Assemble into a voice file of specified length as the clustering result;

In a specific embodiment, according to the result of the vector quantization of the above-mentioned voiceprints, PCA conversion is performed and principal component analysis is performed, and then the K-means algorithm is selected to intelligently cluster the voices, so as to extract the voice features of all persons involved in the case as much as possible, and find out the voice features with the best quality. The speech fragments of the group are assembled into a speech file of a specified length as the clustering result. In order to improve the accuracy of speech feature extraction and clustering, the clustered speech can also be intelligently segmented multiple times according to needs, and then re-clustered again, and it ends when it reaches expectations.

S5. Construct a voiceprint library, extract the voiceprint features of the clustered voice files according to the result of voice clustering, and establish a standardized standard application library.

Specifically, build a voiceprint library: According to the results of voice clustering, extract the voiceprint features of the clustered voice files, and establish a standardized standard application library. The main information of the voiceprint library includes: voiceprint unique ID value, voice clustering Class file storage path, file format, voice duration, voice collection terminal information (mobile phone/computer, model, system information, etc.), voice source information (telephone voice/instant messaging voice, etc., specific application, phone number or IP, voice generation Time), speaker information (speaker UUID, ID number, phone number, etc.).

In the second aspect, the embodiment of the present invention also discloses a device for automatically constructing the voiceprint library involved in the case. As shown in Figure 6, the device specifically includes: a voice extraction module 61, a voice cutting module 62, a voice feature calculation module 63, A clustering module 64 and a module 65 for constructing a voiceprint library.

In a specific embodiment, the voice extraction module 61 is used to obtain and extract the sound files stored in the electronic equipment of the relevant personnel, scan out the audio format files that meet the requirements in the equipment, and save the voice data as the voiceprint library of the relevant personnel Source; voice cutting module 62, for cutting the extracted voice file into voice segments by means of voice recognition engine ASR; calculating voice feature module 63, for utilizing Mel frequency cepstral coefficient MFCC as acoustic features, calculated to obtain Speech frame feature vector and voiceprint vector quantization; Speech clustering module 64, for the voiceprint vector quantization result obtained according to calculation, carry out PCA transformation and carry out principal component analysis, select K-means algorithm to carry out intelligent clustering to speech, extract all According to the voice characteristics of relevant personnel, find the voice segment with the best quality and assemble it into a voice file of specified length as the clustering result; build the voiceprint library module 65, which is used to extract the voiceprint of the clustered voice file according to the voice clustering result Features, establish a standardized standard application library.

The technical solution of the present invention provides a method and device for automatically constructing the voiceprint library involved in the case. The method adopts one-button operation to automatically and quickly extract the voice files in the device (such as confirming fraudulent calls, voice recordings of suspected fraudulent callers, etc.) At the same time, it supports decrypting the chat voice in the current mainstream instant messaging software (including QQ, WeChat, DingTalk, Line, FeiQ, WhatApps, etc.), extracting source voice data, intelligently segmenting and clustering, and extracting voiceprint features to construct voice Texture database, with strong practicability and simple operation, meets the needs of voiceprint recognition and database construction. The disadvantage of this method is that as users pay more and more attention to data security and the types of software such as instant messaging become more and more abundant, it will also require more efforts to extract encrypted chat voice files.

The present invention adds the voiceprint information of the persons involved in the case into the personnel information database, and in the subsequent case detection process, the criminal suspect can be quickly locked through the voiceprint automatic recognition technology, the investigation scope is reduced to a very small size, and the detection efficiency of the case is greatly improved. Applied in Meiya Pico Excalibur reconnaissance integration project.

Referring now to FIG. 7 , it shows a schematic structural diagram of a computer device 700 suitable for implementing an electronic device (such as a server or a terminal device shown in FIG. 1 ) in an embodiment of the present invention. The electronic device shown in FIG. 7 is only an example, and should not limit the functions and scope of use of this embodiment of the present invention.

As shown in FIG. 7, a computer device 700 includes a central processing unit (CPU) 701 and a graphics processing unit (GPU) 702, which can be randomly accessed according to a program stored in a read-only memory (ROM) 703 or loaded from a storage section 709. Various appropriate actions and processes are executed by programs in the memory (RAM) 706 . In the RAM 704, various programs and data necessary for the operation of the device 700 are also stored. The CPU 701 , GPU 702 , ROM 703 , and RAM 704 are connected to each other via a bus 705 . An input/output (I/O) interface 706 is also connected to the bus 705 .

The following components are connected to the I/O interface 706: an input section 707 including a keyboard, a mouse, etc.; an output section 708 including, for example, a liquid crystal display (LCD), etc., and a speaker; a storage section 709 including a hard disk, etc.; The communication part 710 of a network interface card such as a modem or the like. The communication section 710 performs communication processing via a network such as the Internet. A drive 711 can also be connected to the I/O interface 706 as needed. A removable medium 712 such as a magnetic disk, optical disk, magneto-optical disk, semiconductor memory, etc. is mounted on the drive 711 as necessary so that a computer program read therefrom is installed into the storage section 709 as necessary.

In particular, according to the disclosed embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the disclosed embodiments of the present invention include a computer program product, which includes a computer program carried on a computer-readable medium, where the computer program includes program codes for executing the methods shown in the flowcharts. In such an embodiment, the computer program may be downloaded and installed from a network via communication portion 710 and/or installed from removable media 712 . When the computer program is executed by a central processing unit (CPU) 701 and a graphics processing unit (GPU) 702, the above-mentioned functions defined in the method of the present invention are performed.

It should be noted that the computer-readable medium in the present invention may be a computer-readable signal medium or a computer-readable medium or any combination of the above two. A computer readable medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor device, device, or device, or a combination of any of the above. More specific examples of computer readable media may include, but are not limited to, electrical connections with one or more conductors, portable computer diskettes, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable Read only memory (EPROM or flash memory), optical fiber, portable compact disk read only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution device, device, or device. In the present invention, however, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, in which computer-readable program codes are carried. Such propagated data signals may take many forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution apparatus, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

Computer program code for carrying out the operations of the present invention can be written in one or more programming languages, or combinations thereof, including object-oriented programming languages—such as Java, Smalltalk, C++, and conventional A procedural programming language—such as "C" or a similar programming language. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In cases involving a remote computer, the remote computer can be connected to the user computer through any kind of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (such as through an Internet service provider). Internet connection).

The flowchart and block diagrams in the figures illustrate the architecture, functionality and operation of possible implementations of apparatuses, methods and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, program segment, or portion of code that contains one or more logical functions for implementing specified executable instructions. It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or they may sometimes be executed in the reverse order, depending upon the functionality involved. It should also be noted that each block in the block diagrams and/or flowcharts, and combinations of blocks in the block diagrams and/or flowcharts, can be implemented by a dedicated hardware-based device that performs the specified functions or operations , or may be implemented by a combination of dedicated hardware and computer instructions.

The modules involved in the embodiments described in the present invention may be implemented by software or by hardware. The described modules may also be provided in a processor.

As another aspect, the present invention also provides a computer-readable medium. The computer-readable medium may be included in the electronic device described in the above embodiments; or it may exist independently without being assembled into the electronic device. middle. The above-mentioned computer-readable medium carries one or more programs, and when the above-mentioned one or more programs are executed by the electronic device, the electronic device: obtains and extracts the sound files stored in the electronic device of the relevant person, scans out the device The audio format files that meet the requirements are saved as the voice data source of the voiceprint library of relevant personnel; the extracted voice files are divided into voice segments by means of the voice recognition engine ASR; the Mel frequency cepstral coefficient MFCC is used as the acoustic feature , calculate the feature vector of the voice frame and quantize the voiceprint vector; according to the calculated voiceprint vector quantization result, perform PCA conversion for principal component analysis, select the K-means algorithm to intelligently cluster the voice, and extract the voice features of all relevant personnel , find the best-quality speech fragments and assemble them into a speech file of a specified length as the clustering result; according to the speech clustering result, extract the voiceprint features of the clustered speech file, and establish a standardized standard application library.

The above description is only a preferred embodiment of the present invention and an illustration of the applied technical principle. It should be understood by those skilled in the art that the scope of the invention involved in the present invention is not limited to the technical solution formed by the specific combination of the above-mentioned technical features, and should also cover the technical solutions formed by the above-mentioned technical features or Other technical solutions formed by any combination of equivalent features. For example, a technical solution formed by replacing the above-mentioned features with technical features disclosed in the present invention (but not limited to) having similar functions.

Claims (9)

1. A method for automatically constructing the voiceprint library involved in the case, characterized in that the method comprises the following steps: Voice extraction, obtain and extract the sound files stored in the electronic equipment of the relevant personnel, scan out the audio format files that meet the requirements in the equipment and save them as the voice data source of the voiceprint database of the relevant personnel; Voice cutting, by means of the voice recognition engine ASR, the voice file extracted is divided into voice segments; Calculate the speech features, use the Mel frequency cepstrum coefficient MFCC as the acoustic feature, calculate the speech frame feature vector and quantize the voiceprint vector; Speech clustering, according to the calculated results of voiceprint vector quantization, PCA conversion is performed for principal component analysis, K-means algorithm is selected for intelligent clustering of voices, voice features of all relevant personnel are extracted, and voice fragments with the best quality are found and assembled into Voice files of specified length are used as clustering results; Build a voiceprint library, extract the voiceprint features of the clustered voice files according to the results of voice clustering, and establish a standardized standard application library. 2. The method for automatically constructing the voiceprint library involved in the case according to claim 1, characterized in that, in the voice extraction step, if there is instant chat software, the chat voice is encrypted and stored in the chat file, and the chat file needs to be decrypted first and then Speech data is extracted separately. 3. The method for automatically constructing the voiceprint library involved in the case according to claim 2, wherein the separate extraction of the voice data after decrypting the chat file specifically includes: The behavior of the software process is tracked through the dynamic taint analysis method, and the taint propagation path for obtaining the key behavior is constructed. The system automatically records the network data packets and protocol status of each link, and completes the protocol simulation environment by analyzing the data packets and protocols. build; Analyze the memory structure of the software process, and obtain the current login token and other environment variables from the process space; Set the token and variables in the protocol simulation environment, and start to simulate the process of obtaining the key from the server. After the key is successfully extracted, the local chat record file is decrypted and analyzed, and the voice data storage path is found; Scan all files in the device, find out voice files to extract and record. 4. The method for automatically constructing the voiceprint library involved in the case according to claim 3, wherein the step of calculating voice features specifically includes: Preprocess the speech clips to reduce noise; Add Hamming window to reduce Jibbs effect; After the discrete Fourier transform is performed, the sequence triangular filter is used for filtering; Carry out discrete cosine transform DCT to obtain the feature vector of the speech frame, and carry out vector quantization. 5. The method for automatically constructing the voiceprint library involved in the case according to claim 4, characterized in that, in the voice clustering step, the voice after clustering can also be intelligently segmented multiple times as required and then re-clustered again, until it ends as expected. 6. The method for automatically constructing the voiceprint library involved in the case according to claim 5, wherein the main information of the voiceprint library includes: the unique ID value of the voiceprint, the storage path of the voice clustering file, the file format, and the voice duration , voice collection terminal information, voice source information, speaker information; Among them, voice collection terminal information you look at mobile phone or computer, model, system information; voice source information includes telephone voice or instant messaging voice, specific application, phone number or IP, voice generation time; speaker information includes speaker UUID, ID card number, phone number. 7. A device for automatically constructing the voiceprint library involved in the case, characterized in that the device specifically includes: The voice extraction module is used to obtain and extract the sound files stored in the electronic equipment of the relevant personnel, scan out the audio format files that meet the requirements in the equipment, and save them as the voice data source of the voiceprint database of the relevant personnel; Voice cutting module, for cutting the extracted described voice file into voice segments by means of voice recognition engine ASR; Calculating the voice feature module, for using the Mel frequency cepstral coefficient MFCC as the acoustic feature, calculating the voice frame feature vector and quantizing the voiceprint vector; The voice clustering module is used to perform PCA conversion for principal component analysis based on the calculated voiceprint vector quantization results, select the K-means algorithm to intelligently cluster the voices, extract the voice features of all relevant personnel, and find the voice with the best quality The fragments are assembled into a voice file of a specified length as the clustering result; Construct the voiceprint library module, which is used to extract the voiceprint features of the clustered voice files according to the results of voice clustering, and establish a standardized standard application library. 8. An electronic device comprising: one or more processors; storage means 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 are made to implement the method according to any one of claims 1-6. 9. A computer-readable storage medium, on which a computer program is stored, wherein, when the program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images