CN111552832A - Risk user identification method and device based on voiceprint features and associated graph data - Google Patents

Abstract

The invention relates to an artificial intelligence technology, and discloses a risk user identification method based on voiceprint features and associated graph data, comprising: acquiring standard voice information of users; extracting first voiceprint features of the standard voice information; The voiceprint feature is input into the preset correlation map model to obtain the correlation map data related to the first voiceprint feature; the correlation map data is vectorized to obtain the correlation feature vector; if there is a relationship with the first voiceprint in the preset black voiceprint database A voiceprint feature that matches the feature, or a label feature vector matching the associated feature vector exists in the preset black relational graph, and the user is determined to be a risk user. The present invention also provides a risk user identification device, an electronic device, and a computer-readable storage medium based on voiceprint features and associated graph data. The present invention can reduce the missed detection rate of identifying risk users, and is beneficial to enhance the security of information.

Description

Risk user identification method and device based on voiceprint features and associated graph data

technical field

The present invention relates to the technical field of artificial intelligence, and in particular, to a risk user identification method, device, electronic device and computer-readable storage medium based on voiceprint features and associated graph data.

Background technique

At present, information data is increasing exponentially. With the increase of information data, it is necessary to verify the security of user information so as to identify potential risk users. In the prior art, user information is mainly verified based on a single-item verification method to identify risky users. This method has security loopholes, is easy to be missed, and user information is easily stolen by information thieves.

SUMMARY OF THE INVENTION

The present invention provides a method, device, electronic device and computer-readable storage medium for identifying risk users based on voiceprint features and associated graph data, the main purpose of which is to reduce the missed detection rate of identifying risk users, and is conducive to enhancing information security .

To achieve the above purpose, the present invention provides a method for identifying risk users based on voiceprint features and associated graph data, including:

Obtain the user's standard voice information;

extracting the first voiceprint feature of the standard voice information;

Inputting the first voiceprint feature into a preset association atlas model to obtain association atlas data related to the first voiceprint feature;

Vectorizing the association map data to obtain an association feature vector;

Judging whether there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library; and

Judging whether there is a label feature vector that matches the associated feature vector in the preset black relational graph;

If there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library, or if there is a label feature vector matching the associated feature vector in the preset black relationship map, determine The user is a risk user.

Optionally, the obtaining the standard voice information of the user includes:

obtain the original voice information of the user;

Utilize an analog/digital converter to sample the original voice information to obtain a digital voice signal;

performing a pre-emphasis operation on the digital voice signal to obtain a digital filtered voice signal;

A frame-by-frame windowing operation is performed on the digitally filtered voice signal to obtain the standard voice information.

Optionally, the described framing and windowing operation on the digitally filtered speech signal includes:

The frame-by-frame windowing operation is performed on the digitally filtered speech signal by an objective function, and the objective function is:

Wherein, n is the frame number sequence of the digitally filtered voice signal, N is the total number of frames of the digitally filtered voice signal, and w(n) is the single frame data of the standard voice information.

Optionally, the extracting the first voiceprint feature of the standard voice information includes:

The standard voice information is subjected to discrete Fourier transform to obtain spectrum information of the standard voice information;

The triangular filter is used to perform triangular filtering calculation on the standard speech information to obtain the frequency response value of the standard speech information;

Perform logarithmic calculation on the spectrum information and the frequency response value to obtain logarithmic energy;

Perform discrete cosine calculation on the logarithmic energy to obtain the first voiceprint feature.

Optionally, the calculation function included in the discrete Fourier transform is:

Wherein, N is the total number of frames of the digitally filtered voice signal, n is the frame number sequence of the digitally filtered voice signal, w(n) is the single frame data of the standard voice information, and j is the Fourier transform The weight of the transformation, k is the sound frequency of a single frame in the digitally filtered speech signal, and D is the spectrum information.

Optionally, the judging whether there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library includes:

Calculate the first similarity between the first voiceprint feature and the plurality of voiceprint features in the preset black voiceprint library by using the similarity function;

If there is a first similarity greater than a first similarity threshold, it is determined that there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library.

Optionally, the similarity function is:

Wherein, x represents the first voiceprint feature, _yi represents the voiceprint feature in the preset black voiceprint database, n represents the number of voiceprint features in the preset black voiceprint database, sim(x,y _i ) represents the first similarity.

In order to solve the above problems, the present invention also provides a risk user identification device based on voiceprint features and associated graph data, the device comprising:

A voice information acquisition module, used to acquire the user's standard voice information;

a voiceprint feature extraction module for extracting the first voiceprint feature of the standard voice information;

an atlas data acquisition module, configured to input the first voiceprint feature into a preset associated atlas model to obtain associated atlas data related to the first voiceprint feature;

A vector conversion module, for vectorizing the association map data to obtain an association feature vector;

a judgment module for judging whether there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library;

The judging module is also used to judge whether there is a label feature vector that matches the associated feature vector in the preset black relational graph;

A determination module, used for if there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library, or if there is a voiceprint feature matching the associated feature vector in the preset black relational map The label feature vector determines that the user is a risk user.

In order to solve the above problems, the present invention also provides an electronic device, the electronic device includes:

a memory that stores at least one instruction; and

The processor executes the instructions stored in the memory to implement any one of the above-mentioned methods for identifying risk users based on voiceprint features and associated graph data.

In order to solve the above problems, the present invention also provides a computer-readable storage medium, where at least one instruction is stored in the computer-readable storage medium, and the at least one instruction is executed by a processor in an electronic device to implement any one of the above The method for identifying risk users based on voiceprint features and associated graph data described in Item 1.

In the embodiment of the present invention, the standard voice information of the user is obtained; the first voiceprint feature of the standard voice information is extracted; feature-related association map data; vectorize the association map data to obtain an association feature vector; determine whether there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library; and determine the preset Whether there is a label feature vector matching the associated feature vector in the black relational map; if there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library, or the preset There is a label feature vector matching the associated feature vector in the black relationship graph, and the user is determined to be a risk user. Through the double-item verification of the two channels, the purpose of reducing the missed detection rate of identifying risk users and enhancing the security of information is achieved.

Description of drawings

1 is a schematic flowchart of a method for identifying risk users based on voiceprint features and associated graph data according to an embodiment of the present invention;

2 is a schematic block diagram of a device for identifying risk users based on voiceprint features and associated graph data according to an embodiment of the present invention;

3 is a schematic diagram of the internal structure of an electronic device for implementing a method for identifying risk users based on voiceprint features and associated graph data according to an embodiment of the present invention;

The realization, functional characteristics and advantages of the present invention will be further described with reference to the accompanying drawings in conjunction with the embodiments.

Detailed ways

It should be understood that the specific embodiments described herein are only used to explain the present invention, but not to limit the present invention.

The invention provides a risk user identification method based on voiceprint features and associated graph data. Referring to FIG. 1 , it is a schematic flowchart of a method for identifying risk users based on voiceprint features and associated graph data according to an embodiment of the present invention. The method may be performed by an apparatus, which may be implemented in software and/or hardware.

In this embodiment, the method for identifying risk users based on voiceprint features and associated graph data includes:

S1. Acquire the standard voice information of the user.

In this embodiment, the standard voice information of the user may be acquired from a voice database.

Further, the obtaining the standard voice information of the user includes:

obtain the original voice information of the user;

Utilize an analog/digital converter to sample the original voice information to obtain a digital voice signal;

performing a pre-emphasis operation on the digital voice signal to obtain a digital filtered voice signal;

A frame-by-frame windowing operation is performed on the digitally filtered voice signal to obtain the standard voice information.

In this embodiment, the user's original voice information is audio information including the user's voice, and the original voice may be voice information obtained during a voice call with the user.

For example, when a bank loan officer checks a loan user's phone credit, he/she obtains a voice recording of the conversation between the loan reviewer and the loan user, and the voice recording is the original voice information.

In detail, the purpose of sampling the original voice information is to convert the original voice information into a digital signal, so as to facilitate the processing of the voice information.

In this embodiment, an analog-to-digital converter is used to sample the original voice information at a rate of tens of thousands of times per second, and each sampling records the state of the original voice information at a certain time, so that the voice at different times can be obtained. digital voice signal.

Since the pronunciation system of the human voice will suppress the high-frequency part, in this embodiment, the pre-emphasis operation is performed in the above-mentioned manner to increase the energy of the high-frequency part, so that the speech energy of the high-frequency part and the speech energy of the low-frequency part have a similar amplitude , flatten the spectrum of the signal and keep the same signal-to-noise ratio in the entire frequency band from low frequency to high frequency.

In this embodiment, the pre-emphasis operation can compensate the digital speech signal.

Specifically, the pre-emphasis operation can be calculated by y(t)=x(t)-μx(t-1), where x(t) is the digital voice signal, t is time, and y(t) is the The digitally filtered speech signal, μ is the adjustment value of the pre-emphasis operation, and the value range of μ is [0.9, 1.0].

In this embodiment, the frame-by-frame windowing operation is to remove the overlapping part of the speech in the digitally filtered speech signal.

For example, when a bank lender conducts a telephone credit review of a loan user, since the original voice information will have overlapping parts of the voice of the bank lender and the loan user, the framed windowing operation can remove the bank lender's voice and retain the voice of the bank lender and the loan user. The voice of the loan user.

Further, the described framing and windowing operation to the digitally filtered speech signal includes:

The frame-by-frame windowing operation is performed on the digitally filtered speech signal by an objective function, and the objective function is:

Among them, n is the frame number sequence of the digitally filtered voice signal, N is the total number of frames of the digitally filtered voice signal, w(n) is the single frame data of the standard voice information, that is, w(n) represents the standard of each frame voice message.

S2. Extract the first voiceprint feature of the standard voice information.

In detail, extracting the first voiceprint feature of the standard voice information includes:

The standard voice information is subjected to discrete Fourier transform to obtain spectrum information of the standard voice information;

The triangular filter is used to perform triangular filtering calculation on the standard speech information to obtain the frequency response value of the standard speech information;

Perform logarithmic calculation on the spectrum information and the frequency response value to obtain logarithmic energy;

Perform discrete cosine calculation on the logarithmic energy to obtain the first voiceprint feature.

Preferably, the calculation function included in the discrete Fourier transform is:

Wherein, N is the total number of frames of the digitally filtered voice signal, n is the frame number sequence of the digitally filtered voice signal, w(n) is the single frame data of standard voice information, that is, w(n) represents each frame The standard speech information of , j is the weight of the Fourier transform, k is the sound frequency of a single frame in the digitally filtered speech signal, and D is the spectrum information.

Preferably, in this embodiment, a filter bank with M filters (the filters can be triangular filters) is defined, and the center frequency of the filters is f(i), i=1, 2, . . . , M , the center frequency is the cutoff frequency of the filter, and the triangular filter is used to calculate the triangular filter.

Since the triangular filter smoothes the spectrum and removes the effects of harmonics, the formants of the sound are highlighted. Therefore, the pitch or pitch of a piece of sound will not be reflected in the voiceprint feature, that is to say, the voiceprint feature will not be affected by the different pitches of the input sound and thus affect the recognition result.

Preferably, the triangular filtering is calculated as follows:

where f(i) is the center frequency of the triangular filter, i is the group of the triangular filter, H(k) is the frequency response value, and k is the sound frequency of a single frame in the digitally filtered speech signal, that is, k can represent The sound frequency for each frame.

Further, the logarithmic transformation is to calculate the logarithmic energy of each filter bank output.

The average person's response to sound pressure is logarithmic, and people are less sensitive to subtle changes in high sound pressure than low sound pressure. Therefore, using logarithms in this embodiment can reduce the sensitivity of the extracted features to changes in the input sound energy.

Specifically, the logarithmic calculation can be performed by the following formula:

Wherein i is the group of the triangular filter, k is the sound frequency of the single frame of the original voice information, N is the total number of frames of the digitally filtered voice signal, n is the frame number sequence of the digitally filtered voice signal, D is the spectral information, and S(i) is the logarithmic energy output by each filter.

Preferably, S(i) is subjected to discrete cosine transform to obtain voiceprint features, and the discrete cosine transform is as follows:

where n is the sequence of frame numbers of the original speech information, i is the group of triangular filters, M is the total number of groups of triangular filters, S(i) is the logarithmic energy output by each filter, and x is the sound pattern features.

Further, in another embodiment of the present invention, the first feature of extracting the standard voice information includes:

The first feature of the standard speech information is extracted by using an LSTM (Long Short-Term Memory, long short-term memory) network. The LSTM has three "gate" structures, namely, a forget gate, an input gate, and an output gate, which are used to process the input information differently. For the forget gate, as the name implies, part of the information passed through will be forgotten from the neural unit, so that part of the speech features of the previous frame will disappear during transmission, that is, it will no longer enter the next neural unit for training; the input The function of the gate is to add new useful information to the state of the neural unit, that is, after processing the newly learned speech features of this frame, add it to the transmitted information; finally, the output gate is based on the state of the neural unit above. and the processed information output, according to the output at the previous moment and the information to be output in the input at this moment, the output information at this moment is finally obtained as the first voiceprint feature.

S3. Inputting the first voiceprint feature into a preset correlation atlas model to obtain correlation atlas data related to the first voiceprint feature.

In this embodiment, the associated atlas data related to the first voiceprint feature may include, but is not limited to, user tag data corresponding to the first voiceprint feature, and dial records corresponding to the first voiceprint feature. Specifically, the user tag data includes attribute feature data of the user, such as gender, age, region, work data, and the like.

In detail, in this embodiment, the association map model can be constructed by using a convolutional neural network, using the sample voiceprint features as a training set, and using the sample voiceprint features marked by user label data as a label set for training to complete the association map. Model.

For example: inputting the first voiceprint feature of a user into a preset correlation atlas model to obtain the associated atlas data related to the first voiceprint feature, such as the user's information (name, gender, age, region, work, etc.), or the historical dialing time and times corresponding to the first voiceprint feature.

S4. Vectorize the association map data to obtain an association feature vector.

In detail, vectorization is done by the following expression:

Wherein, i represents the serial number of the correlation map data, v ⁱ represents the N-dimensional matrix vector of the correlation map data i, and v _j is the jth element of the N-dimensional matrix vector.

S5. Determine whether there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library; and determine whether there is a label feature vector matching the associated feature vector in the preset black relationship map.

In detail, judging whether there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library includes: calculating the first voiceprint feature and the preset black voiceprint library respectively through a similarity function. The first similarity of multiple voiceprint features; if there is a first similarity greater than the first similarity threshold, it is determined that there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library .

Or, judging whether there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library includes: comparing the first voiceprint feature with the voiceprint feature in the preset black voiceprint library for similarity Calculate to obtain a first similarity set, where the maximum value in the first similarity set is the first target similarity, and if the first target similarity is greater than the first similarity threshold, it is determined that the preset black voiceprint library exists A voiceprint feature matching the first voiceprint feature.

In this embodiment, the blacklist voiceprint database is a voiceprint database obtained by extracting voiceprint feature vectors of the voices of the blacklisted persons.

For example, the blacklist voiceprint database contains the voiceprint features of untrustworthy persons in banks and/or the voiceprint feature database of criminals in public security departments.

Further, the similarity function is:

Wherein, x represents the first voiceprint feature, _yi represents the voiceprint feature in the preset black voiceprint database, n represents the number of voiceprint features in the preset black voiceprint database, sim(x,y _i ) represents the first similarity.

Similarly, judging whether there is a label feature vector that matches the associated feature vector in the preset black relational graph: calculate the number of the associated feature vector and the plurality of label feature vectors in the preset black relational graph by using the similarity function respectively. Second similarity; if there is a second similarity greater than a second similarity threshold, it is determined that there is a voiceprint feature matching the associated feature vector in the preset black relational map.

Or, judging whether there is a label feature vector that matches the associated feature vector in the preset black relational graph: calculating the similarity between the associated feature vector and the label feature vector in the preset black relational graph to obtain a second similarity set, the maximum value in the second similarity set is the second target similarity, if the second target similarity is greater than the second similarity threshold, it is determined that there is a matching feature vector in the preset black relationship map that matches the associated feature vector The label feature vector of .

In this embodiment, the black relational graph database is obtained by extracting the label feature vector of the label data of the blacklisted persons. Therefore, the black relational graph database includes the labelled feature vector of the labelled data of the blacklisted persons.

In this embodiment, the second similarity threshold and the first similarity threshold may be the same or different, the second similarity threshold may be greater than the first similarity threshold, and the second similarity threshold may also be smaller than the first similarity threshold. For example, when the first similarity threshold is 80%, and the second similarity threshold is 90%.

S6. If there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library, or if there is a tag feature vector matching the associated feature vector in the preset black relationship map , and determine that the user is a risk user.

If there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint database, or there is a label feature vector matching the associated feature vector in the preset black relationship map, identifying the user as a risk user can be more comprehensive. Identify risky users and reduce the missed detection of risky users caused by single-item verification.

Further, if it is determined that the user is a risk user, a risk user reminder message is sent.

In the embodiment of the present invention, the standard voice information of the user is obtained; the first voiceprint feature of the standard voice information is extracted; feature-related association map data; vectorize the association map data to obtain an association feature vector; determine whether there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library; and determine the preset Whether there is a label feature vector matching the associated feature vector in the black relational map; if there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library, or the preset There is a label feature vector matching the associated feature vector in the black relationship graph, and the user is determined to be a risk user. Through the double-item verification of the two channels, the purpose of reducing the missed detection rate of identifying risk users and enhancing the security of information is achieved.

As shown in FIG. 2 , it is a functional block diagram of a risk user identification device based on voiceprint features and associated graph data according to the present invention.

The apparatus 100 for identifying risk users based on voiceprint features and associated graph data according to the present invention can be installed in an electronic device. According to the implemented functions, the risk user identification device based on voiceprint features and associated atlas data may include a voice information acquisition module 101, a voiceprint feature extraction module 102, a atlas data acquisition module 103, a vector conversion module 104, a judgment module 105, Determine module 106 . The modules described in the present invention can also be called units, which refer to a series of computer program segments that can be executed by the electronic device processor and can perform fixed functions, and are stored in the memory of the electronic device.

In this embodiment, the functions of each module/unit are as follows:

The voice information obtaining module 101 is used to obtain the standard voice information of the user.

In this embodiment, the standard voice information of the user may be acquired from a voice database.

Further, the obtaining the standard voice information of the user includes:

obtain the original voice information of the user;

Utilize an analog/digital converter to sample the original voice information to obtain a digital voice signal;

performing a pre-emphasis operation on the digital voice signal to obtain a digital filtered voice signal;

A frame-by-frame windowing operation is performed on the digitally filtered voice signal to obtain the standard voice information.

In this embodiment, the user's original voice information is audio information including the user's voice, and the original voice may be voice information obtained during a voice call with the user.

In detail, the purpose of sampling the original voice information is to convert the original voice information into a digital signal, so as to facilitate the processing of the voice information.

In this embodiment, an analog-to-digital converter is used to sample the original voice information at a rate of tens of thousands of times per second, and each sampling records the state of the original voice information at a certain time, so that the voice at different times can be obtained. digital voice signal.

Since the pronunciation system of the human voice will suppress the high-frequency part, in this embodiment, the pre-emphasis operation is performed in the above-mentioned manner to increase the energy of the high-frequency part, so that the speech energy of the high-frequency part and the speech energy of the low-frequency part have a similar amplitude , flatten the spectrum of the signal and keep the same signal-to-noise ratio in the entire frequency band from low frequency to high frequency.

In this embodiment, the pre-emphasis operation can compensate the digital speech signal.

Specifically, the pre-emphasis operation can be calculated by y(t)=x(t)-μx(t-1), where x(t) is the digital voice signal, t is time, and y(t) is the The digitally filtered speech signal, μ is the adjustment value of the pre-emphasis operation, and the value range of μ is [0.9, 1.0].

In this embodiment, the frame-by-frame windowing operation is to remove the overlapping part of the speech in the digitally filtered speech signal.

Further, the described framing and windowing operation to the digitally filtered speech signal includes:

The frame-by-frame windowing operation is performed on the digitally filtered speech signal by an objective function, and the objective function is:

Among them, n is the frame number sequence of the digitally filtered voice signal, N is the total number of frames of the digitally filtered voice signal, w(n) is the single frame data of the standard voice information, that is, w(n) represents the standard of each frame voice message.

The voiceprint feature extraction module 102 is configured to extract the first voiceprint feature of the standard speech information.

In detail, extracting the first voiceprint feature of the standard voice information includes:

The standard voice information is subjected to discrete Fourier transform to obtain spectrum information of the standard voice information;

The triangular filter is used to perform triangular filtering calculation on the standard speech information to obtain the frequency response value of the standard speech information;

Perform logarithmic calculation on the spectrum information and the frequency response value to obtain logarithmic energy;

Perform discrete cosine calculation on the logarithmic energy to obtain the first voiceprint feature.

Preferably, the calculation function included in the discrete Fourier transform is:

Wherein, N is the total number of frames of the digitally filtered voice signal, n is the frame number sequence of the digitally filtered voice signal, w(n) is the single frame data of standard voice information, that is, w(n) represents each frame The standard speech information of , j is the weight of the Fourier transform, k is the sound frequency of a single frame in the digitally filtered speech signal, and D is the spectrum information.

Preferably, in this embodiment, a filter bank with M filters (the filters can be triangular filters) is defined, and the center frequency of the filters is f(i), i=1, 2, . . . , M , the center frequency is the cutoff frequency of the filter, and the triangular filter is used to calculate the triangular filter.

Since the triangular filter smoothes the spectrum and removes the effects of harmonics, the formants of the sound are highlighted. Therefore, the pitch or pitch of a piece of sound will not be reflected in the voiceprint feature, that is to say, the voiceprint feature will not be affected by the different pitches of the input sound and thus affect the recognition result.

Preferably, the triangular filtering is calculated as follows:

where f(i) is the center frequency of the triangular filter, i is the group of the triangular filter, H(k) is the frequency response value, and k is the sound frequency of a single frame in the digitally filtered speech signal, that is, k can represent The sound frequency for each frame.

Further, the logarithmic transformation is to calculate the logarithmic energy of each filter bank output.

The average person's response to sound pressure is logarithmic, and people are less sensitive to subtle changes in high sound pressure than low sound pressure. Therefore, using logarithms in this embodiment can reduce the sensitivity of the extracted features to changes in the input sound energy.

Specifically, the logarithmic calculation can be performed by the following formula:

Wherein i is the group of the triangular filter, k is the sound frequency of the single frame of the original voice information, N is the total number of frames of the digitally filtered voice signal, n is the frame number sequence of the digitally filtered voice signal, D is the spectral information, and S(i) is the logarithmic energy output by each filter.

Preferably, S(i) is subjected to discrete cosine transform to obtain voiceprint features, and the discrete cosine transform is as follows:

where n is the sequence of frame numbers of the original speech information, i is the group of triangular filters, M is the total number of groups of triangular filters, S(i) is the logarithmic energy output by each filter, and x is the sound pattern features.

Further, in another embodiment of the present invention, the first feature of extracting the standard voice information includes:

The first feature of the standard speech information is extracted by using an LSTM (Long Short-Term Memory, long short-term memory) network. The LSTM has three "gate" structures, namely, a forget gate, an input gate, and an output gate, which are used to process the input information differently. For the forget gate, as the name implies, part of the information passed through will be forgotten from the neural unit, so that part of the speech features of the previous frame will disappear during transmission, that is, it will no longer enter the next neural unit for training; the input The function of the gate is to add new useful information to the state of the neural unit, that is, after processing the newly learned speech features of this frame, add it to the transmitted information; finally, the output gate is based on the state of the neural unit above. and the processed information output, according to the output at the previous moment and the information to be output in the input at this moment, the output information at this moment is finally obtained as the first voiceprint feature.

The atlas data acquisition module 103 is configured to input the first voiceprint feature into a preset associated atlas model to obtain associated atlas data related to the first voiceprint feature.

In this embodiment, the associated atlas data related to the first voiceprint feature may include, but is not limited to, user tag data corresponding to the first voiceprint feature, and dial records corresponding to the first voiceprint feature. Specifically, the user tag data includes attribute feature data of the user, such as gender, age, region, work data, and the like.

In detail, in this embodiment, the association map model can be constructed by using a convolutional neural network, using the sample voiceprint features as a training set, and using the sample voiceprint features marked by user label data as a label set for training to complete the association map. Model.

The vector transformation module 104 is configured to vectorize the association map data to obtain an association feature vector.

In detail, vectorization is done by the following expression:

Wherein, i represents the serial number of the correlation map data, v ⁱ represents the N-dimensional matrix vector of the correlation map data i, and v _j is the jth element of the N-dimensional matrix vector.

The judgment module 105 is used for judging whether there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library; The associated feature vector matches the label feature vector.

In detail, judging whether there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library includes: calculating the first voiceprint feature and the preset black voiceprint library respectively through a similarity function. The first similarity of multiple voiceprint features; if there is a first similarity greater than the first similarity threshold, it is determined that there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library .

Or, judging whether there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library includes: comparing the first voiceprint feature with the voiceprint feature in the preset black voiceprint library for similarity Calculate to obtain a first similarity set, where the maximum value in the first similarity set is the first target similarity, and if the first target similarity is greater than the first similarity threshold, it is determined that the preset black voiceprint library exists A voiceprint feature matching the first voiceprint feature.

In this embodiment, the blacklist voiceprint database is a voiceprint database obtained by extracting voiceprint feature vectors of the voices of the blacklisted persons.

For example, the blacklist voiceprint database contains the voiceprint features of untrustworthy persons in banks and/or the voiceprint feature database of criminals in public security departments.

Further, the similarity function is:

Wherein, x represents the first voiceprint feature, _yi represents the voiceprint feature in the preset black voiceprint database, n represents the number of voiceprint features in the preset black voiceprint database, sim(x,y _i ) represents the first similarity.

Similarly, judging whether there is a label feature vector that matches the associated feature vector in the preset black relational graph: calculate the number of the associated feature vector and the plurality of label feature vectors in the preset black relational graph through the similarity function respectively. Second similarity; if there is a second similarity greater than a second similarity threshold, it is determined that there is a voiceprint feature matching the associated feature vector in the preset black relational map.

Or, judging whether there is a label feature vector that matches the associated feature vector in the preset black relational graph: calculating the similarity between the associated feature vector and the label feature vector in the preset black relational graph to obtain a second similarity set, the maximum value in the second similarity set is the second target similarity, if the second target similarity is greater than the second similarity threshold, it is determined that there is a matching feature vector in the preset black relationship map that matches the associated feature vector The label feature vector of .

In this embodiment, the black relational graph database is obtained by extracting the label feature vector of the label data of the blacklisted persons. Therefore, the black relational graph database includes the labelled feature vector of the labelled data of the blacklisted persons.

The determining module 106 is configured to, if there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library, or if there is a voiceprint feature matching the associated feature vector in the preset black relationship map The label feature vector determines that the user is a risk user.

If there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint database, or there is a label feature vector matching the associated feature vector in the preset black relationship map, identifying the user as a risk user can be more comprehensive. And accurately identify risk users.

Further, if it is determined that the user is a risk user, a risk user reminder message is sent.

As shown in FIG. 3 , it is a schematic structural diagram of an electronic device implementing the method for identifying risk users based on voiceprint features and associated graph data according to the present invention.

The electronic device 1 may include a processor 10 , a memory 11 and a bus, and may also include a computer program stored in the memory 11 and executable on the processor 10 .

Wherein, the memory 11 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, mobile hard disk, multimedia card, card-type memory (for example: SD or DX memory, etc.), magnetic memory, magnetic disk, CD etc. The memory 11 may be an internal storage unit of the electronic device 1 in some embodiments, such as a mobile hard disk of the electronic device 1 . In other embodiments, the memory 11 may also be an external storage device of the electronic device 1, such as a pluggable mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, flash memory card (Flash Card), etc. Further, the memory 11 may also include both an internal storage unit of the electronic device 1 and an external storage device. The memory 11 can not only be used to store the application software and various data installed in the electronic device 1, such as the code of the risk user identification program based on the voiceprint feature and the associated map data, etc., but also can be used to temporarily store the output or data to be output.

In some embodiments, the processor 10 may be composed of integrated circuits, for example, may be composed of a single packaged integrated circuit, or may be composed of multiple integrated circuits packaged with the same function or different functions, including one or more integrated circuits. Central processing unit (Central Processing Unit, CPU), microprocessor, digital processing chip, graphics processor and combination of various control chips, etc. The processor 10 is the control core (Control Unit) of the electronic device, and uses various interfaces and lines to connect various components of the entire electronic device, and by running or executing the program or module (for example, executing the program) stored in the memory 11. Risk user identification program based on voiceprint features and associated graph data, etc.), and call the data stored in the memory 11 to execute various functions of the electronic device 1 and process data.

The bus may be a peripheral component interconnect (PCI for short) bus or an extended industry standard architecture (extended industry standard architecture, EISA for short) bus or the like. The bus can be divided into address bus, data bus, control bus and so on. The bus is configured to enable connection communication between the memory 11 and at least one processor 12 and the like.

FIG. 3 only shows an electronic device with components. Those skilled in the art can understand that the structure shown in FIG. 3 does not constitute a limitation on the electronic device 1, and may include fewer or more components than those shown in the figure. components, or a combination of certain components, or a different arrangement of components.

For example, although not shown, the electronic device 1 may also include a power supply (such as a battery) for powering the various components, preferably, the power supply may be logically connected to the at least one processor 10 through a power management device, so that the power management The device implements functions such as charge management, discharge management, and power consumption management. The power source may also include one or more DC or AC power sources, recharging devices, power failure detection circuits, power converters or inverters, power status indicators, and any other components. The electronic device 1 may further include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be repeated here.

Further, the electronic device 1 may also include a network interface, optionally, the network interface may include a wired interface and/or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is usually used in the electronic device 1 Establish a communication connection with other electronic devices.

Optionally, the electronic device 1 may further include a user interface, and the user interface may be a display (Display), an input unit (eg, a keyboard (Keyboard)), optionally, the user interface may also be a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode, organic light-emitting diode) touch device, and the like. The display may also be appropriately called a display screen or a display unit, which is used for displaying information processed in the electronic device 1 and for displaying a visualized user interface.

It should be understood that the embodiments are only used for illustration, and are not limited by this structure in the scope of the patent application.

The risk user identification program 12 based on voiceprint features and associated graph data stored in the memory 11 of the electronic device 1 is a combination of multiple instructions, and when running in the processor 10, can realize:

Obtain the user's standard voice information;

extracting the first voiceprint feature of the standard voice information;

Inputting the first voiceprint feature into a preset association atlas model to obtain association atlas data related to the first voiceprint feature;

Vectorizing the association map data to obtain an association feature vector;

Judging whether there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library; and

Judging whether there is a label feature vector that matches the associated feature vector in the preset black relational graph;

If there is a voiceprint feature matching the first voiceprint feature in the preset black voiceprint library, or if there is a label feature vector matching the associated feature vector in the preset black relationship map, determine The user is a risk user.

Specifically, for the specific implementation method of the above-mentioned instruction by the processor 10, reference may be made to the description of the relevant steps in the corresponding embodiment of FIG. 1, and details are not described herein.

Further, if the modules/units integrated in the electronic device 1 are implemented in the form of software functional units and sold or used as independent products, they may be stored in a computer-readable storage medium. The computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a U disk, a removable hard disk, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM, Read-Only Memory) .

In the several embodiments provided by the present invention, it should be understood that the disclosed apparatus, apparatus and method may be implemented in other manners. For example, the apparatus embodiments described above are only illustrative. For example, the division of the modules is only a logical function division, and there may be other division manners in actual implementation.

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

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

It will be apparent to those skilled in the art that the present invention is not limited to the details of the above-described exemplary embodiments, but that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics of the invention.

Therefore, the embodiments are to be regarded in all respects as illustrative and not restrictive, and the scope of the invention is to be defined by the appended claims rather than the foregoing description, which are therefore intended to fall within the scope of the claims. All changes within the meaning and range of the equivalents of , are included in the present invention. Any reference signs in the claims shall not be construed as limiting the involved claim.

Furthermore, it is clear that the word "comprising" does not exclude other units or steps and the singular does not exclude the plural. Several units or means recited in the system claims can also be realized by one unit or means by means of software or hardware. Second-class terms are used to denote names and do not denote any particular order.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be Modifications or equivalent substitutions can be made without departing from the spirit and scope of the technical solutions of the present invention.

Claims (10)

1. A risk user identification method based on voiceprint features and associated atlas data is characterized by comprising the following steps:

acquiring standard voice information of a user;

extracting a first voiceprint feature of the standard voice information;

inputting the first voiceprint feature into a preset association graph model to obtain association graph data related to the first voiceprint feature;

vectorizing the associated map data to obtain associated feature vectors;

judging whether a preset black voiceprint library has a voiceprint feature matched with the first voiceprint feature; and

judging whether a label feature vector matched with the associated feature vector exists in a preset black relation map;

and if the voiceprint features matched with the first voiceprint features exist in the preset black voiceprint library or the label feature vectors matched with the associated feature vectors exist in the preset black relation atlas, determining that the user is a risk user.

2. The method for identifying a risky user based on voiceprint features and associated atlas data according to claim 1, wherein the obtaining of the standard voice information of the user comprises:

acquiring original voice information of the user;

sampling the original voice information by using an analog/digital converter to obtain a digital voice signal;

carrying out pre-emphasis operation on the digital voice signal to obtain a digital filtering voice signal;

and performing frame division and windowing operation on the digital filtering voice signal to obtain the standard voice information.

3. The method of claim 2, wherein the framing windowing the digitally filtered speech signal comprises:

performing frame-division windowing operation on the digital filtering voice signal through an objective function, wherein the objective function is as follows:

wherein N is the frame number sequence of the digital filtering voice signal, N is the total frame number of the digital filtering voice signal, and w (N) is the single frame data of the standard voice information.

4. The method for identifying a risky user based on voiceprint features and associated atlas data according to claim 1, wherein the extracting the first voiceprint features of the standard speech information comprises:

performing discrete Fourier transform on the standard voice information to obtain frequency spectrum information of the standard voice information;

performing triangular filtering calculation on the standard voice information by using a triangular filter to obtain a frequency response value of the standard voice information;

carrying out logarithmic calculation on the frequency spectrum information and the frequency response value to obtain logarithmic energy;

and performing discrete cosine calculation on the logarithmic energy to obtain the first voiceprint characteristic.

5. The method for at risk user identification based on voiceprint features and associated atlas data of claim 4, wherein the discrete Fourier transform comprises a computational function of:

wherein, N is the total frame number of the digital filtering voice signal, N is the frame number sequence of the digital filtering voice signal, w (N) is the single frame data of the standard voice information, j is the weight of the fourier transform, k is the sound frequency of the single frame in the digital filtering voice signal, and D is the frequency spectrum information.

6. The method for identifying a risky user based on voiceprint features and associated atlas data according to any one of claims 1 to 3, wherein the judging whether the voiceprint features matched with the first voiceprint features exist in the preset black voiceprint library comprises:

respectively calculating first similarity of the first voiceprint features and a plurality of voiceprint features in a preset black voiceprint library through a similarity function;

and if the first similarity which is larger than the first similarity threshold exists, determining that the voiceprint features matched with the first voiceprint features exist in the preset black voiceprint library.

7. The method of claim 6, wherein the similarity function is:

wherein x represents the first voiceprint feature and y_iRepresenting the voiceprint features in the preset black voiceprint library, n representing the number of the voiceprint features in the preset black voiceprint library, sim (x, y)_i) Representing the first similarity.

8. An apparatus for identifying a risky user based on voiceprint features and associated atlas data, the apparatus comprising:

the voice information acquisition module is used for acquiring standard voice information of a user;

the voiceprint feature extraction module is used for extracting a first voiceprint feature of the standard voice information;

the spectrum data acquisition module is used for inputting the first voiceprint characteristics to a preset associated spectrum model to obtain associated spectrum data related to the first voiceprint characteristics;

the vector conversion module is used for vectorizing the associated map data to obtain associated feature vectors;

the judging module is used for judging whether the preset black voiceprint library has the voiceprint characteristics matched with the first voiceprint characteristics;

the judging module is also used for judging whether a label feature vector matched with the associated feature vector exists in a preset black relation map;

and the determining module is used for determining that the user is a risk user if the voiceprint features matched with the first voiceprint features exist in the preset black voiceprint library or the label feature vectors matched with the associated feature vectors exist in the preset black relation atlas.

9. An electronic device, characterized in that the electronic device comprises:

at least one processor; and the number of the first and second groups,

a memory communicatively coupled to the at least one processor; wherein,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform a method of at risk user identification based on voiceprint features and associated atlas data as claimed in any one of claims 1 to 7.

10. A computer-readable storage medium, storing a computer program, wherein the computer program, when executed by a processor, implements a method for at risk user identification based on voiceprint features and associated atlas data as recited in any one of claims 1 to 7.

Images