CN114783444A

CN114783444A - Voiceprint recognition method and device, storage medium and electronic equipment

Info

Publication number: CN114783444A
Application number: CN202210487519.1A
Authority: CN
Inventors: 张宇
Original assignee: Beijing Minglue Zhaohui Technology Co Ltd
Current assignee: Beijing Minglue Zhaohui Technology Co Ltd
Priority date: 2022-05-06
Filing date: 2022-05-06
Publication date: 2022-07-22

Abstract

The invention discloses a voiceprint recognition method and device, a storage medium and electronic equipment. The method comprises the following steps: acquiring a target voice recording of a client; performing voiceprint extraction operation on the target voice recording to obtain a first target voiceprint; comparing the first target voiceprint with the voiceprint database to obtain a first comparison result; under the condition that the first comparison result is that the comparison is successful, marking the client as a risk client, and sending risk early warning information to the user; and under the condition that the first comparison result is that the comparison is not successful, marking the client as a third-party client, and storing the first target voiceprint into a voiceprint database. The invention solves the technical problem of the sharp increase of the marketing expense caused by unclear division of the client attribution.

Description

Voiceprint recognition method and device, storage medium and electronic equipment

Technical Field

The invention relates to the field of computers, in particular to a voiceprint recognition method and device, a storage medium and electronic equipment.

Background

Today, it is an unmet fact that real estate has shifted from the past era of growth to the era of inventory. The change of the external environment of the whole real estate requires the real estate developer to change roles, and the former developer is changed into an operator. A very common scenario problem ensues: "flyer bill". The 'flyer' is a business consultant in the case and a house property intermediary outside the case, and the marketing expense is increased greatly through the modes of single plug-in the hidden guest, mutual introduction of the hidden guest and the like or due to unclear division of the affiliation of the guest. In the prior art, a mode of capturing a face by a camera is adopted, and during signing, the identity card is scanned or the face is captured again, and the face is compared with the stored face to judge whether to fly the bill. But because the adopted camera collects the face and stores the face, the privacy of the person is seriously invaded.

Disclosure of Invention

The embodiment of the invention provides a voiceprint recognition method, a voiceprint recognition device, a storage medium and electronic equipment, and aims to at least solve the technical problem of marketing expense sharp increase caused by unclear client attribution division.

According to an aspect of an embodiment of the present invention, there is provided a voiceprint recognition method, including: acquiring a target voice recording of a client; performing voiceprint extraction operation on the target voice recording to obtain a first target voiceprint; comparing the first target voiceprint with a voiceprint database to obtain a first comparison result; under the condition that the first comparison result is that the comparison is successful, marking the client as a risk client and sending risk early warning information to the user; and under the condition that the first comparison result is that the comparison is not successful, marking the client as a third-party client, and storing the first target voiceprint to the voiceprint database.

According to another aspect of the embodiments of the present invention, there is provided a voiceprint recognition apparatus including: the first acquisition module is used for acquiring a target voice record of a client; the extraction module is used for carrying out voiceprint extraction operation on the target voice record to obtain a first target voiceprint; the first comparison module is used for comparing the first target voiceprint with the voiceprint database to obtain a first comparison result; the first marking module is used for marking the client as a risk client under the condition that the first comparison result is that the comparison is successful, and sending risk early warning information to the user; and the second marking module is used for marking the client as a third-party client under the condition that the first comparison result is that the comparison is not successful, and storing the first target voiceprint to the voiceprint database.

As an optional example, the obtaining module includes: the first acquisition unit is used for acquiring a first voice record of communication among the user, the client and a third party; the detection unit is used for carrying out voice quality detection operation on the first voice record to obtain a second voice record with qualified quality; and the separation unit is used for carrying out voice separation operation on the second voice record to obtain a third voice record of the user, a target voice record of the client and a fourth voice record of the third party.

As an alternative example, the detection unit includes: the detection subunit is used for carrying out voice quality detection operation on the first voice record to obtain a detection result; a deleting subunit, configured to delete the first voice record if the detection result is that the quality is not acceptable; and the reserving subunit is used for reserving the first voice record under the condition that the detection result is qualified.

As an optional example, the extracting module includes a first processing unit, configured to perform a denoising technique and a valid tone extracting operation on the target voice record to obtain a fifth voice record; the second processing unit is used for carrying out Fourier transform and acoustic feature extraction operation on the fifth voice record to obtain a sixth voice record; a training unit, configured to train a voiceprint model according to the sixth voice recording to obtain a trained voiceprint model; and the input unit is used for inputting the target voice record into the trained voiceprint model to obtain a first target voiceprint.

As an optional example, the alignment module includes: a second obtaining unit, configured to obtain a target voiceprint feature vector of the first target voiceprint; a third obtaining unit, configured to obtain all voiceprint feature vectors in the voiceprint database; a scoring unit, configured to score similarity between the target voiceprint feature vector and all the voiceprint feature vectors to obtain a target score group; a third processing unit, configured to obtain the first comparison result as a successful comparison if a score value greater than or equal to a threshold value exists in the target score group; a fourth processing unit, configured to obtain the first comparison result as a successful non-comparison if no score value greater than or equal to the threshold exists in the target score group.

As an alternative example, the scoring unit includes: a scoring unit, configured to score similarity between the target voiceprint feature vector and all the voiceprint feature vectors to obtain an original score group; and the processing subunit is used for multiplying each fraction value in the original fraction array by 100 to obtain the target fraction array.

As an optional example, the first marking module includes: a fourth obtaining unit, configured to obtain a first registration time of the client when the risk early warning record exists in a client database; a first determining unit, configured to determine that the client is the risk client when the first registration time does not exceed a protection period; a second determining unit, configured to determine that the client is the third party client when the first registration time exceeds a protection period.

As an optional example, the apparatus further includes: a first determining module, configured to determine that the client is the third-party client when the client has the third-party client tag; and a second determining module, configured to determine that the client is the third-party client when the risk pre-warning record does not exist in the client database.

As an optional example, the apparatus further includes: a second obtaining module, configured to obtain a second target voiceprint of the client when the client visits naturally; the second comparison module is used for comparing the second target voiceprint with the voiceprint database to obtain a second comparison result; and a third marking module, configured to mark the client as a naturally visited client if the second comparison result is that the comparison is unsuccessful, and store the second target voiceprint in the voiceprint database.

According to still another aspect of the embodiments of the present invention, there is also provided a storage medium having a computer program stored therein, wherein the computer program executes the above voiceprint recognition method when executed by a processor.

According to still another aspect of the embodiments of the present invention, there is also provided an electronic device, including a memory and a processor, where the memory stores a computer program, and the processor is configured to execute the voiceprint recognition method through the computer program.

In the process that the voiceprint recognition method can be used for predicting and optimizing the marketing intelligent technology, the target voice recording of a client is obtained; performing voiceprint extraction operation on the target voice recording to obtain a first target voiceprint; comparing the first target voiceprint with a voiceprint database to obtain a first comparison result; under the condition that the first comparison result is that the comparison is successful, marking the client as a risk client and sending risk early warning information to the user; in the method, whether the first target voiceprint of the client exists or not is searched in the voiceprint database, if yes, the client is divided into risk clients, the purposes of clearly dividing the attribution of the client and giving an early warning in time are achieved, and the technical problem of drastic increase of marketing cost caused by unclear division of the attribution of the client is solved.

Drawings

The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the invention without limiting the invention. In the drawings:

FIG. 1 is a flow diagram of an alternative voiceprint recognition method according to an embodiment of the invention;

FIG. 2 is a comparison diagram of an alternative voiceprint recognition method according to embodiments of the invention;

FIG. 3 is a graph of a first target voiceprint acquisition for an alternative voiceprint recognition method in accordance with embodiments of the present invention;

FIG. 4 is a diagram illustrating the acquisition of an array of target scores for an alternative voiceprint recognition method according to an embodiment of the invention;

FIG. 5 is a subscription flow diagram of an alternative voiceprint recognition method according to an embodiment of the invention;

FIG. 6 is a schematic diagram of an alternative voiceprint recognition apparatus according to an embodiment of the present invention;

FIG. 7 is a schematic diagram of an alternative electronic device according to an embodiment of the invention.

Detailed Description

In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without making any creative effort based on the embodiments in the present invention, shall fall within the protection scope of the present invention.

It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are capable of operation in other sequences than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

According to a first aspect of the embodiments of the present invention, there is provided a voiceprint recognition method, optionally, as shown in fig. 1, the method includes:

s102, acquiring a target voice record of a client;

s104, performing voiceprint extraction operation on the target voice recording to obtain a first target voiceprint;

s106, comparing the first target voiceprint with the voiceprint database to obtain a first comparison result;

s108, under the condition that the first comparison result is that the comparison is successful, marking the client as a risk client, and sending risk early warning information to the user;

s110, under the condition that the first comparison result is that the comparison is not successful, the client is marked as a third-party client, and the first target voiceprint is stored in the voiceprint database.

Optionally, in this embodiment, the client is a client who goes to the real estate sales project and wants to purchase a real estate, and may be a naturally visited client who is not taken through the real estate agency outside the third scenario, or a client who is taken through the real estate agency outside the third scenario. The voiceprint is a sound wave frequency spectrum which carries speech information and is displayed by an electro-acoustic instrument, and the voiceprint not only has specificity, but also has the characteristic of relative stability. A voiceprint database comprising all voiceprints of customers historically visiting a real estate sales arena. The risk client is a client with a risk of 'flyer bill'. The third party customer is a customer brought by the house property intermediary off-site through the third scenario and without the risk of "flyer orders".

Optionally, in this embodiment, the target voice recording generated in the communication between the visiting client and the user may be obtained through a voice card on the user. Wherein the user is a live advisor in a real estate sales arena. And carrying out voiceprint extraction operation on the target voice record of the client to obtain a first target voiceprint of the client. And comparing the first target voiceprint with the voiceprint database, searching whether the first target voiceprint exists in the voiceprint database, marking the client as a risk client with a 'flyer bill' risk if the comparison is successful, namely the first target voiceprint exists in the voiceprint database, and sending risk early warning information to the user to prompt the user. And if the comparison is not successful, namely the first target voiceprint does not exist in the voiceprint database, the client is shown as the first visit without the risk of 'flyer bill', the client is marked as a third-party client, and the first target voiceprint is stored in the voiceprint database.

Optionally, in this embodiment, whether the first target voiceprint of the client exists is searched in the voiceprint database, and if the first target voiceprint of the client exists, the client is divided into risk clients, so that the purposes of clearly dividing the affiliation of the client and timely early warning are achieved, and the technical problem of the sharp increase of the marketing cost caused by unclear division of the affiliation of the client is solved.

As an alternative example, obtaining a target voice recording for a customer includes:

acquiring a first voice record of communication among a user, a client and a third party;

carrying out voice quality detection operation on the first voice recording to obtain a second voice recording with qualified quality;

and carrying out voice separation operation on the second voice record to obtain a third voice record of the user, a target voice record of the client and a fourth voice record of a third party.

Optionally, in this embodiment, the user is a live advisor in a real estate sales arena. The third party is a real estate agency outside the project. The client who takes the case by the third party is accepted by the user's setting consultant, and the first voice recording of the three-party communication is obtained through the voice work card on the user, wherein the first voice recording comprises the voice recording of the user, the voice recording of the client and the voice recording of the third party. And finally, carrying out voice separation operation on the second voice recording to separate out a target voice recording of the client, wherein the target voice recording also comprises a third voice recording of the user and a fourth voice recording of a third party.

As an alternative example, performing a voice quality detection operation on a first voice recording includes:

carrying out voice quality detection operation on the first voice recording to obtain a detection result;

deleting the first voice recording under the condition that the detection result is unqualified;

and under the condition that the detection result is qualified, reserving the first voice recording.

Optionally, in this embodiment, the voice quality detection operation is performed on the first voice record, and it may be detected that the voice is unclear or the voice of which party cannot be distinguished, and these voice records with unqualified quality are deleted, and the voice records with qualified quality are retained, so as to obtain the second voice record with qualified quality.

As an alternative example, performing a voiceprint extraction operation on the target voice recording to obtain the first target voiceprint comprises:

carrying out noise removal technology and effective sound extraction operation on the target voice record to obtain a fifth voice record;

carrying out Fourier transform and acoustic feature extraction operation on the fifth voice record to obtain a sixth voice record;

training the voiceprint model according to the sixth voice recording to obtain a trained voiceprint model;

and inputting the target voice record into the trained voiceprint model to obtain a first target voiceprint.

Optionally, in this embodiment, the target voice record is preprocessed, which includes a noise removal operation and an effective sound extraction operation, to obtain a fifth voice record, and then fourier transform and acoustic feature extraction operations are performed to obtain a sixth voice record, and the sixth voice record is input to an untrained voiceprint model based on the deep neural network for training until a loss function converges, so as to obtain a trained voiceprint model.

As an optional example, comparing the first target voiceprint with the voiceprint database, and obtaining the first comparison result includes:

acquiring a target voiceprint feature vector of a first target voiceprint;

acquiring all voiceprint characteristic vectors in a voiceprint database;

scoring the similarity of the target voiceprint feature vectors and all the voiceprint feature vectors to obtain a target score group;

under the condition that the score value which is greater than or equal to the threshold value exists in the target score group, obtaining a first comparison result as successful comparison;

and under the condition that the score value which is larger than or equal to the threshold value does not exist in the target score group, obtaining the first comparison result as the success of the non-comparison.

Optionally, in this embodiment, a target voiceprint feature vector of a first target voiceprint of a client and all voiceprint feature vectors in a voiceprint database may be extracted through a CNN-based deep neural network, similarity between the target voiceprint feature vector and all voiceprint feature vectors is scored to obtain a target score group, for example, 5 voiceprint data exist in the voiceprint database, similarity scoring is performed on 5 voiceprint feature vectors, the first voiceprint feature vector, the second voiceprint feature vector, the third voiceprint feature vector, the fourth voiceprint feature vector, and the fifth voiceprint feature vector, respectively to obtain a similarity score 85 between the first voiceprint feature vector and the target voiceprint feature vector, and so on to obtain a target score group [ 85,52, 36, 68,98 ]. The threshold value may be selected 95, and in the case where there is a score value 98 in the target score set that is greater than or equal to the threshold value 95, the first comparison result is a successful comparison. If the target score group is found to be [ 85,52, 36, 68,94 ], if no score value greater than or equal to the threshold value 95 is present in the target score group, the first comparison result is found to be a successful mis-comparison.

As an optional example, scoring the similarity between the target voiceprint feature vector and all the voiceprint feature vectors to obtain a target score group includes:

scoring the similarity of the target voiceprint feature vector and all the voiceprint feature vectors to obtain an original score group;

and multiplying each fraction value in the original fraction group by 100 to obtain a target fraction group.

Optionally, in this embodiment, the similarity between the target voiceprint feature vector and all the voiceprint feature vectors is scored to obtain an original score group, for example, 5 voiceprint data exist in the voiceprint database, 5 voiceprint feature vectors are obtained, and a similarity score between the first voiceprint feature vector, the second voiceprint feature vector, the third voiceprint feature vector, the fourth voiceprint feature vector, and the fifth voiceprint feature vector is obtained and is 0.85, so as to obtain an original score group [ 0.85, 0.52, 0.36, 0.68,0.94 ], and each score value in the original score group is multiplied by 100 to obtain an target score group [ 85,52, 36, 68,94 ].

As an optional example, in a case that the first comparison result is that the comparison is successful, marking the client as a risk client, and sending risk early warning information to the user includes:

the method comprises the steps that under the condition that risk early warning records exist in a client database, the first registration time of a client is obtained;

determining that the client is a risk client under the condition that the first registration time does not exceed the protection period;

and determining the client as a third party client when the first registration time exceeds the protection period.

Optionally, in this embodiment, when the first comparison result is that the comparison is successful, that is, a voiceprint of the client exists in the voiceprint database, and the client does not visit for the first time, whether the client has a risk early warning record is queried in the client database, if so, a first registration time of the client, that is, a first visit time, is obtained, and whether the first registration time exceeds a protection period is queried, and if still within the protection period, that is, if not, the client is determined to be a risk client with a risk of "flyer bill", and then no subscription is performed subsequently. If the client is not in the protection period, namely the protection period is exceeded, the client is determined to be a third-party client, and the subsequent subscription can be normally carried out. For example, the first registration time of the client is 2021.01.01, the protection period is one year, and the current time is 2021.12.25, and the client is still in the protection period, if the current time is 2022.04.01, the client is not in the protection period.

As an optional example, the method further includes:

determining that the client is a third-party client under the condition that the client has a third-party client tag;

and under the condition that no risk early warning record exists in the client database, determining that the client is a third-party client.

Optionally, in this embodiment, the client is queried in the client database to have a third-party client tag, and it is determined that the client is a third-party client, and then the subscription may be normal. And (4) no risk early warning record exists in the client database, the client is determined to be a third-party client, and the client can normally sign a contract subsequently.

As an optional example, the method further comprises:

under the condition that the client visits naturally, a second target voiceprint of the client is obtained;

comparing the second target voiceprint with the voiceprint database to obtain a second comparison result;

and under the condition that the second comparison result is that the comparison is not successful, the client is marked as a natural visiting client, and the second target voiceprint is stored in the voiceprint database.

Alternatively, in this embodiment, the client is a naturally-visited client, who is not being brought to the case by a third party, is taken by the user's home advisor, obtaining the voice record of the communication between the two parties through the voice work card on the user, obtaining the voice record with qualified quality through voice quality detection, finally carrying out voice separation operation to separate out a second target voice record of the client, obtaining a second target voiceprint of the client through the trained voiceprint model, comparing the second target voiceprint with the voiceprint database, searching whether the second target voiceprint exists in the voiceprint database, and if the comparison is not successful, namely the second target voiceprint does not exist in the voiceprint database, and marking the client as a naturally visited client, storing the second target voiceprint in a voiceprint database, and then normally signing.

Alternatively, the "flyer bill" mentioned in the present application is a business consultant in the pitch point and a real estate agency outside the pitch point, and the marketing cost is increased dramatically by the method of single external hanging in the hidden guest, mutual introduction of guest interception, etc., or by the unclear attribution division of the guest. The invention relates to a method for preventing 'flyer bill' in a case, which is based on voiceprint recognition technology, and comprises the steps of obtaining voice generated during communication through a voice work card worn by a business consultant in the case, carrying out operations such as quality detection, voice separation, voiceprint extraction and the like to obtain a first target voiceprint of a client, respectively comparing the first target voiceprint with historical voiceprints in a voiceprint database in a ratio of 1: N, and indicating that the client visits for the first time without 'flyer bill' risk if the comparison is not successful, namely the voiceprint database does not have the first target voiceprint.

The equipment characteristics of the voice card: designing a worker card style, integrating into a scene link, wearing the worker card without perception, and weighing only 25 g; only a switch key is needed, recording is carried out when the mobile phone is started, and frequent equipment operation is reduced; recording for continuation of the journey within 10 hours, and storing for 16G to meet the requirement of one-day audio data acquisition; data are returned and charged in a wired mode, and the Bluetooth is supported to check the state of the equipment; the Internet platform composed of the objects directly manages the equipment and supports the optimization and the upgrade of the functions of the equipment firmware.

The specific flow is shown in fig. 2:

a, the client brought by the house property intermediary outside the third scheme:

1. obtaining a voice recording of three parties generated during communication through a voice work card to obtain a first voice recording;

2. performing voice quality detection operation on the first voice record, deleting the voice record with unqualified quality, and reserving the voice record with qualified quality to obtain a second voice record with qualified quality;

3. carrying out voice separation operation on the second voice record to separate out a target voice record of the client, and further comprising a third voice record of the user and a fourth voice record of a third party;

4. the method comprises the steps of carrying out noise removing technology, effective sound extracting operation, Fourier transform and acoustic feature extracting operation on a target voice recording to obtain a sixth voice recording, training a voiceprint model according to the sixth voice recording to obtain a trained voiceprint model, and inputting the target voice recording to the trained voiceprint model to obtain a first target voiceprint. As shown in fig. 3;

5. comparing the first target voiceprint with the voiceprint database, marking the client as a risk client under the condition of successful comparison, sending risk early warning information to the user, marking the client as a third-party client under the condition of unsuccessful comparison, and storing the first target voiceprint into the voiceprint database;

comparing the first target voiceprint to the voiceprint database comprises, as shown in fig. 4:

(1) acquiring a target voiceprint characteristic vector of a first target voiceprint and all voiceprint characteristic vectors in a voiceprint database;

(2) scoring the similarity of the target voiceprint feature vector and all the voiceprint feature vectors to obtain an original score group;

(3) and multiplying each fraction value in the original fraction group by 100 to obtain a target fraction group.

(4) Obtaining a first comparison result as successful comparison under the condition that the target score group has a score value which is greater than or equal to the threshold value; and under the condition that the score value which is larger than or equal to the threshold value does not exist in the target score group, obtaining the first comparison result as the success of the non-comparison.

6. And inquiring whether the client has risk early warning records in a client database, if not, determining that the client is a third-party client, and then normally signing. If the client is in the protection period, determining that the client is a risk client and not signing subsequently. If the client is not in the protection period, the client is determined to be a third party client, and then the client can normally sign up, as shown in fig. 5.

B. Naturally visited customers:

the method comprises the steps of obtaining voice records communicated between two parties through a voice work card on a user body, obtaining the voice records with qualified quality through voice quality detection, finally carrying out voice separation operation to separate a second target voice record of a client, obtaining a second target voiceprint of the client through a trained voiceprint model, comparing the second target voiceprint with a voiceprint database, searching whether the second target voiceprint exists in the voiceprint database, marking the client as a client with natural visit under the condition of unsuccessful comparison, storing the second target voiceprint to the voiceprint database, and enabling follow-up signing to be normal.

It should be noted that, for simplicity of description, the above-mentioned method embodiments are described as a series of acts or combination of acts, but those skilled in the art will recognize that the present invention is not limited by the order of acts, as some steps may occur in other orders or concurrently in accordance with the invention. Further, those skilled in the art should also appreciate that the embodiments described in the specification are preferred embodiments and that the acts and modules referred to are not necessarily required by the invention.

According to another aspect of the embodiments of the present application, there is also provided a voiceprint recognition apparatus, as shown in fig. 6, including:

an obtaining module 602, configured to obtain a target voice recording of a client;

an extracting module 604, configured to perform a voiceprint extraction operation on the target voice recording to obtain a first target voiceprint;

a comparison module 606, configured to compare the first target voiceprint with the voiceprint database to obtain a first comparison result;

the first marking module 608 is configured to mark the client as a risk client and send risk early warning information to the user when the first comparison result is that the comparison is successful;

the second marking module 610 is configured to mark the client as a third-party client and store the first target voiceprint to the voiceprint database when the first comparison result is that the comparison is not successful.

Optionally, in this embodiment, the client is a client who goes to the real estate sales project and wants to purchase a real estate, and may be a naturally visited client who is not taken through the real estate agency outside the third scenario, or a client who is taken through the real estate agency outside the third scenario. The voiceprint is a sound wave frequency spectrum which carries speech information and is displayed by an electro-acoustic instrument, and the voiceprint not only has specificity, but also has the characteristic of relative stability. A voiceprint database comprising all voiceprints of customers historically visiting a real estate sales arena. An at risk customer is a customer who has a "flyer" risk. The third party customer is a customer who is brought by the house property intermediary outside the third scenario and who has no risk of "flyer orders".

Optionally, in this embodiment, the target voice recording generated in the communication between the visiting client and the user may be obtained through a voice card on the user. Wherein the user is a presence advisor in the real estate sales desk. And carrying out voiceprint extraction operation on the target voice record of the client to obtain a first target voiceprint of the client. And comparing the first target voiceprint with the voiceprint database, searching whether the first target voiceprint exists in the voiceprint database, if the comparison is successful, namely the first target voiceprint exists in the voiceprint database, marking the client as a risk client with a 'flyer' risk, and sending risk early warning information to the user to prompt the user. And if the comparison is not successful, namely the first target voiceprint does not exist in the voiceprint database, the client is shown as the first visit without the risk of 'flyer bill', the client is marked as a third-party client, and the first target voiceprint is stored in the voiceprint database.

As an optional example, the obtaining module includes:

the first acquisition unit is used for acquiring a first voice record of communication among the user, the client and a third party;

the detection unit is used for carrying out voice quality detection operation on the first voice record to obtain a second voice record with qualified quality;

and the separation unit is used for carrying out voice separation operation on the second voice record to obtain a third voice record of the user, a target voice record of the client and a fourth voice record of a third party.

Optionally, in this embodiment, the user is a live advisor in a real estate sales arena. The third party is a real estate agency outside the project. The third party takes the client to the case, the client is received by the user's employment consultant, and the first voice recording of the three-party communication is obtained through the voice work card on the user, wherein the first voice recording comprises the voice recording of the user, the voice recording of the client and the voice recording of the third party. And finally, carrying out voice separation operation on the second voice recording to separate out a target voice recording of the client, wherein the target voice recording also comprises a third voice recording of the user and a fourth voice recording of a third party.

As an alternative example, the detection unit includes:

the detection subunit is used for carrying out voice quality detection operation on the first voice record to obtain a detection result;

the deleting subunit is used for deleting the first voice record under the condition that the detection result is unqualified;

and the reservation subunit is used for reserving the first voice recording under the condition that the detection result is qualified.

As an alternative example, the extraction module comprises:

the first processing unit is used for carrying out noise removal technology and effective sound extraction operation on the target voice record to obtain a fifth voice record;

the second processing unit is used for carrying out Fourier transform and acoustic feature extraction operation on the fifth voice record to obtain a sixth voice record;

the training unit is used for training the voiceprint model according to the sixth voice recording to obtain a trained voiceprint model;

and the input unit is used for inputting the target voice record to the trained voiceprint model to obtain a first target voiceprint.

As an optional example, the alignment module comprises:

a second obtaining unit, configured to obtain a target voiceprint feature vector of the first target voiceprint;

a third obtaining unit, configured to obtain all voiceprint feature vectors in the voiceprint database;

the scoring unit is used for scoring the similarity of the target voiceprint feature vector and all the voiceprint feature vectors to obtain a target score group; the third processing unit is used for obtaining a first comparison result as successful comparison under the condition that the score value which is greater than or equal to the threshold value exists in the target score group;

and the fourth processing unit is used for obtaining that the first comparison result is that the non-comparison is successful under the condition that the score value which is larger than or equal to the threshold value does not exist in the target score group.

Optionally, in this embodiment, a target voiceprint feature vector of a first target voiceprint of the client and all voiceprint feature vectors in the voiceprint database may be extracted through the CNN-based deep neural network, and the similarity between the target voiceprint feature vector and all voiceprint feature vectors is scored to obtain a target score group, for example, 5 voiceprint data exist in the voiceprint database, and then similarity scoring is performed on 5 voiceprint feature vectors, a first voiceprint feature vector, a second voiceprint feature vector, a third voiceprint feature vector, a fourth voiceprint feature vector, and a fifth voiceprint feature vector to obtain a similarity score 85 between the first voiceprint feature vector and the target voiceprint feature vector, and so on to obtain a target score group [ 85,52, 36, 68,98 ]. The threshold value may be selected 95, and in the case where there is a score value 98 in the target score group that is greater than or equal to the threshold value 95, the first comparison result is a successful comparison. If the target score group is found to be [ 85,52, 36, 68,94 ], if no score value greater than or equal to the threshold value 95 is present in the target score group, the first comparison result is found to be a successful mis-comparison.

As an alternative example, the scoring unit includes:

the scoring subunit is used for scoring the similarity of the target voiceprint feature vector and all the voiceprint feature vectors to obtain an original score group;

and the processing subunit is used for multiplying each fraction value in the original fraction array by 100 to obtain a target fraction array.

As an alternative example, the first marking module comprises:

the fourth acquisition unit is used for acquiring the first registration time of the client under the condition that the client has risk early warning records in the client database;

the first determining unit is used for determining the client as a risk client under the condition that the first registration time does not exceed the protection period;

and the second determining unit is used for determining that the client is the third party client when the first registration time exceeds the protection period.

Optionally, in this embodiment, when the first comparison result is that the comparison is successful, that is, a voiceprint of the client exists in the voiceprint database, and the client does not visit for the first time, whether the client has a risk early warning record is queried in the client database, if yes, the first registration time of the client, that is, the first visit time, is obtained, and whether the first registration time exceeds the protection period is queried, if still within the protection period, that is, under the condition that the protection period is not exceeded, it is determined that the client is a risk client with a risk of "flyer bill", and then no subscription is performed. If the protection period is not within the protection period, namely the protection period is exceeded, the client is determined to be a third-party client, and then the client can normally sign. For example, the first registration time of the client is 2021.01.01, the protection period is one year, the current time is 2021.12.25, and the client is not in the protection period if the current time is 2022.04.01.

As an optional example, the apparatus further includes:

the first determining module is used for determining that the client is the third-party client under the condition that the client has the third-party client tag;

and the second determining module is used for determining that the client is a third-party client under the condition that the risk early warning record does not exist in the client database.

As an optional example, the apparatus further includes:

the second acquisition module is used for acquiring a second target voiceprint of the client under the condition that the client visits naturally;

the second comparison module is used for comparing the second target voiceprint with the voiceprint database to obtain a second comparison result;

and the third marking module is used for marking the client as a natural visiting client under the condition that the second comparison result is that the comparison is not successful, and storing the second target voiceprint to the voiceprint database.

Alternatively, in this embodiment, the client is a naturally-visited client, who is not being brought to the case by a third party, is taken by the user's home advisor, the voice record communicated between the two parties is obtained through the voice work card on the user, the voice record with qualified quality is obtained through voice quality detection, finally, voice separation operation is carried out to separate out a second target voice record of the client, obtaining a second target voiceprint of the client through the trained voiceprint model, comparing the second target voiceprint with the voiceprint database, searching whether the second target voiceprint exists in the voiceprint database, and if the comparison is not successful, namely the second target voiceprint does not exist in the voiceprint database, and marking the client as a naturally visited client, storing the second target voiceprint in a voiceprint database, and then normally signing.

Please refer to the example for other examples of the embodiment, which is not described herein again.

Fig. 7 is a block diagram of an alternative electronic device according to an embodiment of the present application, as shown in fig. 7, including a processor 702, a communication interface 704, a memory 706 and a communication bus 708, where the processor 702, the communication interface 704 and the memory 706 communicate with each other via the communication bus 708, where,

a memory 706 for storing computer programs;

the processor 702, when executing the computer program stored in the memory 706, performs the following steps:

acquiring a target voice recording of a client;

performing voiceprint extraction operation on the target voice recording to obtain a first target voiceprint;

comparing the first target voiceprint with the voiceprint database to obtain a first comparison result;

under the condition that the first comparison result is that the comparison is successful, marking the client as a risk client, and sending risk early warning information to the user;

and under the condition that the first comparison result is that the comparison is not successful, marking the client as a third-party client, and storing the first target voiceprint to the voiceprint database.

Alternatively, in this embodiment, the communication bus may be a PCI (Peripheral Component Interconnect) bus, an EISA (Extended Industry Standard Architecture) bus, or the like. The communication bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is shown in FIG. 7, but that does not indicate only one bus or one type of bus. The communication interface is used for communication between the electronic equipment and other equipment.

The memory may include RAM, and may also include non-volatile memory (non-volatile memory), such as at least one disk memory. Alternatively, the memory may be at least one memory device located remotely from the processor.

As an example, the memory 706 may include, but is not limited to, the obtaining module 602, the extracting module 604, the comparing module 606, the first marking module 608, and the second marking module 610 of the voiceprint recognition apparatus. In addition, the module may further include, but is not limited to, other module units in the processing apparatus of the request, which is not described in this example again.

The processor may be a general-purpose processor, and may include but is not limited to: a CPU (Central Processing Unit), NP (Network Processor), and the like; but also a DSP (Digital Signal Processing), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other Programmable logic device, discrete Gate or transistor logic device, discrete hardware component.

Optionally, the specific examples in this embodiment may refer to the examples described in the above embodiments, and this embodiment is not described herein again.

It can be understood by those skilled in the art that the structure shown in fig. 7 is only an illustration, and the device implementing the voiceprint recognition method may be a terminal device, and the terminal device may be a terminal device such as a smart phone (e.g., an Android phone, an iOS phone, etc.), a tablet computer, a palmtop computer, and a Mobile Internet Device (MID), a PAD, and the like. Fig. 7 does not limit the structure of the electronic device. For example, the electronic device may also include more or fewer components (e.g., network interfaces, display devices, etc.) than shown in FIG. 7, or have a different configuration than shown in FIG. 7.

Those skilled in the art will appreciate that all or part of the steps in the methods of the above embodiments may be implemented by a program instructing hardware associated with the terminal device, where the program may be stored in a computer-readable storage medium, and the storage medium may include: flash disk, ROM, RAM, magnetic or optical disk, and the like.

According to yet another aspect of the embodiments of the present invention, there is also provided a computer-readable storage medium, in which a computer program is stored, wherein the computer program, when executed by a processor, performs the steps in the above-mentioned voiceprint recognition method.

Alternatively, in this embodiment, a person skilled in the art may understand that all or part of the steps in the various methods in the foregoing embodiments may be implemented by a program instructing hardware related to the terminal device, where the program may be stored in a computer-readable storage medium, and the storage medium may include: flash disks, Read-Only memories (ROMs), Random Access Memories (RAMs), magnetic or optical disks, and the like.

The above-mentioned serial numbers of the embodiments of the present invention are only for description, and do not represent the advantages and disadvantages of the embodiments.

The integrated unit in the above embodiments, if implemented in the form of a software functional unit and sold or used as a separate product, may be stored in the above computer-readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for causing one or more computer devices (which may be personal computers, servers, network devices, etc.) to execute all or part of the steps of the method according to the embodiments of the present invention.

In the above embodiments of the present invention, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.

In the several embodiments provided in the present application, it should be understood that the disclosed client may be implemented in other manners. The above-described apparatus embodiments are merely illustrative, and for example, the division of the units is only one type of logical functional division, and other divisions may be implemented in practice, for example, multiple units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed coupling or direct coupling or communication connection between each other may be an indirect coupling or communication connection through some interfaces, units or modules, and may be electrical or in other forms.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.

The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.

Claims

1. A voiceprint recognition method, comprising:

acquiring a target voice recording of a client;

comparing the first target voiceprint with a voiceprint database to obtain a first comparison result;

under the condition that the first comparison result is successful, marking the client as a risk client, and sending risk early warning information to a user;

2. The method of claim 1, wherein obtaining the target voice recording of the customer comprises:

acquiring a first voice record of communication among the user, the client and a third party;

and carrying out voice separation operation on the second voice record to obtain a third voice record of the user, a target voice record of the client and a fourth voice record of the third party.

3. The method of claim 2 wherein the performing a voice quality detection operation on the first voice recording comprises:

deleting the first voice recording under the condition that the detection result is unqualified in quality;

4. The method of claim 1, wherein performing a voiceprint extraction operation on the target voice recording to obtain a first target voiceprint comprises:

carrying out noise removal technology and effective sound extraction operation on the target voice recording to obtain a fifth voice recording;

training a voiceprint model according to the sixth voice recording to obtain a trained voiceprint model;

and inputting the target voice record to the trained voiceprint model to obtain a first target voiceprint.

5. The method of claim 1, wherein comparing the first target voiceprint to a voiceprint database to obtain a first comparison result comprises:

acquiring a target voiceprint feature vector of the first target voiceprint;

acquiring all voiceprint characteristic vectors in the voiceprint database;

scoring the similarity of the target voiceprint feature vector and all the voiceprint feature vectors to obtain a target score group;

obtaining the first comparison result as successful comparison under the condition that the score value which is greater than or equal to the threshold value exists in the target score group;

and under the condition that the score value which is larger than or equal to the threshold value does not exist in the target score group, obtaining the first comparison result as the non-comparison success.

6. The method according to claim 5, wherein said scoring the similarity of the target voiceprint feature vector and all the voiceprint feature vectors to obtain a target score set comprises:

and multiplying each fraction value in the original fraction array by 100 to obtain the target fraction array.

7. The method of claim 1, wherein if the first comparison result is a successful comparison, marking the client as a risk client, and sending risk early warning information to a user comprises:

determining that the client is the at-risk client if the first registration time does not exceed a protection period;

determining that the customer is the third party customer if the first registration time exceeds a protection period.

8. The method of claim 7, further comprising:

determining that the customer is the third party customer in the event that the customer has the third party customer tag;

and under the condition that the risk early warning record does not exist in the client database, determining that the client is the third-party client.

9. The method of claim 1, further comprising:

under the condition that the client visits naturally, acquiring a second target voiceprint of the client;

10. A voiceprint recognition apparatus comprising:

the acquisition module is used for acquiring a target voice record of a client;

the extraction module is used for carrying out voiceprint extraction operation on the target voice recording to obtain a first target voiceprint;

the comparison module is used for comparing the first target voiceprint with the voiceprint database to obtain a first comparison result;

the first marking module is used for marking the client as a risk client and sending risk early warning information to the user under the condition that the first comparison result is that the comparison is successful;

and the second marking module is used for marking the client as a third-party client under the condition that the first comparison result is that the comparison is not successful, and storing the first target voiceprint to the voiceprint database.

11. A computer-readable storage medium, in which a computer program is stored, which, when being executed by a processor, carries out the method of any one of claims 1 to 9.

12. An electronic device comprising a memory and a processor, characterized in that the memory has stored therein a computer program, the processor being arranged to execute the method of any of claims 1 to 9 by means of the computer program.