Information identification method, device and equipment and storage medium thereof

Abstract

The embodiment of the invention belongs to the technical field of real-time voice processing, is applied to a real-time voice scene of bank business handling, and relates to an information recognition method, device and equipment and a storage medium thereof. Extracting an MFCC feature; inputting the MFCC features into a pre-trained voiceprint feature extraction model, and extracting multi-dimensional voiceprint features; and inputting the multi-dimensional voiceprint features into a pre-trained customer information recognition model, recognizing customer information corresponding to the target object, and outputting the customer information. The client information, such as age, gender, accent and emotion information during speaking, of a target object is recognized through a voice processing technology, the information recognition method is applied to a real-time voice scene of bank business handling, a bank supervision terminal can be helped to know the client information of online clients in real time, and the user experience is improved. And when the emotion of the customer during speaking fluctuates, corresponding adjustment can be made in time, so that the system is more intelligent and humanized.

Description

Technical Field

[0001] The present application relates to the field of real-time voice processing technology, and is applied to real-time voice scenarios in banking business processing, and in particular to an information recognition method, device, equipment and storage medium thereof. Background Art

[0002] With the rapid development of artificial intelligence technology, more and more intelligent devices are being applied to various industries to help users handle business intelligently. Especially in government services or banking business handling, basically every bank hall will be equipped with operating terminals for intelligent business handling.

[0003] However, most of the operating terminals for intelligent business handling installed in every bank lobby currently still use the most basic human-computer interaction click command method to handle business. Even if there are operating terminals that can refer to customer voice to handle business, they are only limited to simple business handling based on simple voice commands. They are not intelligent enough. After the user's voice commands are issued multiple times, if emotional changes occur, they cannot be discovered and reported to the business personnel in the bank lobby in time. This is not user-friendly and can easily cause a bad business experience for customers. Summary of the invention

[0004] The purpose of the embodiments of the present application is to propose an information identification method, device, equipment and storage medium thereof to solve the problem that in the existing real-time voice scenarios of banking business processing, the relevant operation terminals are not intelligent and humanized enough, which easily causes a bad business experience for customers.

[0005] In order to solve the above technical problems, the present application provides an information identification method, which adopts the following technical solutions:

[0006] An information identification method comprises the following steps:

[0007] Obtain the target object's speech to be recognized;

[0008] Inputting the speech to be recognized into a preset speech feature extraction model to extract MFCC features contained in the speech to be recognized;

[0009] Inputting the MFCC features contained in the speech to be recognized into a pre-trained voiceprint feature extraction model to extract the multi-dimensional voiceprint features contained in the MFCC features;

[0010] The multi-dimensional voiceprint feature is input into a pre-trained customer information recognition model to identify and output the customer information corresponding to the target object.

[0011] Furthermore, before executing the step of inputting the MFCC features contained in the speech to be recognized into the pre-trained voiceprint feature extraction model to extract the multi-dimensional voiceprint features contained in the MFCC features, the method further includes:

[0012] Obtaining labeled speech of a batch of users as a training data set, wherein the labeled speech indicators are annotated with gender information, age information, accent information, and emotional information of the corresponding user when speaking;

[0013] Inputting the training data set into the preset speech feature extraction model to extract MFCC features corresponding to all users in the batch of users respectively;

[0014] According to the preset user identifier, construct the gender information, age information, accent information and emotional information when speaking corresponding to each of the MFCC features;

[0015] The MFCC features corresponding to all users in the batch of users are respectively input into the voiceprint feature extraction model to be trained, and the voiceprint feature extraction model to be trained is trained to obtain a pre-trained voiceprint feature extraction model.

[0016] Furthermore, the step of inputting the MFCC features corresponding to all users in the batch of users into the voiceprint feature extraction model to be trained, training the voiceprint feature extraction model to be trained, and obtaining a pre-trained voiceprint feature extraction model specifically includes:

[0017] Using the voiceprint feature extraction model to be trained to perform multi-dimensional voiceprint feature extraction on the MFCC features corresponding to all users in the batch of users;

[0018] According to the user identifier, the multi-dimensional voiceprint features corresponding to each of the MFCC features are determined to obtain a pre-trained voiceprint feature extraction model.

[0019] Furthermore, before executing the step of inputting the multi-dimensional voiceprint feature into the pre-trained customer information recognition model to identify and output the customer information corresponding to the target object, the method further includes:

[0020] Based on the multidimensional voiceprint features corresponding to each of the MFCC features, and the gender information, age information, accent information, and emotional information when speaking corresponding to each of the MFCC features, a probability mapping relationship between the multidimensional voiceprint features and the gender information, age information, accent information, and emotional information when speaking is calculated;

[0021] The probability mapping relationship between multi-dimensional voiceprint features and gender information, the probability mapping relationship between multi-dimensional voiceprint features and age information, the probability mapping relationship between multi-dimensional voiceprint features and accent information, and the probability mapping relationship between multi-dimensional voiceprint features and emotional information during speaking are used as learning knowledge to construct a learning knowledge data set;

[0022] The learning knowledge data set is input into the customer information recognition model to be trained, and the customer information recognition model to be trained is trained to obtain a pre-trained customer information recognition model.

[0023] Furthermore, the customer information recognition model to be trained includes a gender information recognition sub-model, an age information recognition sub-model, an accent information recognition sub-model and an emotion information recognition sub-model, and the step of inputting the learning knowledge data set into the customer information recognition model to be trained, performing learning training on the customer information recognition model to be trained, and obtaining the pre-trained customer information recognition model specifically includes:

[0024] Deploy the probability mapping relationship between the multi-dimensional voiceprint features and the gender information to the gender information recognition sub-model to obtain a gender information recognition sub-model that has been trained and learned;

[0025] Deploy the probability mapping relationship between the multi-dimensional voiceprint feature and the age information to the age information recognition sub-model to obtain the age information recognition sub-model that has been learned and trained;

[0026] Deploy the probability mapping relationship between the multi-dimensional voiceprint features and the accent information to the accent information recognition sub-model to obtain a learned and trained accent information recognition sub-model;

[0027] The probability mapping relationship between the multi-dimensional voiceprint features and the emotional information during speaking is deployed to the emotional information recognition sub-model to obtain the emotional information recognition sub-model that has been learned and trained.

[0028] Furthermore, the step of inputting the multi-dimensional voiceprint feature into a pre-trained customer information recognition model, identifying the customer information corresponding to the target object and outputting the same specifically includes:

[0029] Inputting the multidimensional voiceprint feature into the gender information recognition sub-model that has been trained and learned, and calculating the gender information of the target object according to the probability mapping relationship between the multidimensional voiceprint feature and the gender information in the gender information recognition sub-model;

[0030] Inputting the multidimensional voiceprint feature into the age information recognition sub-model that has been trained, and calculating the age information of the target object according to the probability mapping relationship between the multidimensional voiceprint feature and the age information in the age information recognition sub-model;

[0031] Inputting the multidimensional voiceprint feature into the learned and trained accent information recognition sub-model, and calculating the accent information of the target object according to the probability mapping relationship between the multidimensional voiceprint feature and the accent information in the accent information recognition sub-model;

[0032] Inputting the multidimensional voiceprint features into the emotion information recognition sub-model that has been trained, and calculating the emotion information of the target object when speaking according to the probability mapping relationship between the multidimensional voiceprint features and the emotion information in the emotion information recognition sub-model;

[0033] The target object's gender information, age information, accent information and emotional information when speaking are output as the customer information.

[0034] Furthermore, after executing the step of inputting the multi-dimensional voiceprint feature into the pre-trained customer information recognition model, identifying the customer information corresponding to the target object and outputting it, the method further includes:

[0035] Determining whether business assistance is required for the target object based on a preset business assistance strategy and the customer information;

[0036] If business assistance is required, a business assistance instruction is sent to the target business personnel;

[0037] If business assistance is not needed, the real-time conversation voice of the target object continues to be acquired, and the emotional information of the target object when speaking is continuously calculated based on the real-time conversation voice.

[0038] In order to solve the above technical problems, the embodiment of the present application also provides an information identification device, which adopts the following technical solution:

[0039] An information identification device, comprising:

[0040] A speech acquisition module to be recognized, used to acquire the speech to be recognized of the target object;

[0041] An MFCC feature extraction module, used for inputting the speech to be recognized into a preset speech feature extraction model to extract the MFCC features contained in the speech to be recognized;

[0042] A voiceprint feature extraction module, used for inputting the MFCC features contained in the speech to be recognized into a pre-trained voiceprint feature extraction model, and extracting the multi-dimensional voiceprint features contained in the MFCC features;

[0043] The customer information recognition module is used to input the multi-dimensional voiceprint feature into the pre-trained customer information recognition model, identify the customer information corresponding to the target object and output it.

[0044] In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical solution:

[0045] A computer device includes a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of the above-mentioned information identification method when executing the computer-readable instructions.

[0046] In order to solve the above technical problems, the embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solution:

[0047] A computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the information identification method described above.

[0048] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0049] The information recognition method described in the embodiment of the present application obtains the speech to be recognized of the target object; inputs the speech to be recognized into a preset speech feature extraction model to extract the MFCC features contained in the speech to be recognized; inputs the MFCC features contained in the speech to be recognized into a pre-trained voiceprint feature extraction model to extract the multi-dimensional voiceprint features contained in the MFCC features; inputs the multi-dimensional voiceprint features into a pre-trained customer information recognition model to recognize the customer information corresponding to the target object and output it. The customer information of the target object is recognized through speech processing technology, such as age, gender, accent and emotional information when speaking. The information recognition method described in the present application is applied to the real-time voice scene of banking business processing, which can not only help the bank supervision terminal to understand the customer information of online customers in real time, but also make corresponding adjustments in time when the customer's emotions fluctuate when speaking, which is more intelligent and humane. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0051] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0052] Figure 2 is a flow chart of an embodiment of an information identification method according to the present application;

[0053] Figure 3 yes Figure 2 A flowchart of a specific embodiment of training the voiceprint feature extraction model using a training data set before step 203 is shown;

[0054] Figure 4 yes Figure 3 A flowchart of a specific embodiment of step 304 is shown;

[0055] Figure 5 yes Figure 2 A flowchart of a specific embodiment of performing learning and training on the customer information recognition model before step 204 is shown;

[0056] Figure 6 yes Figure 5 A flowchart of a specific embodiment of step 503 is shown;

[0057] Figure 7 yes Figure 2 A flowchart of a specific embodiment of step 204 is shown;

[0058] Figure 8 yes Figure 2 A flowchart of a specific embodiment of performing business auxiliary judgment after step 204 is shown;

[0059] Fig. 9 is a schematic structural diagram of an embodiment of an information identification device according to the present application;

[0060] Fig.10 It is a structural diagram of an embodiment of a computer device according to the present application. DETAILED DESCRIPTION

[0061] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0062] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0063] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0064] like Figure 1 As shown, the system architecture 100 may include a terminal device 101, a network 102 and a server 103. The terminal device 101 may be a laptop 1011, a tablet computer 1012 or a mobile phone 1013. The network 102 is used to provide a medium for a communication link between the terminal device 101 and the server 103. The network 102 may include various connection types, such as wired, wireless communication links or optical fiber cables.

[0065] The user can use the terminal device 101 to interact with the server 103 through the network 102 to receive or send messages, etc. Various communication client applications can be installed on the terminal device 101, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0066] The terminal device 101 can be various electronic devices with a display screen and supporting web browsing. In addition to the laptop computer 1011, the tablet computer 1012 or the mobile phone 1013, the terminal device 101 can also be an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 (Moving Picture Experts Group Audio Layer IV) player, a laptop computer and a desktop computer, etc.

[0067] The server 103 may be a server that provides various services, such as a background server that provides support for a web page displayed on the terminal device 101 .

[0068] It should be noted that the information identification method provided in the embodiment of the present application is generally executed by a terminal device, and accordingly, the information identification device is generally arranged in the terminal device.

[0069] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0070] Continue to refer Figure 2 , shows a flow chart of an embodiment of an information identification method according to the present application. The information identification method comprises the following steps:

[0071] Step 201, obtaining the speech to be recognized of the target object;

[0072] In this embodiment, the information recognition method can be applied to real-time voice scenarios in banking business processing. Accordingly, the acquisition of the target object's voice to be recognized includes acquiring the voice to be recognized of customers who are conducting business based on the online business processing terminals installed in the bank hall; the target object includes customers who are conducting business through the online business processing terminals, and the voice to be recognized includes voice commands issued by the customers.

[0073] Step 202, inputting the speech to be recognized into a preset speech feature extraction model to extract MFCC features contained in the speech to be recognized;

[0074] Specifically, the preset speech feature extraction model may be only a speech processing component that processes real-time speech into MFCC features, that is, it only includes: a fast Fourier transform component that converts the time domain signal of the real-time speech into a frequency domain signal, a Mel filter group that filters the spectrum corresponding to the converted frequency domain signal, a logarithm calculation component that takes the logarithm of the energy value obtained after processing by the Mel filter group, and a discrete cosine transform component that performs discrete cosine transformation on the logarithm to obtain the MFCC coefficient, and the MFCC coefficient is the MFCC feature.

[0075] Step 203, inputting the MFCC features contained in the speech to be recognized into a pre-trained voiceprint feature extraction model to extract the multi-dimensional voiceprint features contained in the MFCC features;

[0076] Specifically, the MFCC features are input into a pre-trained voiceprint feature extraction model to extract multi-dimensional voiceprint features contained in the MFCC features, such as voiceprint features of multiple dimensions such as tone, speaking speed, pitch, intensity, length, pause, and stress. By extracting the multi-dimensional voiceprint features contained in the MFCC features, customers can be distinguished from multiple dimensions such as speaking tone, speaking speed, pitch, intensity, length, pause, and stress.

[0077] Step 204: input the multi-dimensional voiceprint feature into a pre-trained customer information recognition model, identify the customer information corresponding to the target object, and output it.

[0078] Specifically, the multi-dimensional voiceprint features, such as tone, speaking speed, pitch, intensity, length, pause, stress, etc., are input into the pre-trained customer information recognition model to identify the customer information corresponding to the target object and output it. Here, "customer information" refers to customer information that can be embodied, such as gender, age, accent, and emotional information when speaking. By extracting multi-dimensional voiceprint features from speech and then identifying customer information through multi-dimensional voiceprint features, the bank supervision terminal can learn more about the customer information of the voice customer, and can also make timely adjustments when the customer's emotions fluctuate. For example, when a customer performs multiple operations through the online business terminal and the business is not successfully handled, and when a negative or irritable emotion appears, the business assistant at the bank supervision terminal is promptly reminded to help the customer and provide business handling guidance.

[0079] In this embodiment, the speech to be recognized of the target object is obtained; the speech to be recognized is input into a preset speech feature extraction model to extract the MFCC features contained in the speech to be recognized; the MFCC features contained in the speech to be recognized are input into a pre-trained voiceprint feature extraction model to extract the multi-dimensional voiceprint features contained in the MFCC features; the multi-dimensional voiceprint features are input into a pre-trained customer information recognition model to recognize the customer information corresponding to the target object and output it. The customer information of the target object is recognized through speech processing technology, such as age, gender, accent and emotional information when speaking. The information recognition method described in this application is applied to the real-time voice scene of banking business processing, which can not only help the bank supervision terminal to understand the customer information of online customers in real time, but also make corresponding adjustments in time when the customer's emotions fluctuate when speaking, which is more intelligent and humane.

[0080] Continue to refer Figure 3 In some optional implementations, before step 203, a step of training the voiceprint feature extraction model using a training data set is also included. Figure 3 yes Figure 2 The flowchart of a specific embodiment of using the training data set to train the voiceprint feature extraction model before step 203 includes the following steps:

[0081] Step 301, obtaining the annotated speech of a batch of users as a training data set, wherein the annotated speech indicators annotate the gender information, age information, accent information and emotional information of the corresponding user when speaking;

[0082] By obtaining the labeled speech of batches of users as the training data set, the trained voiceprint feature extraction model is based on the labeled data, ensuring the availability of the model.

[0083] Step 302, inputting the training data set into the preset speech feature extraction model to extract MFCC features corresponding to all users in the batch of users;

[0084] Step 303, constructing gender information, age information, accent information and emotional information when speaking corresponding to each of the MFCC features according to the preset user identifier;

[0085] Specifically, the preset user identifier can be understood as a user's distinguishing identifier, that is, distinguishing number information corresponding to each user.

[0086] Step 304: input the MFCC features corresponding to all users in the batch of users into the voiceprint feature extraction model to be trained, train the voiceprint feature extraction model to be trained, and obtain a pre-trained voiceprint feature extraction model.

[0087] By using the annotated speech of batch users as a training data set, MFCC feature extraction is performed, and then the voiceprint feature extraction model to be trained is trained using MFCC features, which ensures that the probability mapping relationship between multi-dimensional voiceprint features and gender information, age information, accent information and emotional information when speaking can be identified based on the gender information, age information, accent information and emotional information when speaking corresponding to each MFCC feature.

[0088] Continue to refer Figure 4 , Figure 4 yes Figure 3 The flowchart of a specific embodiment of step 304 shown includes the following steps:

[0089] Step 401, using the voiceprint feature extraction model to be trained to perform multi-dimensional voiceprint feature extraction on the MFCC features corresponding to all users in the batch of users;

[0090] Step 402: According to the user identifier, a multi-dimensional voiceprint feature corresponding to each of the MFCC features is determined to obtain a pre-trained voiceprint feature extraction model.

[0091] By performing multi-dimensional voiceprint feature extraction on the MFCC features corresponding to all users in the batch of users, the multi-dimensional voiceprint features corresponding to each of the MFCC features are determined, and a pre-trained voiceprint feature extraction model is obtained, thereby ensuring its usability in the subsequent actual real-time speech multi-dimensional voiceprint feature extraction.

[0092] Continue to refer Figure 5 In some optional implementations, before step 204, a step of learning and training the customer information recognition model is also included. Figure 5 yes Figure 2 The flowchart of a specific embodiment of learning and training the customer information recognition model before step 204 includes the following steps:

[0093] Step 501, based on the multidimensional voiceprint features corresponding to each of the MFCC features, and the gender information, age information, accent information, and emotional information when speaking corresponding to each of the MFCC features, a probability mapping relationship between the multidimensional voiceprint features and the gender information, age information, accent information, and emotional information when speaking is calculated;

[0094] Specifically, for example, the multi-dimensional voiceprint features corresponding to the MFCC features of user A, namely, tone, speaking speed, pitch, intensity, length, pause, and stress, all have corresponding representation values, and the gender information, age information, accent information, and emotional information when speaking corresponding to the MFCC features of user A are female, 26 years old, Sichuan accent, and stable emotions. By obtaining the multi-dimensional voiceprint features corresponding to batch users, and the gender information, age information, accent information, and emotional information when speaking corresponding to each of the MFCC features, since the number of batch users is a known value, the gender information of male and female is also marked. Information can be used to calculate the probability values ​​of different tones, speaking speeds, pitches, sound intensity, sound lengths, pauses, and stresses for males and females, respectively, that is, the probability mapping relationship between the multi-dimensional voiceprint features and gender information can be calculated. Similarly, different age groups can be set to calculate the probability mapping relationship between the multi-dimensional voiceprint features and different age groups. When the set age group is one year, the probability mapping relationship between the multi-dimensional voiceprint features and the age information is calculated. Of course, the probability mapping relationship between the multi-dimensional voiceprint features and the accent information and the emotional information when speaking can also be calculated in this way, which will not be described in detail.

[0095] Step 502, using the probability mapping relationship between the multi-dimensional voiceprint feature and gender information, the probability mapping relationship between the multi-dimensional voiceprint feature and age information, the probability mapping relationship between the multi-dimensional voiceprint feature and accent information, and the probability mapping relationship between the multi-dimensional voiceprint feature and emotional information during speaking as learning knowledge to construct a learning knowledge data set;

[0096] Step 503: input the learning knowledge data set into the customer information recognition model to be trained, perform learning training on the customer information recognition model to be trained, and obtain a pre-trained customer information recognition model.

[0097] In this embodiment, the customer information recognition model is trained by combining the labeled information, so that the customer information recognition model finally obtained after training is more convincing.

[0098] In this embodiment, the customer information recognition model to be trained includes a gender information recognition sub-model, an age information recognition sub-model, an accent information recognition sub-model and an emotion information recognition sub-model.

[0099] Continue to refer Figure 6 , Figure 6 yes Figure 5 The flowchart of a specific embodiment of step 503 shown includes the following steps:

[0100] Step 601, deploying the probability mapping relationship between the multi-dimensional voiceprint features and the gender information to the gender information recognition sub-model to obtain a gender information recognition sub-model that has been trained and learned;

[0101] Step 602, deploying the probability mapping relationship between the multi-dimensional voiceprint feature and the age information to the age information recognition sub-model to obtain the age information recognition sub-model that has been learned and trained;

[0102] Step 603, deploying the probability mapping relationship between the multi-dimensional voiceprint features and the accent information to the accent information recognition sub-model to obtain a learned and trained accent information recognition sub-model;

[0103] Step 604: deploy the probability mapping relationship between the multi-dimensional voiceprint features and the emotional information during speaking to the emotional information recognition sub-model to obtain the emotional information recognition sub-model that has been trained and learned.

[0104] By adopting machine learning methods, the learned knowledge is deployed to the gender information recognition sub-model, age information recognition sub-model, accent information recognition sub-model and emotion information recognition sub-model included in the customer information recognition model to be trained, so as to obtain the customer information recognition model that has been learned and trained. This enables the target object to be represented from multiple angles such as gender, age, accent and emotion when speaking in a more comprehensive way during actual use.

[0105] Continue to refer Figure 7 , Figure 7 yes Figure 2 The flowchart of a specific embodiment of step 204 shown includes the following steps:

[0106] Step 701, inputting the multidimensional voiceprint feature into the trained gender information recognition sub-model, and calculating the gender information of the target object according to the probability mapping relationship between the multidimensional voiceprint feature and the gender information in the gender information recognition sub-model;

[0107] Specifically, when calculating the gender information of the target object according to the probability mapping relationship between the multidimensional voiceprint features and the gender information in the gender information recognition sub-model, the result will have a male probability value and a female probability value, and the gender corresponding to the larger probability value is screened out as the gender information of the target object.

[0108] Step 702, inputting the multidimensional voiceprint feature into the age information recognition sub-model that has been trained, and calculating the age information of the target object according to the probability mapping relationship between the multidimensional voiceprint feature and the age information in the age information recognition sub-model;

[0109] Specifically, according to the probability mapping relationship between the multidimensional voiceprint features and the age information in the age information recognition sub-model, when calculating the age information of the target object, there will be probability values ​​corresponding to multiple ages. Afterwards, the probability values ​​corresponding to the multiple ages are sorted, and the age value corresponding to the maximum probability value is selected as the age information of the target object.

[0110] Step 703, inputting the multi-dimensional voiceprint feature into the trained accent information recognition sub-model, and calculating the accent information of the target object according to the probability mapping relationship between the multi-dimensional voiceprint feature and the accent information in the accent information recognition sub-model;

[0111] Step 704, inputting the multidimensional voiceprint feature into the emotion information recognition sub-model that has been trained, and calculating the emotion information of the target object when speaking according to the probability mapping relationship between the multidimensional voiceprint feature and the emotion information in the emotion information recognition sub-model;

[0112] Step 705: output the target object's gender information, age information, accent information, and emotional information when speaking as the customer information.

[0113] In the above steps 703 to 704, the maximum probability value method can also be used to identify the accent information of the target object and the emotional information of the target object when speaking, which will not be described in detail here.

[0114] By outputting the target object's gender information, age information, accent information and emotional information when speaking as the customer information, it is possible to combine the business assistance strategy within the bank, timely identify customers who need business assistance, and arrange business personnel to provide business guidance, which is more intelligent and humane.

[0115] Continue to refer Figure 8 In some optional implementations, after step 204, a step of performing business auxiliary judgment is also included. Figure 8 yes Figure 2The flowchart of a specific embodiment of performing business auxiliary judgment after step 204 shown includes the following steps:

[0116] Step 801, judging whether business assistance is required for the target object according to a preset business assistance strategy and the customer information;

[0117] Specifically, the preset business assistance strategy, for example: setting an age threshold, customers who exceed the age threshold are customers who need business assistance, to prevent old customers from not being able to operate the smart business processing terminals installed in the bank; for example, setting accent information recognition, if the target object is identified and it is found that the accent information is heavy and it is difficult to hear Mandarin when listening to the target object speak, the business personnel in his hometown can be selected to provide business assistance based on the target object's accent information, which is more intelligent and humane, and improves the customer's service experience.

[0118] Step 802: If business assistance is required, a business assistance instruction is sent to the target business personnel;

[0119] Step 803: If business assistance is not required, continue to obtain the real-time conversation voice of the target object, and continue to calculate the emotional information of the target object when speaking based on the real-time conversation voice.

[0120] By continuing to obtain the real-time conversation voice of the target object and continuously calculating the emotional information of the target object when speaking based on the real-time conversation voice, timely adjustments can be made when the target object has emotional fluctuations, ensuring that business handling problems of customers are handled in a timely manner.

[0121] This application obtains the speech to be recognized of the target object; inputs the speech to be recognized into a preset speech feature extraction model to extract the MFCC features contained in the speech to be recognized; inputs the MFCC features contained in the speech to be recognized into a pre-trained voiceprint feature extraction model to extract the multi-dimensional voiceprint features contained in the MFCC features; inputs the multi-dimensional voiceprint features into a pre-trained customer information recognition model to recognize and output the customer information corresponding to the target object. The customer information of the target object is recognized through speech processing technology, such as age, gender, accent, and emotional information when speaking. The information recognition method described in this application is applied to the real-time voice scene of banking business processing, which can not only help the bank supervision terminal to understand the customer information of online customers in real time, but also make corresponding adjustments in time when the customer's emotions fluctuate when speaking, which is more intelligent and humane.

[0122] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial Intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0123] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0124] In the embodiment of the present application, the speech to be recognized of the target object is obtained; the speech to be recognized is input into a preset speech feature extraction model to extract the MFCC features contained in the speech to be recognized; the MFCC features contained in the speech to be recognized are input into a pre-trained voiceprint feature extraction model to extract the multi-dimensional voiceprint features contained in the MFCC features; the multi-dimensional voiceprint features are input into a pre-trained customer information recognition model to recognize the customer information corresponding to the target object and output it. The customer information of the target object is recognized through speech processing technology, such as age, gender, accent and emotional information when speaking. The information recognition method described in the present application is applied to the real-time voice scene of banking business processing, which can not only help the bank supervision terminal to understand the customer information of online customers in real time, but also make corresponding adjustments in time when the customer's emotions fluctuate when speaking, which is more intelligent and humane.

[0125] Further references Fig. 9 , as a response to the above Figure 2 In order to realize the method shown in the figure, the present application provides an embodiment of an information identification device, and the device embodiment is Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0126] like Fig. 9 As shown, the information recognition device 900 described in this embodiment includes: a speech acquisition module 901 to be recognized, an MFCC feature extraction module 902, a voiceprint feature extraction module 903 and a customer information recognition module 904. Among them:

[0127] The to-be-recognized speech acquisition module 901 is used to acquire the to-be-recognized speech of the target object;

[0128] The MFCC feature extraction module 902 is used to input the speech to be recognized into a preset speech feature extraction model to extract the MFCC features contained in the speech to be recognized;

[0129] The voiceprint feature extraction module 903 is used to input the MFCC features contained in the speech to be recognized into the pre-trained voiceprint feature extraction model to extract the multi-dimensional voiceprint features contained in the MFCC features;

[0130] The customer information recognition module 904 is used to input the multi-dimensional voiceprint feature into the pre-trained customer information recognition model, identify the customer information corresponding to the target object and output it.

[0131] This application obtains the speech to be recognized of the target object; inputs the speech to be recognized into a preset speech feature extraction model to extract the MFCC features contained in the speech to be recognized; inputs the MFCC features contained in the speech to be recognized into a pre-trained voiceprint feature extraction model to extract the multi-dimensional voiceprint features contained in the MFCC features; inputs the multi-dimensional voiceprint features into a pre-trained customer information recognition model to recognize and output the customer information corresponding to the target object. The customer information of the target object is recognized through speech processing technology, such as age, gender, accent, and emotional information when speaking. The information recognition method described in this application is applied to the real-time voice scene of banking business processing, which can not only help the bank supervision terminal to understand the customer information of online customers in real time, but also make corresponding adjustments in time when the customer's emotions fluctuate when speaking, which is more intelligent and humane.

[0132] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0133] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0134] To solve the above technical problems, the present application also provides a computer device. Fig.10 , Fig.10 This is a basic structural block diagram of the computer device in this embodiment.

[0135] The computer device 10 includes a memory 10a, a processor 10b, and a network interface 10c which are interconnected and communicated via a system bus. It should be noted that: Fig.10 Only a computer device 10 having a component memory 10a, a processor 10b, and a network interface 10c is shown, but it should be understood that it is not required to implement all the components shown, and more or fewer components may be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to a microprocessor, an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a digital processor (DSP), an embedded device, etc.

[0136] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0137] The memory 10a includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 10a can be an internal storage unit of the computer device 10, such as a hard disk or memory of the computer device 10. In other embodiments, the memory 10a can also be an external storage device of the computer device 10, such as a plug-in hard disk equipped on the computer device 10, a smart memory card (Smart Med ia Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. Of course, the memory 10a can also include both the internal storage unit of the computer device 10 and its external storage device. In this embodiment, the memory 10a is generally used to store an operating system and various application software installed on the computer device 10, such as a computer-readable instruction of an information recognition method, etc. In addition, the memory 10a can also be used to temporarily store various data that have been output or will be output.

[0138] The processor 10b may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments. The processor 10b is generally used to control the overall operation of the computer device 10. In this embodiment, the processor 10b is used to run the computer-readable instructions stored in the memory 10a or process data, such as computer-readable instructions for running the information recognition method.

[0139] The network interface 10c may include a wireless network interface or a wired network interface. The network interface 10c is generally used to establish a communication connection between the computer device 10 and other electronic devices.

[0140] The computer device proposed in this embodiment belongs to the field of real-time voice processing technology and is applied to real-time voice scenarios in banking business. This application obtains the speech to be recognized by the target object; inputs the speech to be recognized into a preset voice feature extraction model to extract the MFCC features contained in the speech to be recognized; inputs the MFCC features contained in the speech to be recognized into a pre-trained voiceprint feature extraction model to extract the multi-dimensional voiceprint features contained in the MFCC features; inputs the multi-dimensional voiceprint features into a pre-trained customer information recognition model to recognize the customer information corresponding to the target object and output it. The customer information of the target object is recognized by voice processing technology, such as age, gender, accent and emotional information when speaking. The information recognition method described in this application is applied to the real-time voice scenario of banking business, which can not only help the bank supervision terminal to understand the customer information of online customers in real time, but also make corresponding adjustments in time when the customer's emotions fluctuate when speaking, which is more intelligent and humane.

[0141] The present application also provides another implementation, namely, providing a computer-readable storage medium, wherein the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions can be executed by a processor to enable the processor to perform the steps of the information identification method as described above.

[0142] The computer-readable storage medium proposed in this embodiment belongs to the field of real-time speech processing technology and is applied to real-time speech scenarios in banking business. This application obtains the speech to be recognized of the target object; inputs the speech to be recognized into a preset speech feature extraction model to extract the MFCC features contained in the speech to be recognized; inputs the MFCC features contained in the speech to be recognized into a pre-trained voiceprint feature extraction model to extract the multi-dimensional voiceprint features contained in the MFCC features; inputs the multi-dimensional voiceprint features into a pre-trained customer information recognition model to recognize the customer information corresponding to the target object and output it. The customer information of the target object is recognized through speech processing technology, such as age, gender, accent and emotional information when speaking. The information recognition method described in this application is applied to the real-time speech scenario of banking business, which can not only help the bank supervision terminal to understand the customer information of online customers in real time, but also make corresponding adjustments in time when the customer's emotions fluctuate when speaking, which is more intelligent and humane.

[0143] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0144] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. An information identification method, characterized in that: The steps include: Obtain the target object's speech to be recognized; Inputting the speech to be recognized into a preset speech feature extraction model to extract MFCC features contained in the speech to be recognized; Inputting the MFCC features contained in the speech to be recognized into a pre-trained voiceprint feature extraction model to extract the multi-dimensional voiceprint features contained in the MFCC features; The multi-dimensional voiceprint feature is input into a pre-trained customer information recognition model to identify and output the customer information corresponding to the target object.

2. The information identification method according to claim 1, characterized in that: Before executing the step of inputting the MFCC features contained in the speech to be recognized into the pre-trained voiceprint feature extraction model to extract the multi-dimensional voiceprint features contained in the MFCC features, the method further includes: Obtaining labeled speech of a batch of users as a training data set, wherein the labeled speech indicators are annotated with gender information, age information, accent information, and emotional information of the corresponding user when speaking; Inputting the training data set into the preset speech feature extraction model to extract MFCC features corresponding to all users in the batch of users respectively; According to the preset user identifier, construct the gender information, age information, accent information and emotional information when speaking corresponding to each of the MFCC features; The MFCC features corresponding to all users in the batch of users are respectively input into the voiceprint feature extraction model to be trained, and the voiceprint feature extraction model to be trained is trained to obtain a pre-trained voiceprint feature extraction model.

3. The information identification method according to claim 2, characterized in that: The step of inputting the MFCC features corresponding to all users in the batch of users into the voiceprint feature extraction model to be trained, training the voiceprint feature extraction model to be trained, and obtaining a pre-trained voiceprint feature extraction model specifically includes: Using the voiceprint feature extraction model to be trained to perform multi-dimensional voiceprint feature extraction on the MFCC features corresponding to all users in the batch of users; According to the user identifier, the multi-dimensional voiceprint features corresponding to each of the MFCC features are determined to obtain a pre-trained voiceprint feature extraction model.

4. The information identification method according to claim 3, characterized in that: Before executing the step of inputting the multi-dimensional voiceprint feature into a pre-trained customer information recognition model to identify and output the customer information corresponding to the target object, the method further includes: Based on the multidimensional voiceprint features corresponding to each of the MFCC features, and the gender information, age information, accent information, and emotional information when speaking corresponding to each of the MFCC features, a probability mapping relationship between the multidimensional voiceprint features and the gender information, age information, accent information, and emotional information when speaking is calculated; The probability mapping relationship between multi-dimensional voiceprint features and gender information, the probability mapping relationship between multi-dimensional voiceprint features and age information, the probability mapping relationship between multi-dimensional voiceprint features and accent information, and the probability mapping relationship between multi-dimensional voiceprint features and emotional information during speaking are used as learning knowledge to construct a learning knowledge data set; The learning knowledge data set is input into the customer information recognition model to be trained, and the customer information recognition model to be trained is trained to obtain a pre-trained customer information recognition model.

5. The information identification method according to claim 4, characterized in that: The customer information recognition model to be trained includes a gender information recognition sub-model, an age information recognition sub-model, an accent information recognition sub-model and an emotion information recognition sub-model. The step of inputting the learning knowledge data set into the customer information recognition model to be trained, performing learning training on the customer information recognition model to be trained, and obtaining a pre-trained customer information recognition model specifically includes: Deploy the probability mapping relationship between the multi-dimensional voiceprint features and the gender information to the gender information recognition sub-model to obtain a gender information recognition sub-model that has been learned and trained; Deploy the probability mapping relationship between the multi-dimensional voiceprint feature and the age information to the age information recognition sub-model to obtain the age information recognition sub-model that has been learned and trained; Deploy the probability mapping relationship between the multi-dimensional voiceprint features and the accent information to the accent information recognition sub-model to obtain a learned and trained accent information recognition sub-model; The probability mapping relationship between the multi-dimensional voiceprint features and the emotional information during speaking is deployed to the emotional information recognition sub-model to obtain the emotional information recognition sub-model that has been learned and trained.

6. The information identification method according to claim 5, characterized in that: The step of inputting the multi-dimensional voiceprint feature into a pre-trained customer information recognition model, identifying the customer information corresponding to the target object and outputting it specifically includes: Inputting the multidimensional voiceprint feature into the gender information recognition sub-model that has been trained and learned, and calculating the gender information of the target object according to the probability mapping relationship between the multidimensional voiceprint feature and the gender information in the gender information recognition sub-model; Inputting the multidimensional voiceprint feature into the age information recognition sub-model that has been trained, and calculating the age information of the target object according to the probability mapping relationship between the multidimensional voiceprint feature and the age information in the age information recognition sub-model; Inputting the multidimensional voiceprint feature into the learned and trained accent information recognition sub-model, and calculating the accent information of the target object according to the probability mapping relationship between the multidimensional voiceprint feature and the accent information in the accent information recognition sub-model; Inputting the multidimensional voiceprint features into the emotion information recognition sub-model that has been trained, and calculating the emotion information of the target object when speaking according to the probability mapping relationship between the multidimensional voiceprint features and the emotion information in the emotion information recognition sub-model; The target object's gender information, age information, accent information and emotional information when speaking are output as the customer information.

7. The information identification method according to claim 1 or 6, characterized in that: After executing the step of inputting the multi-dimensional voiceprint feature into the pre-trained customer information recognition model, identifying the customer information corresponding to the target object and outputting it, the method further includes: Determining whether business assistance is required for the target object based on a preset business assistance strategy and the customer information; If business assistance is required, a business assistance instruction is sent to the target business personnel; If business assistance is not needed, the real-time conversation voice of the target object continues to be acquired, and the emotional information of the target object when speaking is continuously calculated based on the real-time conversation voice.

8. An information recognition device, characterized in that: include: A speech acquisition module to be recognized, used to acquire the speech to be recognized of the target object; An MFCC feature extraction module, used for inputting the speech to be recognized into a preset speech feature extraction model to extract the MFCC features contained in the speech to be recognized; A voiceprint feature extraction module, used for inputting the MFCC features contained in the speech to be recognized into a pre-trained voiceprint feature extraction model, and extracting the multi-dimensional voiceprint features contained in the MFCC features; The customer information recognition module is used to input the multi-dimensional voiceprint feature into the pre-trained customer information recognition model, identify the customer information corresponding to the target object and output it.

9. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of the information identification method according to any one of claims 1 to 7 when executing the computer-readable instructions.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by a processor, the steps of the information identification method according to any one of claims 1 to 7 are implemented.

Images