Identity verification method, device and equipment and storage medium thereof

Abstract

The embodiment of the invention belongs to the technical field of financial science and technology, is applied to an identity verification scene of both parties of a claim settlement service dialogue, and relates to an identity verification method, device and equipment and a storage medium thereof. Carrying out merging processing to obtain merged voices corresponding to all the voice objects respectively; according to the combined voice, performing identity category prediction on all the voice objects to obtain an identity category prediction result; and performing identity authenticity verification on all the voice objects in combination with the identity category prediction result to obtain a verification result. The identity verification method is applied to a financial claim settlement service dialogue two-party identity verification scene, especially during insurance claim settlement, an artificial intelligence prediction model mode is adopted firstly, the identities of customer service and customers are predicted in combination with dialogue voices of the two parties, and then identity authenticity verification is carried out in combination with an identity prediction result; the security of the financial claim settlement service is ensured, and the risk control of the claim settlement service is improved.

Description

Technical Field

[0001] The present application relates to the field of financial technology, and is applied to identity authentication scenarios of both parties in a claims business dialogue, and in particular to an identity authentication method, apparatus, device and storage medium thereof. Background Art

[0002] When there are multiple speakers in an audio clip, and the audios of different speakers are mixed in a single channel, it is difficult to use traditional voiceprint recognition directly. To solve this problem, the traditional approach is to first separate the speakers and then perform voiceprint recognition based on this.

[0003] For example, when making insurance claims, the common speaker separation methods in the conversation between customer service and customers are mostly based on deep learning. However, such methods have the problem of generalization, and the accuracy of voiceprint recognition is highly dependent on the effect of speaker separation. In addition, some methods can achieve voice separation, but cannot determine whether the separated voice is the customer or the customer service, which in turn affects the subsequent voiceprint recognition and identity determination, and cannot control the risks of the claims business. Summary of the invention

[0004] The purpose of the embodiments of the present application is to propose an identity authentication method, device, equipment and storage medium thereof to solve the problem that when the existing speaker is separated, the subsequent voiceprint recognition and identity determination are affected, and the risk control for the claims business cannot be performed.

[0005] In order to solve the above technical problems, the embodiment of the present application provides an identity authentication method, which adopts the following technical solutions:

[0006] An identity verification method comprises the following steps:

[0007] Obtain all dialogue voice segments in the target mixed channel;

[0008] Merging all the dialogue voice segments to obtain merged voices corresponding to all voice objects;

[0009] According to the combined speech corresponding to all speech objects, the identity category of all speech objects is predicted to obtain the identity category prediction result;

[0010] The identity authenticity of all the voice objects is verified in combination with the identity category prediction result to obtain a verification result.

[0011] Furthermore, all the voice objects include customer service objects and client objects, and the step of merging all the conversation voice segments to obtain the merged voices corresponding to all the voice objects specifically includes:

[0012] Extracting acoustic feature parameters of all the dialogue speech segments respectively to extract the acoustic feature parameters of each speech segment;

[0013] Performing similarity calculation on the acoustic feature parameters corresponding to all the dialogue voice segments, to obtain similarity calculation results, wherein the Euclidean distance algorithm is used when performing the similarity calculation;

[0014] Based on the similarity calculation result and a preset merging metric algorithm, all the dialogue voice segments are merged to obtain a voice segment merging result;

[0015] According to the voice segment merging result, the merged voices corresponding to all the voice objects are obtained.

[0016] Furthermore, the step of performing identity category prediction on all voice objects according to the merged voices corresponding to all voice objects to obtain identity category prediction results specifically includes:

[0017] Using speech recognition technology, the conversational speech corresponding to all speech objects is processed into speech text, and the speech text corresponding to all speech objects is obtained;

[0018] The speech texts corresponding to all speech objects are input into a pre-trained identity prediction model to obtain predicted identities of all speech objects, wherein the identity prediction model includes a text semantic recognition model based on natural language.

[0019] Furthermore, before executing the step of inputting the speech texts corresponding to all speech objects into the pre-trained identity prediction model to obtain the predicted identities of all speech objects, the method further includes:

[0020] Acquire batch training samples, wherein the training samples include conversation speech texts between a customer service object and a client object;

[0021] Inputting the batch training samples into the natural language-based text semantic recognition model to perform model learning training to obtain a pre-trained identity prediction model, wherein the purpose of performing model learning training is to enable the pre-trained identity prediction model to predict the identity type of the target voice object based on the input voice text;

[0022] The step of inputting the speech texts corresponding to all speech objects into the pre-trained identity prediction model to obtain the predicted identities of all speech objects specifically includes:

[0023] Perform semantic analysis on the speech texts corresponding to all speech objects;

[0024] According to the semantic parsing results, the identity types of all speech objects are predicted.

[0025] Furthermore, the step of performing identity authenticity verification on all the voice objects in combination with the identity category prediction result to obtain the verification result specifically includes:

[0026] Filtering out voice objects whose identity category is a first target category from the identity category prediction results, wherein the first target category is a customer service category;

[0027] Obtaining the voiceprint features corresponding to the voice object of the customer service category;

[0028] Inputting the voiceprint feature into a preset first voiceprint comparison library for voiceprint recognition comparison, wherein the preset first voiceprint comparison library is a voiceprint comparison library established based on the customer service voiceprint feature;

[0029] If the voiceprint recognition comparison passes, then the voice object whose identity category is the second target category is screened out, wherein the second target category is the customer category;

[0030] Obtaining voiceprint features corresponding to the voice objects of the customer category;

[0031] Inputting the voiceprint feature into a preset second voiceprint comparison library for voiceprint recognition comparison, wherein the preset second voiceprint comparison library is a voiceprint comparison library established based on the voiceprint feature of the customer;

[0032] If the voiceprint recognition comparison passes, a prompt message indicating successful identity authentication is sent to the target risk control end.

[0033] Furthermore, after executing the step of inputting the voiceprint feature into a preset first voiceprint comparison library and performing voiceprint recognition comparison, the method further includes:

[0034] If the voiceprint recognition comparison fails, the voice object that is in dialogue with the voice object of the customer service category is identified, and the identity types of the two are exchanged;

[0035] Obtain the voiceprint features corresponding to the voice object of the customer service category after the identity type is swapped;

[0036] Inputting the voiceprint feature into the first voiceprint comparison library to perform voiceprint recognition comparison;

[0037] If the voiceprint recognition comparison still fails, a prompt message indicating identity authentication failure is sent to the target risk control terminal;

[0038] If the voiceprint recognition comparison passes, then the voiceprint features corresponding to the voice object of the customer category after the identity type is swapped are obtained;

[0039] The voiceprint feature is input into a preset second voiceprint comparison library for voiceprint recognition and comparison.

[0040] Furthermore, after executing the step of inputting the voiceprint feature into a preset second voiceprint comparison library and performing voiceprint recognition comparison, the method further includes:

[0041] If the voiceprint recognition comparison fails, a prompt is sent to the target service end that the corresponding voice object is a customer whose voiceprint feature has not been entered.

[0042] In order to solve the above technical problems, the embodiment of the present application also provides an identity authentication device, which adopts the following technical solution:

[0043] An identity verification device, comprising:

[0044] A dialogue voice segment acquisition module, used to acquire all dialogue voice segments in the target mixed channel;

[0045] A dialogue voice segment merging processing module is used to merge all the dialogue voice segments to obtain merged voices corresponding to all voice objects;

[0046] An identity category prediction module is used to predict the identity category of all voice objects according to the combined voices corresponding to all voice objects, and obtain an identity category prediction result;

[0047] The identity result verification module is used to verify the identity authenticity of all the voice objects in combination with the identity category prediction result to obtain a verification result.

[0048] 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:

[0049] 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 identity authentication method when executing the computer-readable instructions.

[0050] 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:

[0051] A computer-readable storage medium stores computer-readable instructions, which implement the steps of the identity authentication method described above when executed by a processor.

[0052] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0053] The identity authentication method described in the embodiment of the present application obtains all the dialogue voice segments in the target mixed channel; merges all the dialogue voice segments to obtain the merged voices corresponding to all the voice objects; predicts the identity category of all the voice objects based on the merged voices corresponding to all the voice objects to obtain the identity category prediction results; verifies the identity authenticity of all the voice objects in combination with the identity category prediction results to obtain the verification results. The identity authentication method described in the present application is applied to the identity authentication scenario of both parties in the financial claims business dialogue, especially in insurance claims. First, an artificial intelligence prediction model is used to predict the identity of the customer service and the customer in combination with the dialogue voices of both parties. Then, the identity authenticity is verified in combination with the identity prediction results, which ensures the security of the financial claims business and improves the risk control of the claims business. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] 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.

[0055] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0056] Figure 2 is a flow chart of an embodiment of an identity authentication method according to the present application;

[0057] Figure 3 yes Figure 2 A flowchart of a specific embodiment of step 202 is shown;

[0058] Figure 4 yes Figure 3 A flowchart of a specific embodiment of step 303 is shown;

[0059] Figure 5 yes Figure 2 A flowchart of a specific embodiment of step 203 is shown;

[0060] Figure 6 It is a flowchart of a specific embodiment of pre-training an identity prediction model in the identity authentication method described in the present application;

[0061] Figure 7 yes Figure 5 A flowchart of a specific embodiment of step 502 is shown;

[0062] Figure 8 yes Figure 2 A flowchart of a specific embodiment of step 204 is shown;

[0063] Fig. 9 It is a flowchart of a specific embodiment of pre-training an identity prediction model in the identity authentication method described in the present application;

[0064] Fig.10 is a schematic structural diagram of an embodiment of an identity verification device according to the present application;

[0065] Fig.11 It is a structural diagram of an embodiment of a computer device according to the present application. DETAILED DESCRIPTION

[0066] 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.

[0067] 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.

[0068] 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.

[0069] 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.

[0070] 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.

[0071] The terminal device 101 can be any electronic device with a display screen and supporting web browsing. In addition to a laptop computer 1011, a tablet computer 1012 or a 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), a laptop computer, a desktop computer, etc.

[0072] 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 .

[0073] It should be noted that the identity authentication method provided in the embodiment of the present application is generally executed by a server, and accordingly, the identity authentication device is generally set in the server.

[0074] 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.

[0075] Continue to refer Figure 2 , shows a flow chart of an embodiment of an identity authentication method according to the present application. The identity authentication method comprises the following steps:

[0076] Step 201: Acquire all dialogue voice segments in the target mixed channel.

[0077] In this embodiment, the target mixed channel includes a voice output channel shared by the conversation voice between the customer service and the customer. The target mixed channel may include voice mixing between mono channels according to the different conversation microphones used during the conversation. If the microphones of the conversation devices are all mono channels when the customer service and the customer are talking, the customer service conversation content and the customer conversation content corresponding to the two mono channels are obtained and output through one voice output channel. Of course, it also includes voice mixing between stereo channels. If the microphones of the conversation devices are all stereo channels when the customer service and the customer are talking, the corresponding voice content of the two stereo channels are obtained. The customer service conversation content and the customer service conversation content are output through a voice output channel, wherein the stereo channel includes at least a left channel and a right channel, the left channel and the right channel respectively obtain voice signals, and then output the voice content. When outputting, the customer service conversation content and the customer service conversation content are output through a voice output channel; of course, it also includes voice mixing between stereo channels and mono channels. Assuming that when the customer service staff is having a conversation with the customer, one microphone of the conversation device is a stereo channel and the other is a mono channel, then the customer service conversation content and the customer conversation content corresponding to the two channels are obtained and output through a voice output channel.

[0078] It should be understood that the above-mentioned mixing not only refers to the mixing of left and right channels to form stereo sound, but also includes the mixed output of speech output to different speakers.

[0079] In actual conversation scenarios, the two parties will not directly state "Hello, I am a customer" or "Hello, I am a customer service representative" during the conversation. Moreover, even if there is a direct statement, when obtaining all the conversation voice segments in the target mixed channel, it is not strictly followed that questions and answers are alternated one by one. There will also be a mixture of conversation segments. It can be understood that the output voice segment may be the voice of the customer service representative or the voice of the customer. This leads to the need for identity authentication work in the speech recognition and intent recognition stages.

[0080] The identity authentication method described in this application is applied to the identity authentication scenario of both parties in the financial claims business dialogue, especially in insurance claims. When the voice clips are obtained, all voice object identities can be authenticated first, thereby ensuring that only the customer's intentions are identified separately in the future, or only the customer service response content is analyzed and identified separately, ensuring that business separation processing is carried out according to the different identities of the interlocutors in the future.

[0081] Step 202: merge all the dialogue voice segments to obtain merged voices corresponding to all voice objects.

[0082] In this embodiment, by merging all the conversation voice segments, the merged voices corresponding to all the voice objects are obtained, that is, the output voices in a mixed state are merged and sorted, that is, all the conversation voice segments issued by customer service are sorted together, and all the conversation voices issued by customers are sorted together.

[0083] Specifically, when all the dialogue voice segments are merged to obtain the merged voices corresponding to all voice objects, they can be sorted according to the dialogue order corresponding to the dialogue voice segments to ensure faster business intent recognition during subsequent business intent recognition.

[0084] Step 203: perform identity category prediction on all voice objects according to the merged voices corresponding to all voice objects, and obtain identity category prediction results.

[0085] Through the merged voices corresponding to all voice objects, identity category prediction is performed on all voice objects to obtain identity category prediction results, so as to facilitate subsequent business separation processing according to the different identities of the interlocutors, and also help to quickly identify the business needs of customers in insurance claims business scenarios.

[0086] Step 204: verify the identity authenticity of all the voice objects in combination with the identity category prediction result to obtain a verification result.

[0087] In this embodiment, the identity authenticity of all the voice objects is verified in combination with the identity category prediction result to obtain a verification result. The purpose is to identify the identities of both parties in the conversation. Especially in insurance claims business, by performing identity verification, the risk of insurance fraud can be reduced to a certain extent, and the reliability and security of the business can be improved.

[0088] In this embodiment, all dialogue voice segments in the target mixed channel are obtained; all dialogue voice segments are merged to obtain the merged voices corresponding to all voice objects; identity categories are predicted for all voice objects based on the merged voices corresponding to all voice objects to obtain identity category prediction results; identity authenticity is verified for all voice objects in combination with the identity category prediction results to obtain verification results. The identity authentication method described in this application is applied to the identity authentication scenario of both parties in the financial claims business dialogue, especially in insurance claims. First, an artificial intelligence prediction model is used to predict the identity of the customer service and the customer in combination with the dialogue voices of both parties. Then, the identity authenticity is verified in combination with the identity prediction results, which ensures the security of the financial claims business and improves the risk control of the claims business.

[0089] In this embodiment, all voice objects include customer service objects and customer objects.

[0090] Continue to refer Figure 3 , Figure 3 yes Figure 2 The flowchart of a specific embodiment of step 202 shown includes the following steps:

[0091] Step 301, extracting acoustic feature parameters of all the dialogue speech segments respectively, and extracting acoustic feature parameters of each speech segment;

[0092] Specifically, the acoustic feature parameters include Mel-spectrogram acoustic feature parameters.

[0093] Step 302, performing similarity calculation on the acoustic feature parameters corresponding to all the dialogue voice segments, to obtain a similarity calculation result, wherein the Euclidean distance algorithm is used when performing the similarity calculation;

[0094] Specifically, assuming that after the Mel-spectrogram acoustic feature parameters are obtained, it should be understood that a series of acoustic feature parameters are obtained for each dialogue voice segment, and according to the different dimensions of the acoustic feature parameters, the Euclidean distance algorithm is used to calculate the similarity of the acoustic feature parameters between each two dialogue voice segments.

[0095] Step 303, based on the similarity calculation result and a preset merging metric algorithm, merging all the dialogue voice segments to obtain a voice segment merging result;

[0096] Step 304: obtaining the merged speech corresponding to all the speech objects according to the speech segment merging result.

[0097] In this embodiment, the acoustic feature parameters of all dialogue voice segments are extracted, and then the similarity of the acoustic feature parameters is calculated, and a preset merging metric algorithm is used to merge all the dialogue voice segments to obtain a voice segment merging result, thereby realizing the merging and sorting of voice segments of different voice objects through acoustic feature parameters.

[0098] Continue to refer Figure 4 , Figure 4 yes Figure 3 The flowchart of a specific embodiment of step 303 shown includes the following steps:

[0099] Step 401, selecting two speech segments with the highest similarity from the similarity calculation results;

[0100] Step 402, judging whether to merge the two speech segments according to the Bayesian information criterion and a preset score threshold;

[0101] Step 403: if the two voice segments need to be merged, a merging operation is performed, and the latest voice segment after the merging process is obtained, and acoustic feature parameters of the latest voice segment are re-extracted;

[0102] Step 404, calculating the similarity between the re-extracted acoustic feature parameters and the acoustic feature parameters corresponding to other speech segments, and updating the similarity calculation result;

[0103] In this embodiment, after each merging operation is performed, acoustic feature parameter extraction and similarity recalculation are performed again, so as to facilitate merging in combination with the re-obtained similarity calculation results.

[0104] Step 405, iteratively execute steps 401 to 404 until there is no need to merge the two voice segments, then the merging is completed, and the merged voices corresponding to all voice objects are obtained.

[0105] In this embodiment, the step of determining whether to merge the two speech segments according to the Bayesian information criterion and a preset score threshold specifically includes:

[0106] According to the Bayesian Information Criterion:

[0107] BIC = -2log(L) + klog(n),

[0108] Calculate the BIC value, where L represents the maximum likelihood estimate when merging speech segments, k represents the number of all speech objects, and n represents the total number of all dialogue speech segments after the current merging process;

[0109] If the BIC value does not exceed the score threshold, performing merging processing;

[0110] If the BIC value exceeds the score threshold, the merging process is completed.

[0111] Specifically, the Bayesian information criterion and iterative merging method are used to merge the dialogue voice segments, ensuring the scientific nature of the voice segment merging.

[0112] It should be understood that, in general, it is a single customer service representative and a single customer, such as a car insurance customer, that is, a claims customer service representative and an insurance policyholder. In this case, the k value is 2. However, it does not exclude the case of a single customer service representative and multiple customers, such as an accident insurance customer, where the insurance beneficiaries include multiple people. When making a call, there is only one customer service representative, but there may be multiple customer service personnel communicating. In this case, the corresponding conversation voice clips are conversation voice clips of multiple people. Assuming that the number of customer personnel communicating is 3, the k value is 4.

[0113] Continue to refer Figure 5, Figure 5 yes Figure 2 The flowchart of a specific embodiment of step 203 shown includes the following steps:

[0114] Step 501, using speech recognition technology, respectively converting the conversational speech corresponding to all speech objects into text, and obtaining speech texts corresponding to all speech objects;

[0115] Step 502, input the speech texts corresponding to all speech objects into the pre-trained identity prediction model to obtain the predicted identities of all the speech objects, wherein the identity prediction model includes a text semantic recognition model based on natural language.

[0116] Specifically, when performing identity prediction for all voice objects, the merged voice corresponding to different voice objects is first converted into text, and then the identity prediction is performed based on the pre-trained identity prediction model. An artificial intelligence prediction model is used, which is more intelligent and automated, and combined with text semantic recognition to ensure the accuracy of the prediction.

[0117] Continue to refer Figure 6 In some optional implementations, a step of pre-training the identity prediction model is also included before step 502. Figure 6 It is a flowchart of a specific embodiment of pre-training the identity prediction model in the identity authentication method described in the present application, comprising the following steps:

[0118] Step 601, obtaining batch training samples, wherein the training samples include conversation speech texts between a customer service object and a customer object;

[0119] Step 602: Input the batch training samples into the natural language-based text semantic recognition model to perform model learning training to obtain a pre-trained identity prediction model, wherein the purpose of performing model learning training is to enable the pre-trained identity prediction model to predict the identity type of the target voice object based on the input voice text.

[0120] Specifically, a machine learning method is adopted to pre-train the identity prediction model, and batch training samples are input, in which the conversation voice text corresponding to the customer service object and the conversation voice text corresponding to the customer object are respectively marked. Through machine learning training, the training model can extract the text semantic content corresponding to the customer service object for identity identification, as well as the text semantic content corresponding to the customer object for identity identification, to ensure that in subsequent actual predictions, only natural language understanding and semantic recognition are required based on the input voice text content to preliminarily predict the identity type of all voice objects.

[0121] Continue to refer Figure 7 , Figure 7 yes Figure 5 The flowchart of a specific embodiment of step 502 shown includes the following steps:

[0122] Step 701, semantically analyzing the speech texts corresponding to all speech objects;

[0123] Step 702: predict the identity types of all speech objects based on the semantic analysis results.

[0124] Specifically, by acquiring batches of conversational voice texts of customer service objects and customer objects, we train the identity prediction model so that we can predict the identity types of all voice objects based on the semantic analysis results of the voice texts. This ensures that in the claims business scenario, the identity types of customer service and customers can be identified based on the conversational voice content between them.

[0125] Continue to refer Figure 8 , Figure 8 yes Figure 2 The flowchart of a specific embodiment of step 204 shown includes the following steps:

[0126] Step 801, filtering out voice objects whose identity category is a first target category from the identity category prediction results, wherein the first target category is a customer service category;

[0127] Step 802, obtaining the voiceprint feature corresponding to the voice object of the customer service category;

[0128] Specifically, the acoustic feature parameters corresponding to the voice object in the above steps can be set as the voiceprint features, which can be used directly without the need for secondary extraction.

[0129] Step 803: input the voiceprint feature into a preset first voiceprint comparison library to perform voiceprint recognition comparison, wherein the preset first voiceprint comparison library is a voiceprint comparison library established based on the customer service voiceprint feature;

[0130] Step 804: if the voiceprint recognition comparison passes, then filter out the voice object whose identity category is the second target category, wherein the second target category is the customer category;

[0131] Step 805, obtaining the voiceprint features corresponding to the voice object of the customer category;

[0132] Step 806, inputting the voiceprint feature into a preset second voiceprint comparison library to perform voiceprint recognition comparison, wherein the preset second voiceprint comparison library is a voiceprint comparison library established based on the voiceprint feature of the customer;

[0133] Step 807: If the voiceprint recognition comparison passes, a prompt message indicating successful identity authentication is sent to the target risk control terminal.

[0134] Specifically, based on the preliminary prediction results of the identities of the two parties in the conversation, identity authentication is performed from the corresponding voiceprint feature comparison library, ensuring the identity authentication of customer service and the identity authentication of the customer, further ensuring the security of financial claims business and improving the risk control of claims business.

[0135] Continue to refer Fig. 9 In some optional implementations, after step 803, a step of pre-training the identity prediction model is also included. Fig. 9 It is a flowchart of a specific embodiment of pre-training the identity prediction model in the identity authentication method described in the present application, comprising the following steps:

[0136] Step 901, if the voiceprint recognition comparison fails, then identify the voice object that is in dialogue with the voice object of the customer service category, and exchange the identity types of the two;

[0137] Step 902, obtaining the voiceprint features corresponding to the voice object of the customer service category after the identity type is exchanged;

[0138] Step 903, inputting the voiceprint feature into the first voiceprint comparison library to perform voiceprint recognition comparison;

[0139] Step 904: If the voiceprint recognition comparison still fails, a prompt message indicating identity authentication failure is sent to the target risk control terminal;

[0140] Step 905, if the voiceprint recognition comparison passes, then obtain the voiceprint feature corresponding to the voice object of the customer category after the identity type is exchanged;

[0141] Step 906: input the voiceprint feature into a preset second voiceprint comparison library to perform voiceprint recognition comparison.

[0142] Specifically, if after the first voiceprint comparison library comparison, the preliminary identity prediction result has a prediction error, that is, the customer's identity is predicted as customer service, at this time, the identity type is swapped and then the voiceprint comparison is performed. It can be understood as adding a bottom-line verification, which further ensures the prudence and meticulousness of identity authentication, ensures the security of financial claims business, and improves the risk control of claims business.

[0143] In this embodiment, after executing the step of inputting the voiceprint feature into the preset second voiceprint comparison library and performing voiceprint recognition comparison, the method further includes: if the voiceprint recognition comparison fails, sending a prompt to the target service end that the corresponding voice object is a customer whose voiceprint feature has not been recorded, so as to facilitate the subsequent customer identity verification by the target service end, and update the voiceprint feature of the customer whose voiceprint feature has not been recorded to the second voiceprint comparison library.

[0144] It should be understood that this step corresponds to the situation where the customer service verification has passed, but the customer verification has not passed, that is, there is a case of intentional harassment, and there is also a case where the customer has not yet entered the voiceprint feature. Therefore, a prompt is sent to the target business end that the corresponding voice object is a customer whose voiceprint feature has not been entered, so that the target business end can perform identity verification and voiceprint feature entry later.

[0145] This application obtains all the dialogue voice segments in the target mixed channel; merges all the dialogue voice segments to obtain the merged voice corresponding to all the voice objects; predicts the identity category of all the voice objects based on the merged voice corresponding to all the voice objects to obtain the identity category prediction result; verifies the identity authenticity of all the voice objects in combination with the identity category prediction result to obtain the verification result. The identity authentication method described in this application is applied to the identity authentication scenario of both parties in the financial claims business dialogue, especially in insurance claims. First, an artificial intelligence prediction model is used to predict the identity of the customer service and the customer in combination with the dialogue voice of both parties. Then, the identity authenticity is verified in combination with the identity prediction result, which ensures the security of the financial claims business and improves the risk control of the claims business.

[0146] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, 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.

[0147] 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.

[0148] In an embodiment of the present application, all dialogue voice segments in the target mixed channel are obtained; all dialogue voice segments are merged to obtain the merged voices corresponding to all voice objects; identity categories are predicted for all voice objects based on the merged voices corresponding to all voice objects to obtain identity category prediction results; identity authenticity is verified for all voice objects in combination with the identity category prediction results to obtain verification results. The identity authentication method described in the present application is applied to the identity authentication scenario of both parties in the financial claims business dialogue, especially in insurance claims. First, an artificial intelligence prediction model is used to predict the identities of the customer service and the customer in combination with the dialogue voices of both parties. Then, identity authenticity verification is performed in combination with the identity prediction results, thereby ensuring the security of the financial claims business and improving the risk control of the claims business.

[0149] Further references Fig.10 , as a response to the above Figure 2 The present application provides an embodiment of an identity verification device, and the device embodiment is similar to Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0150] like Fig.10 As shown, the identity authentication device 10 described in this embodiment includes: a conversation voice segment acquisition module 10a, a conversation voice segment merging processing module 10b, an identity category prediction module 10c and an identity result verification module 10d. Among them:

[0151] A dialogue voice segment acquisition module 10a, used to acquire all dialogue voice segments in the target mixed channel;

[0152] A dialogue voice segment merging processing module 10b is used to merge all the dialogue voice segments to obtain merged voices corresponding to all voice objects;

[0153] The identity category prediction module 10c is used to predict the identity category of all voice objects according to the combined voices corresponding to all voice objects, and obtain the identity category prediction result;

[0154] The identity verification module 10d is used to verify the identity authenticity of all the voice objects in combination with the identity category prediction results to obtain verification results.

[0155] This application obtains all the dialogue voice segments in the target mixed channel; merges all the dialogue voice segments to obtain the merged voice corresponding to all the voice objects; predicts the identity category of all the voice objects based on the merged voice corresponding to all the voice objects to obtain the identity category prediction result; verifies the identity authenticity of all the voice objects in combination with the identity category prediction result to obtain the verification result. The identity authentication method described in this application is applied to the identity authentication scenario of both parties in the financial claims business dialogue, especially in insurance claims. First, an artificial intelligence prediction model is used to predict the identity of the customer service and the customer in combination with the dialogue voice of both parties. Then, the identity authenticity is verified in combination with the identity prediction result, which ensures the security of the financial claims business and improves the risk control of the claims business.

[0156] 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).

[0157] 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.

[0158] To solve the above technical problems, the present application also provides a computer device. Fig.11 , Fig.11 This is a basic structural block diagram of the computer device in this embodiment.

[0159] The computer device 11 includes a memory 11a, a processor 11b, and a network interface 11c which are interconnected through a system bus. It should be noted that: Fig.11Only a computer device 11 having a component memory 11a, a processor 11b, and a network interface 11c 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.

[0160] 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.

[0161] The memory 11a 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, disk, optical disk, etc. In some embodiments, the memory 11a can be an internal storage unit of the computer device 11, such as a hard disk or memory of the computer device 11. In other embodiments, the memory 11a can also be an external storage device of the computer device 11, such as a plug-in hard disk equipped on the computer device 11, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. Of course, the memory 11a can also include both the internal storage unit of the computer device 11 and its external storage device. In this embodiment, the memory 11a is generally used to store the operating system and various application software installed on the computer device 11, such as a computer-readable instruction of an identity authentication method, etc. In addition, the memory 11a can also be used to temporarily store various types of data that have been output or are to be output.

[0162] The processor 11b may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip in some embodiments. The processor 11b is generally used to control the overall operation of the computer device 11. In this embodiment, the processor 11b is used to run the computer-readable instructions stored in the memory 11a or process data, such as computer-readable instructions for running the identity authentication method.

[0163] The network interface 11c may include a wireless network interface or a wired network interface. The network interface 11c is generally used to establish a communication connection between the computer device 11 and other electronic devices.

[0164] The computer device proposed in this embodiment belongs to the field of financial technology and is applied to identity authentication scenarios of both parties in claims business conversations. This application obtains all conversation voice segments in the target mixed channel; merges all conversation voice segments to obtain merged voices corresponding to all voice objects; predicts the identity category of all voice objects based on the merged voices corresponding to all voice objects to obtain identity category prediction results; verifies the identity authenticity of all voice objects in combination with the identity category prediction results to obtain verification results. The identity authentication method described in this application is applied to the identity authentication scenario of both parties in financial claims business conversations, especially in insurance claims. First, an artificial intelligence prediction model is used to predict the identity of the customer service and the customer in combination with the conversation voices of both parties. Then, the identity authenticity is verified in combination with the identity prediction results, which ensures the security of the financial claims business and improves the risk control of the claims business.

[0165] 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 identity authentication method as described above.

[0166] The computer-readable storage medium proposed in this embodiment belongs to the field of financial technology and is applied to identity authentication scenarios of both parties in claims business conversations. This application obtains all conversation voice segments in the target mixed channel; merges all conversation voice segments to obtain merged voices corresponding to all voice objects; predicts the identity category of all voice objects based on the merged voices corresponding to all voice objects to obtain identity category prediction results; verifies the identity authenticity of all voice objects in combination with the identity category prediction results to obtain verification results. The identity authentication method described in this application is applied to the identity authentication scenario of both parties in financial claims business conversations, especially in insurance claims. First, an artificial intelligence prediction model is used to predict the identity of the customer service and the customer in combination with the conversation voices of both parties. Then, the identity authenticity is verified in combination with the identity prediction results, which ensures the security of the financial claims business and improves the risk control of the claims business.

[0167] 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.

[0168] 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 identity authentication method, characterized in that: The steps include: Obtain all dialogue voice segments in the target mixed channel; Merging all the dialogue voice segments to obtain merged voices corresponding to all voice objects; According to the combined speech corresponding to all speech objects, the identity category of all speech objects is predicted to obtain the identity category prediction result; The identity authenticity of all the voice objects is verified in combination with the identity category prediction result to obtain a verification result.

2. The identity authentication method according to claim 1, characterized in that: The all voice objects include customer service objects and customer objects, and the step of merging all the conversation voice segments to obtain the merged voices corresponding to all the voice objects specifically includes: Extracting acoustic feature parameters of all the dialogue speech segments respectively to extract the acoustic feature parameters of each speech segment; Performing similarity calculation on the acoustic feature parameters corresponding to all the dialogue voice segments, to obtain similarity calculation results, wherein the Euclidean distance algorithm is used when performing the similarity calculation; Based on the similarity calculation result and a preset merging metric algorithm, all the dialogue voice segments are merged to obtain a voice segment merging result; According to the voice segment merging result, the merged voices corresponding to all the voice objects are obtained.

3. The identity authentication method according to claim 1, characterized in that: The step of performing identity category prediction on all voice objects according to the combined voices corresponding to all voice objects to obtain identity category prediction results specifically includes: Using speech recognition technology, the conversational speech corresponding to all speech objects is processed into speech text, and the speech text corresponding to all speech objects is obtained; The speech texts corresponding to all speech objects are input into a pre-trained identity prediction model to obtain predicted identities of all speech objects, wherein the identity prediction model includes a text semantic recognition model based on natural language.

4. The identity authentication method according to claim 3, characterized in that: Before executing the step of inputting the speech texts corresponding to all the speech objects into the pre-trained identity prediction model to obtain the predicted identities of all the speech objects, the method further includes: Acquire batch training samples, wherein the training samples include conversation speech texts between a customer service object and a client object; Inputting the batch training samples into the natural language-based text semantic recognition model to perform model learning training to obtain a pre-trained identity prediction model, wherein the purpose of performing model learning training is to enable the pre-trained identity prediction model to predict the identity type of the target voice object based on the input voice text; The step of inputting the speech texts corresponding to all speech objects into the pre-trained identity prediction model to obtain the predicted identities of all speech objects specifically includes: Perform semantic analysis on the speech texts corresponding to all speech objects; According to the semantic parsing results, the identity types of all speech objects are predicted.

5. The identity authentication method according to claim 1, characterized in that: The step of performing identity authenticity verification on all the voice objects in combination with the identity category prediction result to obtain the verification result specifically includes: Filtering out voice objects whose identity category is a first target category from the identity category prediction results, wherein the first target category is a customer service category; Obtaining the voiceprint features corresponding to the voice object of the customer service category; Inputting the voiceprint feature into a preset first voiceprint comparison library for voiceprint recognition comparison, wherein the preset first voiceprint comparison library is a voiceprint comparison library established based on the customer service voiceprint feature; If the voiceprint recognition comparison passes, then the voice object whose identity category is the second target category is screened out, wherein the second target category is the customer category; Obtaining voiceprint features corresponding to the voice objects of the customer category; Inputting the voiceprint feature into a preset second voiceprint comparison library for voiceprint recognition comparison, wherein the preset second voiceprint comparison library is a voiceprint comparison library established based on the voiceprint feature of the customer; If the voiceprint recognition comparison passes, a prompt message indicating successful identity authentication is sent to the target risk control end.

6. The identity authentication method according to claim 5, characterized in that: After executing the step of inputting the voiceprint feature into a preset first voiceprint comparison library to perform voiceprint recognition comparison, the method further includes: If the voiceprint recognition comparison fails, the voice object that is in dialogue with the voice object of the customer service category is identified, and the identity types of the two are exchanged; Obtain the voiceprint features corresponding to the voice object of the customer service category after the identity type is swapped; Inputting the voiceprint feature into the first voiceprint comparison library to perform voiceprint recognition comparison; If the voiceprint recognition comparison still fails, a prompt message indicating identity authentication failure is sent to the target risk control terminal; If the voiceprint recognition comparison passes, then the voiceprint features corresponding to the voice object of the customer category after the identity type is swapped are obtained; The voiceprint feature is input into a preset second voiceprint comparison library for voiceprint recognition and comparison.

7. The identity authentication method according to claim 5 or 6, characterized in that: After executing the step of inputting the voiceprint feature into a preset second voiceprint comparison library to perform voiceprint recognition comparison, the method further includes: If the voiceprint recognition comparison fails, a prompt is sent to the target service end that the corresponding voice object is a customer whose voiceprint feature has not been entered.

8. An identity verification device, characterized in that: include: A dialogue voice segment acquisition module, used to acquire all dialogue voice segments in the target mixed channel; A dialogue voice segment merging processing module is used to merge all the dialogue voice segments to obtain merged voices corresponding to all voice objects; An identity category prediction module is used to predict the identity category of all voice objects according to the combined voices corresponding to all voice objects, and obtain an identity category prediction result; The identity result verification module is used to verify the identity authenticity of all the voice objects in combination with the identity category prediction result to obtain a verification result.

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 identity authentication method as described in 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, which, when executed by a processor, implement the steps of the identity authentication method according to any one of claims 1 to 7.

Images