Data verification method and device, electronic equipment and computer readable storage medium

Abstract

The application relates to the technical field of data processing, is applied to the financial technology and smart medical treatment scene, and provides a data verification method and device, an electronic device and a computer readable storage medium, the method comprises the following steps: obtaining an original voice signal; performing similarity calculation processing on a voiceprint feature and a user registered voiceprint to obtain voiceprint similarity information; performing physiological consistency determination processing on the voiceprint feature and a pathological sound biological feature to obtain a physiological authenticity score; performing structured extraction on a voice text corresponding to the original voice signal to obtain symptom description information; and performing medical consistency determination processing on the symptom description information and a pathological sound segment to obtain a medical authenticity score; and determining a verification result according to the voiceprint similarity information, the physiological authenticity score and the medical authenticity score. Through the technical scheme, the authenticity of the user identity can be verified, and it can be judged whether the user is in a real pathological state, so that the risk of fraud is reduced.

Description

Technical Field

[0001] The embodiments of this application relate to, but are not limited to, the field of data processing technology, and are applied to financial technology and smart healthcare scenarios. In particular, they relate to a data verification method, apparatus, electronic device, and computer-readable storage medium. Background Technology

[0002] With the continuous development of society and economy, and the continuous progress of science and technology, people's living standards are constantly improving. In order to better protect their lives, insurance business has been widely promoted and applied. Among them, in the telephone consultation and initial review of health insurance claims, users need to describe their symptoms verbally to support their application for hospitalization allowance or outpatient reimbursement. Similarly, when making an appointment by telephone, users also need to describe their symptoms verbally to apply for an appointment. However, the current application system mainly relies on voiceprint recognition to verify identity and uses simple keywords to match and judge the authenticity of the illness. However, this method can only verify the authenticity of the user's identity, but cannot accurately determine whether the user is actually in a pathological state, which poses a risk of deception. Summary of the Invention

[0003] The following is an overview of the subject matter described in detail herein. This overview is not intended to limit the scope of the claims.

[0004] To address the problems mentioned in the background section, this application provides a data verification method, apparatus, electronic device, and computer-readable storage medium, which can not only verify the authenticity of a user's identity but also determine whether the user is truly in a pathological state, effectively reducing the risk of deception.

[0005] In a first aspect, embodiments of this application provide a data verification method, the data verification method comprising: Acquire the raw speech signal; The original speech signal is segmented to obtain multiple speech segments; Multiple speech segments are subjected to sound event detection processing to obtain multiple pathological sound segments; For each of the pathological sound segments, feature extraction processing is performed to obtain voiceprint features and pathological acoustic biological features; and the voiceprint features are compared with preset user-registered voiceprints to obtain voiceprint similarity information. The voiceprint features and the pathological acoustic biological features are subjected to physiological consistency judgment processing to obtain a physiological authenticity score. The original speech signal is subjected to structured extraction of the corresponding speech text to obtain symptom description information; and the symptom description information and the pathological sound segment are subjected to medical consistency judgment processing to obtain a medical authenticity score. The verification result is determined based on the voiceprint similarity information, the physiological authenticity score, and the medical authenticity score.

[0006] Secondly, embodiments of this application also provide a data verification device, the data verification device comprising: The acquisition unit is used to acquire the raw speech signal; The segmentation unit is used to segment the original speech signal to obtain multiple speech segments; The detection unit is used to perform sound event detection processing on multiple speech segments to obtain multiple pathological sound segments; The execution unit is used to perform feature extraction processing on each of the pathological sound segments to obtain voiceprint features and pathological sound biological features; and to perform similarity calculation processing on the voiceprint features and the preset user-registered voiceprints to obtain voiceprint similarity information. The first determination unit is used to perform physiological consistency determination processing on the voiceprint features and the pathological acoustic biological features to obtain a physiological authenticity score. The second determination unit is used to perform structured extraction of the speech text corresponding to the original speech signal to obtain symptom description information; and to perform medical consistency determination processing on the symptom description information and the pathological sound segment to obtain a medical authenticity score. The calculation unit is used to determine the verification result based on the voiceprint similarity information, the physiological authenticity score, and the medical authenticity score.

[0007] Thirdly, embodiments of this application also provide an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the data verification method described in the first aspect above.

[0008] Fourthly, embodiments of this application also provide a computer-readable storage medium storing computer-executable instructions for performing the data verification method described in the first aspect above.

[0009] The data verification method according to the embodiments provided in this application has at least the following beneficial effects: In the data verification process, firstly, the original speech signal is acquired; then, the acquired original speech signal is segmented to obtain multiple speech segments; next, sound event detection processing is performed on the multiple speech segments to obtain multiple pathological sound segments; then, feature extraction processing is performed on each pathological sound segment to obtain voiceprint features and pathological acoustic biological features; furthermore, similarity calculation processing is performed between the voiceprint features and the preset user-registered voiceprint to obtain corresponding voiceprint similarity information; physiological consistency judgment processing is also performed on the voiceprint features and pathological acoustic biological features to obtain a physiological authenticity score; then, structured extraction processing is performed on the speech text corresponding to the original speech signal to obtain symptom description information; and finally, medical consistency judgment processing is performed on the symptom description information and the pathological sound segments to obtain a medical authenticity score; finally, the corresponding verification result can be determined based on the voiceprint similarity information, the physiological authenticity score, and the medical authenticity score. By using the above technical solution, relevant data is verified through three aspects: voiceprint similarity information, physiological authenticity score, and medical authenticity score. This not only verifies the authenticity of the user's identity but also determines whether the user is actually in a pathological state, effectively reducing the risk of deception. Attached Figure Description

[0010] The accompanying drawings are used to provide a further understanding of the technical solutions of this application and constitute a part of the specification. They are used together with the embodiments of this application to explain the technical solutions of this application and do not constitute a limitation on the technical solutions of this application.

[0011] Figure 1 This is a schematic diagram of an application environment for a data verification method in one embodiment of this application; Figure 2 This is a flowchart illustrating a data verification method provided in one embodiment of this application; Figure 3 yes Figure 2 A schematic diagram of a specific implementation method of step S200; Figure 4 yes Figure 2 A schematic diagram of a specific implementation of step S300; Figure 5 yes Figure 2 A schematic diagram of a specific implementation of step S500; Figure 6 yes Figure 2 A schematic diagram of a specific implementation method of step S600; Figure 7 yes Figure 2 A flowchart illustrating another specific implementation of step S600; Figure 8 yes Figure 2 A schematic diagram of a specific implementation of step S700; Figure 9 This is a schematic diagram of a data verification device provided in one embodiment of this application; Figure 10 This is a schematic diagram of an electronic device provided in one embodiment of this application. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0013] It should be noted that although functional modules are divided in the device schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first," "second," etc., in the specification, claims, and the aforementioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be noted that, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0014] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0015] AI is a new technical science that studies and develops theories, methods, technologies, and application systems for simulating, extending, and expanding human intelligence. Artificial intelligence is a branch of computer science that attempts to understand the essence of intelligence and produce new intelligent machines that can react in a way similar to human intelligence. Research in this field includes robotics, speech recognition, image recognition, natural language processing, and expert systems. Artificial intelligence can simulate the information processes of human consciousness and thought. Furthermore, artificial intelligence utilizes digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceiving the environment, acquiring knowledge, and using that knowledge to achieve optimal results—the theories, methods, technologies, and application systems available for use.

[0016] Foundational technologies for artificial intelligence generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies mainly encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0017] Artificial intelligence, or 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 that knowledge to obtain optimal results.

[0018] The servers involved in artificial intelligence technology can be standalone servers or cloud servers that provide basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0019] It should be noted that in all specific embodiments of this application, when processing data related to user identity or characteristics, such as user information, user behavior data, user historical data, and user location information, user permission or consent is obtained first. Furthermore, the collection, use, and processing of this data comply with relevant laws, regulations, and standards. In addition, when embodiments of this application require access to sensitive personal information of users, separate permission or consent from the user is obtained through pop-ups or redirection to confirmation pages. Only after obtaining the user's separate permission or consent is the necessary user-related data required for the proper functioning of these embodiments acquired.

[0020] The data verification method provided in this application embodiment can be applied to, for example, Figure 1In this application environment, the client communicates with the server via a network. The server can obtain the original speech signal from the client, then segment the obtained original speech signal to obtain multiple speech segments; then, it performs sound event detection processing on the multiple speech segments to obtain multiple pathological sound segments; then, it performs feature extraction processing on each pathological sound segment to obtain voiceprint features and pathological acoustic biological features; and then it performs similarity calculation processing between the voiceprint features and the preset user registration voiceprint to obtain the corresponding voiceprint similarity information; it also performs physiological consistency judgment processing on the voiceprint features and pathological acoustic biological features to obtain a physiological authenticity score; then, it performs structured extraction processing on the speech text corresponding to the original speech signal to obtain symptom description information; and then, it performs medical consistency judgment processing on the symptom description information and the pathological sound segments to obtain a medical authenticity score; finally, the corresponding verification result can be determined based on the voiceprint similarity information, physiological authenticity score, and medical authenticity score. The above technical solution verifies relevant data through three aspects: voiceprint similarity information, physiological authenticity score, and medical authenticity score. This not only verifies the authenticity of the user's identity but also determines whether the user is truly in a pathological state, effectively reducing the risk of deception. The client can be, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. The server can be implemented using a standalone server or a server cluster consisting of multiple servers. The following detailed description of specific embodiments of this application further illustrates the implementation details.

[0021] like Figure 2 As shown, Figure 2 This is a flowchart illustrating a data verification method provided in one embodiment of this application. The data verification method includes the following steps: Step S100: Acquire the raw speech signal.

[0022] The data verification method provided in this application first acquires the original voice signal during the data verification process. The acquired original voice signal is then segmented to obtain multiple voice segments. Specifically, in the field of smart healthcare, the original voice signal is generated when a patient registers for an appointment via voice using a smart medical robot, or when a patient makes an appointment by phone. Similarly, in the field of fintech, the original voice signal is generated when a user makes a claim through a smart claims robot, or during the initial review of health insurance claims via telephone consultation, where the user needs to describe symptoms verbally to support a hospitalization allowance or outpatient reimbursement application.

[0023] It is worth noting that during the acquisition of raw voice signals, when processing data related to user identity or characteristics, such as user information, user behavior data, user historical data, and user location information, user permission or consent is always obtained first. Furthermore, the collection, use, and processing of this data comply with relevant laws, regulations, and standards. In addition, when this application embodiment needs to obtain sensitive personal information of users, separate permission or consent from the user is obtained through pop-ups or redirects to confirmation pages. Only after explicitly obtaining the user's separate permission or consent is the necessary user-related data for the normal operation of this application embodiment acquired.

[0024] Step S200: Segment the original speech signal to obtain multiple speech segments.

[0025] The data verification method provided in this application embodiment can segment the original speech signal after obtaining it to obtain multiple speech segments, in order to prepare for subsequent sound event detection processing.

[0026] It is worth noting that in the process of segmenting the original speech signal to obtain multiple speech segments, the original speech signal is first divided into multiple consecutive short-time frames; then, short-time energy calculation is performed on each short-time frame to obtain the short-time energy, and zero-crossing detection is performed on each short-time frame to obtain the short-time zero-crossing rate; finally, the original speech signal can be segmented according to the short-time energy, the short-time zero-crossing rate, and the preset energy threshold and zero-crossing rate threshold to obtain multiple speech segments, which prepares for subsequent sound event detection processing.

[0027] like Figure 3 As shown, segmenting the original speech signal to obtain multiple speech segments can include the following steps: Step S210: Divide the original speech signal into multiple consecutive short time frames; Step S220: Perform short-time energy calculation processing on each short-time frame to obtain the short-time energy; and perform zero-crossing detection processing on each short-time frame to obtain the short-time zero crossover rate. Step S230: Based on short-time energy, short-time zero-crossing rate, and preset energy threshold and zero-crossing rate threshold, the original speech signal is segmented to obtain multiple speech segments.

[0028] For steps S210 to S230, in the process of segmenting the original speech signal to obtain multiple speech segments, the original speech signal is first divided into multiple consecutive short-time frames; then, short-time energy calculation is performed on each short-time frame to obtain the short-time energy, and zero-crossing detection is performed on each short-time frame to obtain the short-time zero-crossing rate; finally, the original speech signal can be segmented according to the short-time energy, the short-time zero-crossing rate, and preset energy thresholds and zero-crossing rate thresholds to obtain multiple speech segments. Through the above technical solution, the original speech signal can be segmented into multiple speech segments, allowing for subsequent sound event detection processing on these multiple speech segments, preparing for subsequent data verification.

[0029] It is worth noting that by dividing the original speech signal into multiple consecutive short frames, short-time energy calculation can be performed on each short frame to obtain the short-time energy; zero-crossing detection can be performed on each short frame to obtain the short-time zero-crossing rate; finally, based on the calculated short-time energy, short-time zero-crossing rate, and pre-set energy threshold and zero-crossing rate threshold, the original speech signal can be segmented to obtain multiple speech segments, thereby enabling better sound event detection processing of the original speech signal.

[0030] Step S300: Perform sound event detection processing on multiple speech segments to obtain multiple pathological sound segments.

[0031] The data verification method provided in this application segmentes the original speech signal to obtain multiple speech segments, and then performs sound event detection processing on the obtained multiple speech segments to obtain multiple pathological sound segments, providing data basis for subsequent calculation of physiological authenticity score and medical authenticity score.

[0032] It is worth noting that in the process of performing sound event detection processing on multiple speech segments to obtain multiple pathological sound segments, the process first performs pathological event probability prediction processing on multiple speech segments based on a pre-set sound event detection module, thereby obtaining multiple pathological event probabilities; then, based on a pre-set peak clustering algorithm, the multiple pathological event probabilities are processed to extract continuous probability intervals, thereby obtaining multiple pathological segments, where each pathological segment carries a confidence level; subsequently, pathological segments with a confidence level greater than a pre-set confidence level threshold can be used as the corresponding pathological sound segments, preparing for subsequent confirmation of voiceprint features and pathological acoustic biological features.

[0033] like Figure 4 As shown, performing sound event detection processing on multiple speech segments to obtain multiple pathological sound segments can include the following steps: Step S310: Based on the preset sound event detection module, perform pathological event probability prediction processing on multiple speech segments to obtain multiple pathological event probabilities; Step S320: Based on a preset peak clustering algorithm, the probabilities of multiple pathological events are extracted into a continuous probability interval to obtain multiple pathological segments, wherein each pathological segment carries a confidence level. Step S330: Pathological segments with a confidence level greater than a preset confidence level threshold are used as pathological sound segments.

[0034] For steps S310 to S330, in the process of performing sound event detection processing on multiple speech segments to obtain multiple pathological sound segments, firstly, based on a pre-set sound event detection module, the multiple speech segments are processed to predict the probability of pathological events, thereby obtaining multiple pathological event probabilities; then, based on a pre-set peak clustering algorithm, the multiple pathological event probabilities are processed to extract continuous probability intervals, thereby obtaining multiple pathological segments, where each pathological segment carries a confidence level; subsequently, pathological segments with a confidence level greater than a pre-set confidence level threshold can be used as the corresponding pathological sound segments; through the above technical solution, the corresponding pathological sound segments can be determined, preparing for subsequent data verification.

[0035] For example, in the field of smart healthcare, when a patient consults a smart healthcare robot about disease prevention measures and registers for an appointment, raw voice signals are generated. After the raw voice signals are segmented to obtain multiple voice segments, a pathological event probability prediction process can be performed on the multiple voice segments based on a pre-set sound event detection module to obtain multiple pathological event probabilities. Then, based on a pre-set peak clustering algorithm, continuous probability interval extraction processing is performed on the multiple pathological event probabilities to obtain multiple pathological segments. Finally, the pathological segments with a confidence level greater than a pre-set confidence level threshold can be used as pathological sound segments in the field of smart healthcare. Alternatively, in the fintech field, when users are using voice to describe their symptoms during the initial review of health insurance claims via telephone consultation, raw voice signals can be generated. After segmenting the raw voice signals into multiple voice segments, a pre-defined sound event detection module can be used to predict the probability of pathological events from these segments, resulting in multiple pathological event probabilities. Then, a pre-defined peak clustering algorithm is used to extract continuous probability intervals from these pathological event probabilities, thus obtaining multiple pathological segments. Finally, pathological segments with a confidence level greater than a pre-defined confidence threshold can be used as pathological voice segments in the fintech field.

[0036] It is worth noting that the confidence threshold in this application embodiment can be set according to actual needs. By flexibly setting the confidence threshold, the determination of pathological sound segments can be more flexible and accurate.

[0037] For example, in the process of anti-fraud during health insurance telephone consultations, the 16kHz telephone voice stream is segmented based on energy and zero crossover rate to obtain N voice segments {U1, U2, ..., U_N}; a multi-task sound event detection module is run in parallel, and based on an improved residual neural network architecture, it outputs the probability time series of three types of pathological events: cough (further subdivided into dry / wet cough), wheezing (inspiratory / expiratory), and nasal congestion; for each type of event, a peak clustering algorithm is used to extract continuous high-probability intervals to form a set of pathological segments P = {p1, p2, ..., p_M}, where each p_j contains start and end time, event type, and confidence level; if a voice segment U_i contains ≥2 high-confidence pathological segments (confidence level > 0.8), it is marked as a high-risk consultation segment.

[0038] Step S400: Perform feature extraction processing on each pathological sound segment to obtain voiceprint features and pathological sound biological features; and perform similarity calculation processing between the voiceprint features and the preset user-registered voiceprints to obtain voiceprint similarity information.

[0039] The data verification method provided in this application, after processing multiple speech segments to obtain multiple pathological speech segments through sound event detection, can then perform feature extraction processing on each pathological speech segment to obtain voiceprint features and pathological acoustic biological features. Next, the obtained voiceprint features and pre-defined user-registered voiceprints can be compared to calculate voiceprint similarity information. This voiceprint similarity information characterizes the similarity between the voiceprint of the speaker and the voiceprint of a registered user, used to determine the authenticity and reliability of the user's identity. Furthermore, physiological consistency determination can be performed based on the voiceprint features and pathological acoustic biological features.

[0040] It is worth noting that by performing feature extraction processing on each pathological sound segment, we can obtain voiceprint features and pathological acoustic biological features, which can prepare for the subsequent confirmation of voiceprint similarity information, physiological authenticity score and medical authenticity score.

[0041] It is worth noting that a user's registered voiceprint contains multiple registered voiceprints. Therefore, the voiceprint features obtained from feature extraction can be compared with each registered voiceprint in the user's registered voiceprint. The most similar value is then used as the voiceprint similarity information, which can prepare for subsequent identity verification and improve the accuracy of subsequent identity verification.

[0042] Step S500: Perform physiological consistency judgment processing on voiceprint features and pathological acoustic biological features to obtain physiological authenticity score.

[0043] The data verification method provided in this application, after performing feature extraction processing on each pathological sound segment to obtain voiceprint features and pathological acoustic biological features, can then perform physiological consistency judgment processing on the voiceprint features and pathological acoustic biological features to obtain a physiological authenticity score. The physiological authenticity score is used to characterize whether the voiceprint features and pathological acoustic biological features are physiologically consistent. The higher the physiological authenticity score, the more physiologically consistent the voiceprint features and pathological acoustic biological features are.

[0044] like Figure 5 As shown, the physiological consistency assessment of voiceprint features and pathological acoustic biological features to obtain a physiological authenticity score can include the following steps: Step S510: Construct joint embedded information from voiceprint features and pathological acoustic biological features; Step S520: Based on a preset binary classifier, the joint embedding information is subjected to consistency judgment processing to obtain the physiological authenticity score.

[0045] For steps S510 to S520, in the process of obtaining a physiological authenticity score by performing physiological consistency judgment on voiceprint features and pathological biological features, joint embedding information is first constructed from voiceprint features and pathological acoustic biological features; then, consistency judgment processing is performed on the joint embedding information based on a pre-set binary classifier to obtain a physiological authenticity score, so that subsequent data verification can be more accurate.

[0046] For example, in the process of anti-fraud during health insurance telephone consultations, for each pathological segment p_j, its voiceprint embedding e_s,j ∈ ^{512}, using the same ECAPA-TDNN architecture as the main voiceprint registration model; simultaneously extracting pathological acoustic biological features f_patho,j ∈ ^{128} includes 12 medical-related indicators such as cough duration, fundamental frequency jitter, harmonic-to-noise ratio, and centroid of wheezing spectrum; calculate the cosine similarity s_id,j between e_s,j and the user's registered voiceprint e_reg; if s_id,j<0.65, it is determined that the pathological voice is not made by the person; construct a joint embedding z_j = [e_s,j; f_patho,j], and use a binary classifier to determine whether "voiceprint-pathology" is physiologically consistent, and output the consistency score s_physio,j ∈ [0,1].

[0047] Step S600: Perform structured extraction on the speech text corresponding to the original speech signal to obtain symptom description information; and perform medical consistency judgment processing on the symptom description information and pathological sound segments to obtain a medical authenticity score.

[0048] The data verification method provided in this application extracts the structured text of the original speech signal to obtain symptom description information. Then, the symptom description information and pathological sound segments are subjected to medical consistency judgment processing to obtain a medical authenticity score. This can determine whether the speaker of the speech actually has the relevant symptoms, and thus can effectively judge whether the user is actually in a pathological state, effectively reducing the risk of deception.

[0049] It is worth noting that in the process of extracting symptom description information from the structured speech text corresponding to the original speech signal, the original speech signal is first processed by speech recognition and decoding to obtain the speech text; then the obtained speech text is processed by structured information extraction to obtain the corresponding symptom description information, which provides a data foundation for the subsequent determination of medical authenticity score.

[0050] It is worth noting that in the process of obtaining a medical authenticity score by performing medical consistency judgment on symptom description information and pathological sound segments, the symptom description information is first extracted to obtain medical symptom labels, and the pathological sound segments are feature extracted to obtain biomarker information. Then, the medical symptom labels and biomarker information are matched according to a pre-set medical consistency judgment model to determine the degree of logical consistency between the medical status labels and biomarker information, thereby obtaining a medical authenticity score and further improving the accuracy of data verification.

[0051] like Figure 6 As shown, the structured extraction of symptom description information from the original speech signal can include the following steps: Step S610: Perform speech recognition and decoding processing on the original speech signal to obtain the speech text; Step S620: Perform structured information extraction processing on the voice text to obtain symptom description information.

[0052] For steps S610 to S620, in the process of obtaining symptom description information by performing structured extraction processing on the speech text corresponding to the original speech signal, the original speech signal is first subjected to speech recognition and decoding processing to obtain speech text; then the speech text can be subjected to structured information extraction processing to obtain the corresponding symptom description information, which provides a basis for subsequent determination of medical authenticity score.

[0053] For example, in the field of smart healthcare, when a patient registers for an appointment by phone, the original voice signal is generated. This original voice signal can then undergo speech recognition and decoding to obtain speech-to-text. Subsequently, structured information extraction processing can be performed on the speech-to-text to obtain relevant symptom description information, providing a basis for subsequent intelligent registration. Alternatively, in the field of fintech, when a user describes symptoms during a health insurance phone consultation and initial claims review, the original voice signal is also generated. This original voice signal can then undergo speech recognition and decoding to obtain speech-to-text. Subsequently, structured information extraction processing can be performed on the speech-to-text to obtain relevant symptom description information, providing a basis for subsequent claims processing.

[0054] like Figure 7 As shown, the medical consistency assessment of symptom description information and pathological sound clips to obtain a medical authenticity score may include the following steps: Step S630: Extract and process the symptom description information to obtain medical symptom tags; and extract features from the pathological sound segments to obtain biomarker information; Step S640: Based on a preset medical consistency judgment model, the medical symptom labels and biomarker information are matched to determine the degree of logical consistency between the medical status labels and biomarker information, and a medical authenticity score is obtained.

[0055] For steps S630 to S640, in the process of performing medical consistency judgment processing on symptom description information and pathological sound segments to obtain a medical authenticity score, the symptom description information is first extracted to obtain medical symptom labels; and the pathological sound segments are feature extracted to obtain biomarker information. Then, the medical symptom labels and biomarker information are matched based on a pre-set medical consistency judgment model to determine the degree of logical consistency between the medical status labels and biomarker information, and obtain the corresponding medical authenticity score. Based on the medical authenticity score, it can be determined whether the voice speaker actually has the relevant symptoms, and thus it can be well judged whether the user is really in a pathological state, which can effectively reduce the risk of deception.

[0056] For example, in the process of anti-fraud during health insurance telephone consultations, the entire call text T is structurally extracted to obtain the chief symptom S_symptom (such as "dry cough" or "paroxysmal nocturnal dyspnea"). Then, S_symptom and the pathological fragment type set {type(p_j)} are input into a fine-tuned large model (based on medical knowledge graph enhancement). The model outputs a medical consistency score s_med ∈ [0,1]. For example, "dry cough" matches "dry cough fragment" and gets a high score, while "chest pain" has no corresponding sound and gets a low score. If the user describes "severe asthma" but only mild nasal sounds are detected, s_med is significantly reduced. At the same time, the large model evaluates whether the text is contradictory or overly dramatic (such as "severe cough for three days but in good spirits") and generates a semantic credibility s_credible.

[0057] Step S700: Determine the verification result based on voiceprint similarity information, physiological authenticity score, and medical authenticity score.

[0058] The data verification method provided in this application, after performing similarity calculation processing between voiceprint features and preset user-registered voiceprints to obtain voiceprint similarity information, performing physiological consistency judgment processing between voiceprint features and pathological acoustic biological features to obtain a physiological authenticity score, and performing medical consistency judgment processing between symptom description information and pathological sound segments to obtain a medical authenticity score, can determine the final verification result based on the voiceprint similarity information, physiological authenticity score, and medical authenticity score. Through the above technical solution, by comprehensively considering voiceprint similarity information, physiological authenticity score, and medical authenticity score during the data verification process, the final verification result is more reasonable and accurate.

[0059] It is worth noting that in determining the verification result based on voiceprint similarity information, physiological authenticity score, and medical authenticity score, the risk score is first calculated based on voiceprint similarity, physiological authenticity score, medical authenticity score, and preset weight coefficients. Then, the risk score is compared with a preset risk threshold to determine the corresponding verification result. This not only enables identity verification but also effectively determines whether the user is truly in a pathological state, thus significantly reducing the risk of deception.

[0060] like Figure 8 As shown, the verification result is determined based on voiceprint similarity information, physiological authenticity score, and medical authenticity score, and may include the following steps: Step S710: Calculate the risk score based on voiceprint similarity, physiological authenticity score, medical authenticity score, and preset weighting coefficients. Step S720: The risk score is compared with a preset risk threshold to determine the verification result.

[0061] For steps S710 to S720, in the process of determining the verification result based on voiceprint similarity information, physiological authenticity score, and medical authenticity score, a risk score is first calculated based on voiceprint similarity, physiological authenticity score, medical authenticity score, and preset weight coefficients. Subsequently, the risk score can be compared with a preset risk threshold to determine the corresponding verification result. This not only enables identity verification but also effectively determines whether the user is truly in a pathological state, thus significantly reducing the risk of deception.

[0062] It is worth noting that both the weighting coefficient and the risk threshold can be set according to actual needs, thereby making the determination of verification results more flexible and accurate.

[0063] It is worth noting that if the verification result indicates low risk, the verification can be approved; if the verification result indicates medium risk, it can be transferred to manual review; and if the verification result indicates high risk, the application can be directly rejected and the verification will fail.

[0064] In addition, such as Figure 9 As shown, one embodiment of this application also provides a data verification device 10, which includes: Acquisition unit 100 is used to acquire the original speech signal; The segmentation unit 200 is used to segment the original speech signal to obtain multiple speech segments; The detection unit 300 is used to perform sound event detection processing on multiple speech segments to obtain multiple pathological sound segments. The execution unit 400 is used to perform feature extraction processing on each pathological sound segment to obtain voiceprint features and pathological sound biological features; and to perform similarity calculation processing between the voiceprint features and the preset user-registered voiceprints to obtain voiceprint similarity information. The first judgment unit 500 is used to perform physiological consistency judgment processing on voiceprint features and pathological acoustic biological features to obtain a physiological authenticity score. The second judgment unit 600 is used to perform structured extraction of the speech text corresponding to the original speech signal to obtain symptom description information; and to perform medical consistency judgment processing on the symptom description information and pathological sound segments to obtain a medical authenticity score. The calculation unit 700 is used to determine the verification result based on voiceprint similarity information, physiological authenticity score and medical authenticity score.

[0065] The specific implementation of the data verification device 10 is basically the same as the specific embodiment of the data verification method described above, and will not be repeated here.

[0066] In addition, such as Figure 10 As shown, one embodiment of this application also provides an electronic device 800, which includes: a memory 820, a processor 810, and a computer program stored on the memory 820 and executable on the processor 810.

[0067] The processor 810 and memory 820 can be connected via a bus or other means.

[0068] The non-transient software program and instructions required to implement the data verification method of the above embodiments are stored in the memory 820. When executed by the processor 810, the data verification method of each of the above embodiments is executed.

[0069] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0070] Furthermore, one embodiment of this application also provides a computer-readable storage medium storing computer-executable instructions that are executed by a processor 810 or a controller, for example, by a processor 810 in the above-described device embodiment, which causes the processor 810 to perform the data verification method in the above-described embodiment.

[0071] The above embodiments can be used in combination, and modules with the same name in different embodiments may be the same or different.

[0072] The foregoing has described specific embodiments of this application; other embodiments are within the scope of the appended claims. In some cases, the actions or steps described in the claims may be performed in a different order than those shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily have to follow the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0073] The various embodiments in this application are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments of apparatus, devices, and computer-readable storage media are basically similar to the method embodiments, so the descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0074] The apparatus, device, computer-readable storage medium and method provided in the embodiments of this application are corresponding. Therefore, the apparatus, device and non-volatile computer storage medium also have similar beneficial technical effects as the corresponding method. Since the beneficial technical effects of the method have been described in detail above, the beneficial technical effects of the corresponding apparatus, device and computer storage medium will not be described again here.

[0075] In the 1990s, improvements to a technology could be clearly distinguished as either hardware improvements (e.g., improvements to the circuit structure of diodes, transistors, switches, etc.) or software improvements (improvements to the methodology). However, with technological advancements, many methodological improvements today can be considered direct improvements to the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved methodology into the hardware circuit. Therefore, it cannot be said that a methodological improvement cannot be implemented using hardware physical modules. For example, a Programmable Logic Device (PLD) (such as a Field Programmable Gate Array (FPGA)) is such an integrated circuit whose logic function is determined by the user programming the device. Designers can program a digital system themselves to "integrate" it onto a PLD, without needing chip manufacturers to design and manufacture dedicated integrated circuit chips. Moreover, nowadays, instead of manually manufacturing integrated circuit chips, this programming is mostly implemented using "logic compiler" software, which is similar to the software compiler used when writing program development code. The original code before compilation must also be written in a specific programming language, which is called a Hardware Description Language (HDL). There is not just one HDL, but many kinds, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. Currently, the most commonly used are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also understand that by simply performing some logic programming on the method flow using the aforementioned hardware description languages ​​and programming it into an integrated circuit, the hardware circuit that implements the logic method flow can be easily obtained.

[0076] The controller can be implemented in any suitable manner. For example, it can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers. Examples of controllers include, but are not limited to, the following microcontrollers: ARC625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicon Labs C8051F320. A memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art will also recognize that, in addition to implementing the controller in purely computer-readable program code form, the same functionality can be achieved by logically programming the method steps to make the controller take the form of logic gates, switches, ASICs, programmable logic controllers, and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the means included therein for implementing various functions can also be considered as structures within the hardware component. Alternatively, the means for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0077] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0078] For ease of description, the above apparatus is described by dividing it into various functional units. Of course, in implementing the embodiments of this application, the functions of each unit can be implemented in one or more software and / or hardware.

[0079] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, embodiments of this application can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of this application can take the form of computer program products implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0080] This specification is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0081] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0082] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0083] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0084] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash memory (FlashRAM). Memory is an example of computer-readable media.

[0085] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0086] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0087] In this application embodiment, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent the existence of A alone, A and B simultaneously, or B alone. A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of singular or plural items. For example, at least one of a, b, and c can represent: a, b, c, a and b, a and c, b and c, or a and b and c, where a, b, and c can be single or multiple.

[0088] The embodiments of this application can be described in the general context of computer-executable instructions, such as program modules, that are executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. The embodiments of this application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In a distributed computing environment, program modules can reside in local and remote computer storage media, including storage devices.

[0089] The various embodiments in this application are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0090] It should be noted that any AI models, software tools, or components not belonging to this company appearing in the embodiments of this application are merely illustrative examples and do not represent actual use. All user personal information involved in the embodiments of this application has been authorized (with the knowledge and consent) by the relevant parties or has been fully authorized by all parties, and the executing entity may obtain it through various legal and compliant means. The collection, storage, use, processing, transmission, provision, and disclosure of the information, data, and signals involved all comply with relevant laws and regulations and do not violate public order and good morals.

[0091] The above description is merely an embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of this application should be included within the scope of the claims of this application.

Claims

1. A data verification method, characterized in that, The data verification method includes: Acquire the raw speech signal; The original speech signal is segmented to obtain multiple speech segments; Multiple speech segments are subjected to sound event detection processing to obtain multiple pathological sound segments; For each of the pathological sound segments, feature extraction processing is performed to obtain voiceprint features and pathological acoustic biological features; and the voiceprint features are compared with preset user-registered voiceprints to obtain voiceprint similarity information. The voiceprint features and the pathological acoustic biological features are subjected to physiological consistency judgment processing to obtain a physiological authenticity score. The original speech signal is subjected to structured extraction of the corresponding speech text to obtain symptom description information; and the symptom description information and the pathological sound segment are subjected to medical consistency judgment processing to obtain a medical authenticity score. The verification result is determined based on the voiceprint similarity information, the physiological authenticity score, and the medical authenticity score.

2. The data verification method according to claim 1, characterized in that, The original speech signal is segmented to obtain multiple speech segments, including: The original speech signal is divided into multiple consecutive short time frames; For each of the short-time frames, a short-time energy calculation is performed to obtain the short-time energy; and for each of the short-time frames, a zero-crossing count detection is performed to obtain the short-time zero crossover rate. The original speech signal is segmented based on the short-time energy, the short-time zero-crossing rate, and preset energy and zero-crossing rate thresholds to obtain multiple speech segments.

3. The data verification method according to claim 1, characterized in that, The process of performing sound event detection on multiple speech segments yields multiple pathological sound segments, including: Based on a preset sound event detection module, multiple speech segments are processed to predict the probability of pathological events, thereby obtaining multiple probabilities of pathological events. Based on a preset peak clustering algorithm, the probabilities of multiple pathological events are extracted into a continuous probability interval to obtain multiple pathological segments, wherein each pathological segment carries a confidence level. The pathological audio segments whose confidence level is greater than a preset confidence level threshold are designated as the pathological audio segments.

4. The data verification method according to claim 1, characterized in that, The physiological consistency determination process for the voiceprint features and the pathological acoustic biological features, to obtain a physiological authenticity score, includes: The voiceprint features and the pathological acoustic biological features are constructed into joint embedded information; The joint embedding information is processed by a preset binary classifier to determine consistency, thereby obtaining the physiological authenticity score.

5. The data verification method according to claim 1, characterized in that, The step of extracting structured speech text corresponding to the original speech signal to obtain symptom description information includes: The original speech signal is subjected to speech recognition and decoding processing to obtain the speech text; The symptom description information is obtained by performing structured information extraction processing on the spoken text.

6. The data verification method according to claim 1, characterized in that, The step of performing medical consistency determination on the symptom description information and the pathological sound fragments to obtain a medical authenticity score includes: The symptom description information is extracted and processed to obtain medical symptom tags; and the pathological sound segments are feature extracted to obtain biomarker information. The medical symptom labels and biomarker information are matched based on a pre-defined medical consistency judgment model to determine the degree of logical consistency between the medical status labels and the biomarker information, thereby obtaining the medical authenticity score.

7. The data verification method according to claim 1, characterized in that, The verification result is determined based on the voiceprint similarity information, the physiological authenticity score, and the medical authenticity score, including: A risk score is calculated based on the voiceprint similarity, the physiological authenticity score, the medical authenticity score, and a preset weighting coefficient. The risk score is compared with a preset risk threshold to determine the verification result.

8. A data verification device, characterized in that, The data verification device includes: The acquisition unit is used to acquire the raw speech signal; The segmentation unit is used to segment the original speech signal to obtain multiple speech segments; The detection unit is used to perform sound event detection processing on multiple speech segments to obtain multiple pathological sound segments; The execution unit is used to perform feature extraction processing on each of the pathological sound segments to obtain voiceprint features and pathological sound biological features; and to perform similarity calculation processing on the voiceprint features and the preset user-registered voiceprints to obtain voiceprint similarity information. The first determination unit is used to perform physiological consistency determination processing on the voiceprint features and the pathological acoustic biological features to obtain a physiological authenticity score. The second determination unit is used to perform structured extraction of the speech text corresponding to the original speech signal to obtain symptom description information; and to perform medical consistency determination processing on the symptom description information and the pathological sound segment to obtain a medical authenticity score. The calculation unit is used to determine the verification result based on the voiceprint similarity information, the physiological authenticity score, and the medical authenticity score.

9. An electronic device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, when the processor executes the computer program, it implements the data verification method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing computer-executable instructions, characterized in that, The computer-executable instructions are used to execute the data verification method according to any one of claims 1 to 7.

Images