An identity authentication method, device and equipment

Abstract

The embodiment of the specification discloses an identity verification method, device and equipment, the method comprises: in response to the identity verification request triggered by the target user for the verification event, determining the risk level corresponding to the verification event, and generating the dynamic verification instruction matched with the risk level; based on the mapping relationship between the preset risk level and the feature dimension, determine the first feature dimension corresponding to the verification event; based on the device information of the target user using device, determine the feature dimension matched with the device information as the second feature dimension corresponding to the verification event; obtain the verification video and / or verification audio generated by the target user executing the verification action under the guidance of the dynamic verification instruction, based on the common feature dimension between the first feature dimension and the second feature dimension, extract at least one verification feature from the verification audio and / or verification video, and verify the identity of the target user by using the extracted verification feature, and obtain the identity verification result of the target user.

Description

Technical Field

[0001] This document relates to the field of information security technology, and in particular to an authentication method, device, and equipment. Background Technology

[0002] Identity verification is a fundamental aspect of information security and access control, with its core objective being to accurately confirm a user's true identity in a non-cooperative environment. With the increasing prevalence of digital services, identity verification is being used in more and more scenarios, placing higher demands on its accuracy, robustness, and resistance to attacks.

[0003] Currently, identity verification is mainly conducted through methods such as CAPTCHA, facial recognition, or fingerprint recognition. CAPTCHA is easily stolen, facial recognition performance drops sharply under occlusion and mask attacks, and fingerprint recognition is highly sensitive to the moisture and wear of the finger surface. It is evident that these verification methods are not only vulnerable to specific types of forgery attacks due to their reliance on a single method, but also prone to excluding some users from the service due to their fixed nature (e.g., disabled users may not be able to perform fingerprint recognition), thus affecting the accuracy of identity verification. Therefore, a technical solution is needed that can adaptively select a verification method suitable for the verification scenario to improve the accuracy of identity verification. Summary of the Invention

[0004] The purpose of the embodiments in this specification is to provide a technical solution that can perform identity verification more accurately.

[0005] To achieve the above technical solution, the embodiments in this specification are implemented as follows: This specification provides an authentication method comprising: responding to an authentication request triggered by a target user for a verification event, determining a risk level corresponding to the verification event, and generating a dynamic verification instruction adapted to the risk level; determining a first feature dimension mapped to the risk level corresponding to the verification event based on a preset mapping relationship between risk levels and feature dimensions; determining a second feature dimension corresponding to the verification event based on device information of the device used by the target user, using the feature dimension matching the device information; acquiring a verification video and / or verification audio generated by the target user performing a verification action under the guidance of the dynamic verification instruction, and extracting at least one verification feature corresponding to the common feature dimension between the first feature dimension and the second feature dimension from the verification audio and / or verification video; and verifying the identity of the target user using the at least one verification feature to obtain the authentication result of the target user.

[0006] This specification provides an authentication device comprising: a request parsing module, which, in response to an authentication request triggered by a target user for a verification event, determines the risk level corresponding to the verification event and generates a dynamic verification instruction adapted to the risk level; a feature processing module, which, based on a preset mapping relationship between risk levels and feature dimensions, determines a first feature dimension mapped to the risk level corresponding to the verification event; based on device information of the device used by the target user, determines a feature dimension matching the device information as a second feature dimension corresponding to the verification event; acquires a verification video and / or verification audio generated by the user performing a verification action under the guidance of the dynamic verification instruction, and extracts at least one verification feature corresponding to the common feature dimension between the first feature dimension and the second feature dimension from the verification audio and / or the verification video; and an authentication module, which uses the at least one verification feature to verify the identity of the target user and obtain the authentication result of the target user.

[0007] This specification provides an authentication device comprising: a processor; and a memory arranged to store computer-executable instructions, wherein the executable instructions, when executed, cause the processor to: respond to an authentication request triggered by a target user for a verification event, determine the risk level corresponding to the verification event, and generate a dynamic verification instruction adapted to the risk level; determine a first feature dimension mapped to the risk level corresponding to the verification event based on a preset mapping relationship between risk levels and feature dimensions; determine a second feature dimension corresponding to the verification event based on device information of the device used by the target user, using the feature dimension matching the device information; acquire a verification video and / or verification audio generated by the target user performing a verification action under the guidance of the dynamic verification instruction, and extract at least one verification feature corresponding to the common feature dimension between the first feature dimension and the second feature dimension from the verification audio and / or verification video; and verify the identity of the target user using the at least one verification feature to obtain the authentication result of the target user.

[0008] This specification also provides a storage medium for storing computer-executable instructions. When executed by a processor, the executable instructions implement the following process: in response to an authentication request triggered by a target user for a verification event, determining the risk level corresponding to the verification event and generating a dynamic verification instruction adapted to the risk level; determining a first feature dimension mapped to the risk level corresponding to the verification event based on a preset mapping relationship between risk levels and feature dimensions; determining a second feature dimension corresponding to the verification event based on device information of the device used by the target user and matching the device information; acquiring a verification video and / or verification audio generated by the target user performing a verification action under the guidance of the dynamic verification instruction, and extracting at least one verification feature corresponding to the common feature dimension between the first feature dimension and the second feature dimension from the verification audio and / or verification video; and verifying the identity of the target user using the at least one verification feature to obtain the authentication result of the target user.

[0009] This specification also provides a computer program product, including a computer program that, when executed by a processor, implements the following process: in response to an authentication request triggered by a target user for a verification event, determining the risk level corresponding to the verification event and generating a dynamic verification instruction adapted to the risk level; based on a preset mapping relationship between risk levels and feature dimensions, determining a first feature dimension mapped to the risk level corresponding to the verification event; based on device information of the device used by the target user, determining the feature dimension matching the device information as a second feature dimension corresponding to the verification event; acquiring a verification video and / or verification audio generated by the target user performing a verification action under the guidance of the dynamic verification instruction, and extracting at least one verification feature corresponding to the common feature dimension between the first feature dimension and the second feature dimension from the verification audio and / or verification video; using the at least one verification feature to verify the identity of the target user, obtaining the authentication result of the target user. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the drawings used in the description of the embodiments or prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Figure 1 This is a flowchart illustrating one authentication method described in this specification. Figure 2 This is a flowchart illustrating a process for determining the risk level corresponding to a verification event, as described in this specification. Figure 3 This is a flowchart illustrating another method for determining the risk level of a verification event in this specification. Figure 4 This is a schematic diagram of a process for extracting verification features according to this specification; Figure 5 This is a schematic diagram illustrating a process for verifying the identity of a target user based on multiple verification features, as described in this specification. Figure 6 This is a flowchart illustrating a process for determining the comprehensive test results at the same point in time, as described in this specification. Figure 7 This is a flowchart illustrating the execution of an authentication method in one application scenario described in this specification. Figure 8 This is a schematic diagram of the structure of an authentication device described in this specification; Figure 9 This is a schematic diagram of the structure of an authentication device described in this specification. Detailed Implementation

[0011] This specification provides an authentication method, apparatus, and device through its embodiments.

[0012] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0013] Currently, identity verification is mainly conducted through methods such as CAPTCHA, facial recognition, or fingerprint recognition. CAPTCHA is easily stolen, facial recognition performance drops sharply under occlusion and mask attacks, and fingerprint recognition is highly sensitive to the moisture and wear of the finger surface. These verification methods are not only vulnerable to specific types of forgery attacks due to their reliance on a single method, but their fixed nature can also exclude some users from the service, thus affecting the accuracy of identity verification.

[0014] To address this, the present invention proposes a superior identity verification scheme. In response to an identity verification request triggered by a target user for a verification event, the scheme determines the risk level corresponding to the verification event and generates a dynamic verification instruction adapted to the risk level. It acquires the verification video and / or verification audio generated by the target user performing verification actions under the guidance of the dynamic verification instruction. Based on a preset mapping relationship between risk levels and feature dimensions, it determines the first feature dimension mapped to the risk level of the verification event. Then, based on the device information of the target user's device, it determines the feature dimension matching the device information as the second feature dimension corresponding to the verification event. Finally, based on the shared feature dimension between the first and second feature dimensions, it extracts at least one verification feature from the verification audio and / or verification video to verify the target user's identity using at least one verification feature, thus obtaining the target user's identity verification result. In this way, by identifying the verification event, determining its risk level, and adaptively determining the corresponding verification feature for the target user based on the determined risk level, the scheme reduces verification errors caused by using general verification methods and effectively improves the accuracy of identity verification. Specific processing details can be found in the following embodiments.

[0015] like Figure 1 As shown in the embodiments of this specification, an authentication method is provided. The execution subject of this method can be a terminal device or a server, etc. The terminal device can be a mobile terminal device such as a mobile phone or tablet computer, a computer device such as a laptop or desktop computer, or an IoT device (specifically, a smartwatch, in-vehicle device, etc.). The server can be a single server or a server cluster composed of multiple servers. The server can be a backend server in fields such as finance or online shopping, or a backend server of an application. This embodiment uses a server as the execution subject for detailed description. For the case where the execution subject is a terminal device, please refer to the following server-side processing, which will not be repeated here. The method may specifically include the following steps: In step S102, in response to the authentication request triggered by the target user for the authentication event, the risk level corresponding to the authentication event is determined, and a dynamic authentication instruction adapted to the risk level is generated.

[0016] In this context, an event refers to a state change or action that occurs within a system or application and is worthy of being perceived and responded to. A verification event refers to an event that presents a risk and requires confirmation of the operator's true identity before execution can proceed. Examples of verification events include transaction actions and requests to modify sensitive information. The operator in a verification event is the target user. Dynamic verification instructions are control information containing specific verification task parameters, which can be used to drive the target user to complete identity verification according to specified requirements.

[0017] In implementation, various verification events can be preset, such as payment events and password change events. When a target user triggers a verification event, an authentication request corresponding to the verification event is generated. Then, in response to the authentication request, the risk level corresponding to the verification event can be determined. The risk level reflects the potential loss caused by the verification event; for example, the risk level can be high-risk, medium-risk, or low-risk. There are several ways to determine the corresponding risk level for a verification event. For example, a mapping relationship between verification events and risk levels can be preset, and the risk level corresponding to the verification event triggered by the target user can be found based on the mapping relationship; alternatively, the risk of the verification event can be assessed in real time, and the risk level corresponding to the verification event triggered by the target user can be determined based on the assessment results. These are merely examples, and this embodiment does not limit the scope of the implementation.

[0018] Then, different instruction generation rules can be set for different risk levels to generate dynamic verification instructions that are adapted to the risk level according to the corresponding instruction generation rules. A dynamic verification instruction can be a single instruction or multiple consecutive instructions. It is important to note that dynamic verification instructions for the same risk level may have the same type, but their content may not be exactly the same. For example, dynamic verification instructions for a low-risk level might instruct the target user to read different numeric verification codes.

[0019] In step S104, based on the preset mapping relationship between risk level and feature dimension, the first feature dimension mapped to the risk level corresponding to the verification event is determined.

[0020] Among them, feature dimension refers to the type of biometric feature involved in identity verification, such as facial feature dimension, voiceprint feature dimension, etc.

[0021] In implementation, corresponding feature dimensions can be pre-configured for different risk levels to pre-generate a mapping relationship between risk levels and feature dimensions. Based on this, after determining the risk level corresponding to the verification event, the feature dimension mapped to the risk level of the verification event can be found by retrieving the mapping relationship, and the found feature dimension is used as the first feature dimension. The number of first feature dimensions can be one or more.

[0022] In step S106, based on the device information of the target user's device, the feature dimension that matches the device information is determined as the second feature dimension corresponding to the verification event.

[0023] In implementation, the device used by the target user is identified, and its device information is obtained. Device information is attribute information used to uniquely identify the device and provide its biometric data collection capabilities; it may include at least hardware identification information and device function description information. Based on the device information, the biometric data collection capabilities of the device can be determined. The feature dimension corresponding to the device's biometric data collection capabilities is defined as the second feature dimension, and there may be one or more second feature dimensions.

[0024] In step S108, the verification video and verification audio generated by the target user performing verification actions under the guidance of dynamic verification instructions are obtained, and at least one verification feature corresponding to the common feature dimension is extracted from the verification audio and / or verification video based on the common feature dimension between the first feature dimension and the second feature dimension.

[0025] As mentioned above, the number of first feature dimensions can be one or more, and the number of second feature dimensions can also be one or more. The intersection of the first feature dimension and the second feature dimension can be determined, and the obtained intersection is determined as the common feature dimension between the first feature dimension and the second feature dimension. That is, the common feature dimension can be understood as the result of the intersection of two feature dimension sets, and the number of common feature dimensions can be one or more.

[0026] In implementation, dynamic verification commands can be issued to target users, displaying the required verification actions. Users can complete these actions according to the commands. The server can simultaneously capture video and audio streams of the user's actions, designating the video stream as the verification video and the audio stream as the verification audio. Then, based on the identified common feature dimensions, the extraction object corresponding to each dimension can be determined. This object could be either verification audio or verification video. Verification features corresponding to the common feature dimensions are extracted from these objects. For example, if the common feature dimension is a voiceprint feature dimension, the extracted verification feature could be a voiceprint feature. Each common feature dimension corresponds to one verification feature. Feature extraction is performed on at least one common feature dimension to obtain at least one verification feature extracted from the verification audio and / or verification video.

[0027] In step S110, the identity of the target user is verified using at least one verification feature to obtain the identity verification result of the target user.

[0028] Verification features can be understood as biometric vectors of a specified dimension extracted from verification video or audio data. Verification features can be used to verify the identity of a target user from a specified dimension.

[0029] In implementation, at least one extracted verification feature can be combined for comprehensive judgment. Especially when there are multiple verification features, by deeply fusing multi-dimensional biometric vectors, the identity of the target user can be comprehensively verified from multiple dimensions, generating the target user's identity verification result. Deeply fusing multi-dimensional features for identity verification can more comprehensively identify forgery attacks, thereby significantly improving the accuracy of identity verification.

[0030] This specification provides an authentication method that, in response to an authentication request triggered by a target user for a verification event, determines the risk level corresponding to the verification event and generates a dynamic verification instruction adapted to the risk level. It acquires verification video and / or verification audio generated when the target user performs a verification action guided by the dynamic verification instruction. Based on a preset mapping relationship between risk levels and feature dimensions, it determines the first feature dimension mapped to the risk level of the verification event. Then, based on the device information of the target user's device, it determines the feature dimension matching the device information as the second feature dimension corresponding to the verification event. Finally, based on the shared feature dimension between the first and second feature dimensions, it extracts at least one verification feature from the verification audio and / or verification video to verify the target user's identity using at least one verification feature, thus obtaining the target user's authentication result. In this way, by identifying the verification event, determining its risk level, and adaptively determining the corresponding verification feature for the target user based on the determined risk level, it reduces the verification error caused by using general verification methods and effectively improves the accuracy of authentication.

[0031] In the above or following embodiments, the processing method for determining the risk level corresponding to the verification event in step S102 in response to the authentication request triggered by the target user for the verification event can be varied. An exemplary processing method is provided below, such as... Figure 2 As shown, this processing method may specifically include the following steps.

[0032] Step S202: In response to the authentication request triggered by the target user for the authentication event, determine at least one risk assessment dimension that is compatible with the authentication event.

[0033] In implementation, multiple risk assessment dimensions can be pre-defined for risk levels. Different risk assessment dimensions can reflect different evaluation perspectives on verification events. When an authentication request triggered by a target user for a verification event is received, at least one risk assessment dimension that matches the verification event can be selected from the pre-defined multiple risk assessment dimensions.

[0034] There are various ways to select risk assessment dimensions. For example, a preset risk assessment dimension can be associated with an event type. Upon receiving an authentication request, the event type corresponding to the verification event is first determined, and then the risk assessment dimension associated with that event type is set as the risk assessment dimension suitable for the verification event. Alternatively, fixed risk assessment dimensions can be pre-associated with applications, and the risk assessment dimension associated with the application that triggers the verification event is set as the risk assessment dimension suitable for the verification event. Of course, these are just examples, and other selection methods can be used, which will not be exhaustively listed in this embodiment.

[0035] Step S204: Calculate the risk score of the verification event under each risk assessment dimension, and determine the risk level corresponding to the verification event based on the risk score corresponding to at least one risk assessment dimension.

[0036] In implementation, after identifying at least one risk assessment dimension corresponding to the verification event, the risk score of the verification event under each risk assessment dimension can be calculated. The risk scores corresponding to at least one risk assessment dimension are weighted to determine the comprehensive score corresponding to the verification event. Then, based on the preset mapping relationship between risk level and comprehensive score, the risk level mapped by the comprehensive score corresponding to the verification event can be determined, and the mapped risk level is determined as the risk level corresponding to the verification event.

[0037] Of course, other methods can also be used to determine the risk level of the verification event. For example, for the risk score obtained under each risk assessment dimension, a risk sub-level can be determined for that risk assessment dimension. Then, a comprehensive decision can be made based on the risk sub-levels corresponding to at least one risk assessment dimension to obtain the risk level of the verification event. Specifically, when there is only one risk assessment dimension, the risk sub-level corresponding to that risk assessment dimension is the risk level of the verification event; when there are multiple risk assessment dimensions, the risk sub-level corresponding to the risk assessment dimension with the highest weight can be determined as the risk level of the verification event.

[0038] like Figure 3 As shown, in one optional implementation, the implementation method of calculating the risk score of the verification event under each risk assessment dimension in step S204, and determining the risk level corresponding to the verification event based on the risk scores corresponding to multiple risk assessment dimensions, may specifically include the following steps.

[0039] Step S302: Determine the event type of the verification event, and determine the event risk score of the verification event based on the risk assessment rules corresponding to the event type.

[0040] This allows for the pre-setting of various event types, such as transactions, payments, password changes, and account binding, with different risk assessment rules set for each event type. For example, for the "payment" event type, the risk of the event type can be assessed based on the other party's credit score; the lower the other party's credit score, the higher the risk of this event type.

[0041] In implementation, the event type of the verification event can be identified first. Then, the risk of the verification event can be assessed using the risk assessment rules corresponding to that event type, and an event risk score can be generated based on the assessment results. The event risk score represents the probability that the verification event itself may cause a loss.

[0042] Alternatively, manually labeled verification events and event risk scores can be used to train the event assessment model, so that the trained event assessment model contains risk assessment rules corresponding to multiple event types. The verification event is input into the trained event assessment model, the event assessment model identifies the event type of the verification event, and evaluates the verification event using the risk assessment rules corresponding to the event type, and outputs the event risk score of the verification event.

[0043] It is important to note that when a verification event is associated with multiple event types, the risk assessment rules for each event type can be used to evaluate the verification event separately. The evaluation results of multiple event types can then be combined to make a decision to determine the event risk score of the verification event.

[0044] Step S304: Determine the device used by the target user when triggering the verification event, assess the security of the device, and determine the device risk score of the verification event based on the assessment results.

[0045] The equipment risk score represents the probability of loss caused by a device that triggers a verification event. Multiple evaluation indicators can be pre-defined for each device, such as whether it is an original manufacturer device, whether there are signs of modification, whether the serial number is valid, and whether the operating environment is normal.

[0046] In implementation, for each device used when a verification event is triggered, the device can be tested using preset evaluation indicators to assess its security. Then, the device risk score of the verification event can be determined based on the test results of multiple evaluation indicators. For example, if the test result of each evaluation indicator is "yes" or "no", the device risk score of the verification event can be determined based on the number of "yes" test results or the proportion of the number of "yes" test results in the total number of indicators.

[0047] Step S306: Obtain the risky operation behaviors performed by the target user within a preset historical period, and determine the behavioral risk score of the verification event based on the risk level and frequency of the risky operation behaviors.

[0048] The behavioral risk score represents the probability of loss due to user behavior. In practical applications, risky operations refer to actions that pose an attack risk. Depending on the needs of the actual scenario, certain types of actions can be pre-defined as risky operations, such as changing passwords or making transactions.

[0049] In implementation, the user's actions within a preset historical period can be acquired. Based on preset risky action types, risky actions are extracted from the acquired actions; the number of risky actions can be one or more. Then, the frequency of each risky action and its risk level within that historical period can be determined. Based on the frequency and risk level, a risk sub-score for each risky action is determined. Finally, the risk sub-scores of the extracted one or more risky actions are weighted and calculated to obtain the behavioral risk score for the verification event.

[0050] Step S308: Determine the risk level corresponding to the verification event based on the event hazard score, equipment hazard score, and behavioral risk score.

[0051] In implementation, the event hazard score, equipment hazard score, and behavioral risk score can be normalized first. Then, the event hazard score, equipment hazard score, and behavioral risk score can be weighted and calculated to obtain the comprehensive score corresponding to the verification event. Based on the preset mapping relationship between risk level and comprehensive score, the risk level mapped by the comprehensive score corresponding to the verification event can be determined, and the mapped risk level can be determined as the risk level corresponding to the verification event.

[0052] certainly, Figure 3 The risk assessment dimensions included in the implementation shown are merely exemplary, and other risk assessment dimensions can also be used to determine the risk level corresponding to the verification event. For example, risk assessment dimensions may also include event context, triggering operation behavior, etc., to assess the risk of the verification event based on the contextual information in the verification event (such as the specific amount) and the user's operation behavior that triggers the verification event (such as mouse trajectory). This embodiment does not limit the risk assessment dimensions.

[0053] In the above or following embodiments, various implementation methods can be used to extract at least one verification feature corresponding to a common feature dimension from the verification audio and / or verification video. For example, when the risk level corresponding to the verification event is high-risk, the common feature dimension may include a voiceprint feature dimension, a speech content feature dimension, a facial feature dimension, and a lip movement feature dimension. Based on this, such as... Figure 4As shown, the specific implementation method for extracting at least one verification feature corresponding to the common feature dimension from verification audio and / or verification video may include the following steps.

[0054] Step S402: Extract the voiceprint features and speech content features of the target user from the verification audio.

[0055] Voiceprint features refer to the phonological features contained in a voice that can represent and identify the speaker. Different users have different voiceprint features, making voiceprints unique. Speech content features are used to describe the content of the target user's speech.

[0056] In implementation, voiceprint features and speech content features can be extracted through two parallel paths. Preprocessing such as noise reduction and silence removal can be applied to the verification audio to obtain valid verification audio. In one path, feature extraction can be performed on the valid verification audio to obtain the target user's voiceprint features, which may include spectral features, prosodic features, etc. In the other path, speech recognition can be performed on the valid verification audio, and the recognized audio content can be converted into text to obtain the target user's speech-text features.

[0057] Step S404: Extract the facial features and lip movement features of the target user from the verification video.

[0058] Among them, facial features are used to describe the appearance of the target user, and lip movement features are used to describe the trend of lip shape changes of the target user.

[0059] In implementation, facial features and lip motion features can be extracted through two parallel paths. In one path, keyframes containing faces are selected from the verification video. The face position is located within these keyframes, and the face pose is corrected to align with a standard frontal face image. The target user's facial features are then extracted from these corrected keyframes. In the other path, the lip region is located frame-by-frame from the verification video, and lip contour points are marked within this region to obtain a lip marker sequence. This lip marker sequence is then modeled to generate lip motion features.

[0060] Step S406: Identify at least one verification feature corresponding to the common feature dimension, including voiceprint features, speech content features, facial features, and lip movement features.

[0061] Among them, facial features are static image features, while voiceprint features, speech content features, and lip movement features are dynamic features related to time series. In practical applications, multimodal features such as voiceprint features, speech content features, facial features, and lip movement features extracted from verification videos and audio can all be used as verification features for authenticating target users.

[0062] It should be understood that voiceprint features originate from audio signals, and their time series is determined by the start and end times of speech; speech content features originate from the text output of ASR (Automatic Speech Recognition), and their timestamps depend on the processing latency of the recognition engine; lip movement features originate from video frame sequences, and their timestamps depend on the camera's frame rate and the video encoding time base. The sampling rates, time bases, and start and end times of these three features are all different, resulting in a lack of direct correspondence between speech content and lip movements at the same moment. Therefore, when using multiple verification features to verify the identity of a target user, it is necessary to time-align voiceprint features, speech content features, and lip movement features to improve verification accuracy.

[0063] Based on this, such as Figure 5 As shown, the process of verifying the identity of a target user using at least one verification feature to obtain the target user's authentication result may specifically include the following steps: Step S502: Extract the time sequence of the verification action from the voiceprint features and determine the time sequence as the time axis.

[0064] The voiceprint feature is a vector sequence composed of voiceprint feature vectors from multiple audio frames. The frame indices of this sequence are used as time scales, and a time axis is formed based on these multiple time scales. Each point on the time axis corresponds to a voiceprint feature vector. For example, if the verification audio duration is 10 seconds, and the processed valid verification audio is from second 1.2 to second 8.5, then the time axis ranges from second 1.2 to second 8.5. Furthermore, the voiceprint feature has already filtered out silent segments, and its time sequence accurately reflects the actual time period during which the user spoke, making it cleaner than the original audio timeline and more accurate than the video timeline.

[0065] Step S504: Determine the original timestamps corresponding to the speech content features and lip movement features respectively.

[0066] Wherein, both the voice content feature and the lip movement feature are vector sequences that change over time, and they are dynamic just like the voiceprint feature. The voice content feature may include a plurality of token vectors, which form a vector sequence in the order of audio frames. The start time corresponding to each token vector is recorded as an original timestamp, and each token vector corresponds to one original timestamp. For example, the voice content feature may include token vectors "I want", "to transfer", "three thousand yuan to", and "Zhang San", and the original timestamps corresponding to these token vectors are 1s, 1.5s, 2s, and 3s respectively. During the process in which the target user performs the verification action, the shape of the lips changes in each frame. The lip movement feature may include a plurality of mouth shape vectors, which form a vector sequence in the order of video frames. The change moment of the mouth shape vector is recorded as an original timestamp, and each mouth shape vector corresponds to one original timestamp. For example, the lip movement feature may include mouth shape vectors "wide open", "slightly open", "wide open", and "slightly closed", and the original timestamps corresponding to these mouth shape vectors are 1s, 2s, 3s, and 4s respectively.

[0067] Step S506: Based on the determined original timestamps, map the voice content feature and the lip movement feature onto the main time axis to obtain a time-aligned multi-modal feature sequence.

[0068] In implementation, taking the moment Ti on the main time axis as an example, if the original timestamps of the voice content feature and the lip movement feature include the moment Ti, the voiceprint feature, the token vector and the mouth shape vector at the moment Ti can be directly concatenated. If the original timestamps of the voice content feature and the lip movement feature do not include the moment Ti, the two time nodes before and after Ti that are closest to Ti in the original timestamps of the voice content feature can be found, and the token vector at Ti can be calculated through linear interpolation. If Ti exceeds the time range of all words, nearest neighbor extrapolation is used or the token vector is set to a zero vector. Similarly, the two time nodes before and after Ti that are closest to Ti are found, and the mouth shape vector at Ti is calculated through linear interpolation. The three feature vectors at the same point on the main time axis are concatenated to finally obtain a time-aligned multi-modal feature sequence. The feature vector at each time node contains information of three modalities: voiceprint, content, and mouth shape, and is strictly aligned in the time dimension.

[0069] Step S508: Verify the identity of the target user based on the multi-modal feature sequence, and obtain an identity verification result of the target user.

[0070] There may be various ways to verify the identity of the target user based on the multi-modal feature sequence.

[0071] In one optional verification method, the verification logic corresponding to each of the voiceprint features, speech content features, and lip movement features can be obtained. Simultaneously, the verification logic corresponding to each of these features is invoked to perform parallel verification of the voiceprint features, speech content features, and lip movement features at the same time point in the multimodal feature sequence. When the verification results of multiple modal features all show that they pass the verification, it can be determined whether the three modal features of voiceprint, content, and lip movement at that time point are interdependent. The verification score for that time point can be obtained based on the dependency relationship among the three. Then, based on the verification scores of all time points along the time axis, the identity verification result of the target user is comprehensively derived. For example, the identity verification of the target user can be determined to be successful when the number of time points with verification scores exceeding a preset threshold exceeds a preset percentage; or, the average verification score of all time points can be calculated, and the identity verification of the target user can be determined to be successful when the average value exceeds a preset threshold.

[0072] In another optional implementation, the verification logic corresponding to each of the voiceprint features, speech content features, and lip movement features can be obtained. Then, the verification logic corresponding to each of these features can be called simultaneously to perform parallel verification of the voiceprint features, speech content features, and lip movement features at the same time point in the multimodal feature sequence. Based on the verification results of multiple verification features at the same time point, a comprehensive verification result at the same time point is determined. For example, different weights can be assigned to features of different modalities, and a comprehensive verification result can be obtained by weighted calculation of the verification results of multiple verification features. Then, a fusion decision is made on the comprehensive verification results at multiple time points in the multimodal feature sequence to obtain the identity verification result of the target user. In practical applications, a fusion decision model can be pre-trained so that the model can output the identity verification result of the target user based on the coherence between the comprehensive verification results at multiple time points and the number of time points where verification is successful.

[0073] Based on this, such as Figure 6 As shown, parallel verification is performed on the voiceprint features, speech content features, and lip movement features at the same time point in the multimodal feature sequence. Based on the verification results of multiple verification features at the same time point, the specific implementation method of the comprehensive verification result at the same time point can include the following steps: Step S602: Extract the target user's identity data from the preset user database. The identity data includes reserved voiceprint data.

[0074] The target user's identity data is collected in advance with the target user's authorization, and the identity data includes, but is not limited to, reserved voiceprint data, reserved facial data, and reserved fingerprint data.

[0075] Step S604: Extract text content from dynamic verification instructions and determine the standard lip change trend corresponding to the text content.

[0076] Among them, high-risk dynamic verification instructions may include execution information displayed to the target user in text form, requiring the target user to read the text content contained in the execution information. When generating dynamic verification instructions containing text content, a standard video of reading the text content can be generated simultaneously, and the standard lip change trend corresponding to the text content can be extracted from the standard video.

[0077] Step S606: Simultaneously check whether the voiceprint features at the same time point match the reserved voiceprint data, whether the speech content features are consistent with the text content, and whether the lip movement features are consistent with the standard lip change trend.

[0078] In implementation, verification processes targeting different features can be executed in parallel. By verifying whether the voiceprint features match the pre-reserved voiceprint data, it can be confirmed whether the user actually performing the verification action is the claimed user, preventing impersonation or substitution. By verifying the synchronicity between lip movement features and standard lip change trends, it can be determined whether the user actually performing the verification action is a real person speaking in real time, rather than a pre-recorded playback or AI synthesis. By verifying whether the speech content features are consistent with the text content, it can be determined whether the user actually performing the verification action accurately uttered the specified content, preventing induced tampering or content replacement.

[0079] Step S608: Determine the user dimension score based on the test results for voiceprint features, determine the command dimension score based on the test results for speech content features, and determine the liveness dimension score based on the test results for lip movement features.

[0080] In practice, the probability that the user actually performing the verification action is the claimed user can be determined as the user dimension score based on the fit between voiceprint features and reserved voiceprint data; the probability that the user actually performing the verification action accurately speaks the specified content can be determined as the instruction dimension score based on the consistency between speech content features and text content; and the probability that the user actually performing the verification action is a real person can be determined as the liveness dimension score based on the consistency between lip movement features and standard lip change trends.

[0081] Step S610: Weighted fusion of user dimension score, instruction dimension score and liveness dimension score to determine the comprehensive test result at the same time point.

[0082] In implementation, corresponding weights can be set for the user dimension, command dimension, and liveness dimension. The user dimension score, command dimension score, and liveness dimension score are weighted and summed according to the weights corresponding to different dimensions, and the summation result is determined as the comprehensive test result at that time point.

[0083] In addition, while performing parallel tests on voiceprint features, speech content features, and lip movement features, the extracted facial features can also be tested. By testing whether the facial features match the reserved facial data, it can be confirmed whether the user who actually performs the verification action is the claimed user. Then, by combining the test results of facial features and voiceprint features, the probability that the user who actually performs the verification action is the claimed user can be re-determined, and a user dimension score can be obtained.

[0084] In the above or below embodiments, other authentication methods can be introduced on the basis of the above authentication methods to achieve multi-factor authentication of user identity.

[0085] In one alternative implementation, the behavioral characteristics of the target user can be determined based on the operation behavior related to the verification event performed by the target user; the habitual characteristics of the target user can be obtained, and the compatibility between the behavioral characteristics and habitual characteristics can be tested, so as to generate a hidden verification result of the target user based on the test result. When both the hidden verification result and the identity verification result show that the verification is successful, it is determined that the target user has passed the identity verification.

[0086] Among them, habitual characteristics can be summarized based on the target user's historical operational behavior, or preset by the target user.

[0087] In implementation, the system can collect the target user's actions related to the verification event, including the actions that trigger the verification event and the actions performed during the verification process. Behavioral characteristics of the target user can then be extracted from the collected actions, including typing rhythm, swiping speed, and grip angle. Habitual characteristics of the target user can also be obtained, including typing rhythm, swiping speed, and grip angle. By verifying the compatibility between behavioral and habitual characteristics, the system can determine whether the operator is the user based on their unique operating habits. Furthermore, the target user is only considered successfully verified if both the hidden verification result and the identity verification result show as successful. If either verification method fails, a secondary verification is triggered.

[0088] It should be noted that the target user is unaware of this implicit verification. Implicit verification can be performed simultaneously with the verification features to improve verification efficiency. Furthermore, behavioral features are unconscious, continuous, and highly personalized. It is difficult for attackers to completely replicate the muscle memory that a person has developed over many years. Through this implicit verification, the accuracy of identity verification can be effectively improved.

[0089] In another alternative implementation, when the authentication result shows that the authentication has passed, the target user's authentication result can be transmitted to the hardware layer, so that the hardware layer can sign the authentication result using the built-in private key; the signature result returned by the hardware layer is received, and the correctness of the signature result is verified to generate a hardware verification result. When the hardware verification result shows that the authentication has passed, it is determined that the target user has passed the authentication.

[0090] In implementation, after receiving the authentication result, the hardware layer can independently verify it again. If the verification passes, it signs the authentication result using its built-in private key. It's important to note that the hardware layer does not re-verify features such as voiceprint, facial features, voice content, or lip movements; rather, it verifies that the received content is complete and untampered with. Upon receiving the signature result from the hardware layer, the public key can be used to verify the signature's correctness, generating a hardware verification result. If the signature result can be decrypted using the public key, the hardware verification result is considered successful. When both the hardware verification result and the authentication result indicate successful verification, the target user is considered authenticated. If the hardware verification result fails, secondary verification can be triggered.

[0091] Of course, the target user's identity can also be dual-verified based on the identity verification result, the hidden verification result, and the hardware verification result. When both the identity verification result and the hidden verification result show that the verification is successful, the hardware verification is performed, and when all three verification results show that the verification is successful, the target user is confirmed to have passed the identity verification.

[0092] The following describes in detail an authentication method provided by this specification, using specific application scenarios. In this method, an authentication toolkit can be installed on the client side. The client authenticates the user by calling the authentication toolkit, and the client can interact with the authentication server through a gateway. The authentication event triggered by the target user is a high-risk transaction event, and the risk level of this authentication event is high. For example... Figure 7 As shown, the method may specifically include the following steps: Step S1: The target user initiates a high-risk transaction (such as a large transfer) on the client. Step S2: The gateway identifies high-risk transactions and generates an authentication request, which is then sent to the authentication server. Step S3: The identity verification server randomly generates a dynamic verification command adapted to the high-risk level (such as reading "Departing from Baidi in the morning amidst colorful clouds"), and sends the dynamic verification command and the configuration of the collection parameters to the identity verification toolkit in the client. Step S4: The identity verification toolkit displays dynamic verification instructions to the target user, so that the target user needs to read the dynamic verification instructions displayed on the screen in real time. Step S5: The identity verification toolkit determines the first feature dimension corresponding to the verification event based on the preset mapping relationship between risk level and feature dimension, determines the second feature dimension corresponding to the verification event based on the device information of the target user's device, and collects the data stream corresponding to the common feature dimension between the first feature dimension and the second feature dimension: video stream (including face and lip movement) and audio stream (including voice content and voiceprint features) based on the common feature dimension between the first feature dimension and the second feature dimension, and encrypts and uploads the collected data stream to the identity verification server. Step S6: The identity verification server performs deep identity verification based on the collected data stream and integrates the verification results from multiple dimensions to obtain the identity verification result of the target user. The verification process includes: verifying whether the current face is consistent with the reserved facial data, verifying whether the current voice features match the user's reserved voiceprint, converting the collected audio into text and verifying whether it is consistent with the issued random Tang poem command, using a multimodal model to analyze the trend of mouth shape changes in the video and the issued text command, and whether the collected speech spectrum is highly aligned. Step S7: The identity verification server returns the authentication result to the gateway through the gateway; Step S8: The identity verification toolkit identifies the identity verification result and returns the corresponding operation instructions to the user. If the verification is successful, it returns an instruction to continue the transaction; if the verification fails, it returns an instruction to stop the transaction.

[0093] The above are the authentication methods provided in the embodiments of this specification. Based on the same idea, the embodiments of this specification also provide an authentication device, such as... Figure 8 As shown.

[0094] The authentication device includes: a request parsing module 801, a feature processing module 802, and an authentication module 803, wherein: The request parsing module 801, in response to the identity verification request triggered by the target user for the verification event, determines the risk level corresponding to the verification event and generates a dynamic verification instruction adapted to the risk level. The feature processing module 802 determines the first feature dimension mapped to the risk level corresponding to the verification event based on the preset mapping relationship between risk level and feature dimension; based on the device information of the target user's device, it determines the feature dimension matching the device information as the second feature dimension corresponding to the verification event; it acquires the verification video and / or verification audio generated by the user performing verification actions under the guidance of the dynamic verification instruction, and extracts at least one verification feature corresponding to the common feature dimension from the verification audio and / or the verification video based on the common feature dimension between the first feature dimension and the second feature dimension; The identity verification module 803 uses the at least one verification feature to verify the identity of the target user and obtain the identity verification result of the target user.

[0095] In this embodiment of the specification, the request parsing module 801 includes: In response to an authentication request triggered by a target user for a verification event, at least one risk assessment dimension adapted to the verification event is determined; Calculate the risk score of the verification event under each risk assessment dimension, and determine the risk level corresponding to the verification event based on the risk scores corresponding to multiple risk assessment dimensions.

[0096] In this embodiment of the specification, the request parsing module 801 includes: The event type of the verification event is determined, and the event risk score of the verification event is determined based on the risk assessment rules corresponding to the event type. Determine the device used by the target user when triggering the verification event, assess the security of the device, and determine the device risk score of the verification event based on the assessment results; The risky operational behaviors performed by the target user within a preset historical period are obtained, and the behavioral risk score of the verification event is determined based on the risk level and frequency of the risky operational behaviors. The risk level corresponding to the verification event is determined based on the event hazard score, the equipment hazard score, and the behavior risk score.

[0097] In the implementation of this specification, when the risk level corresponding to the verification event is high-risk, the common feature dimensions include voiceprint feature dimension, speech content feature dimension, facial feature dimension, and lip movement feature dimension. The feature processing module 802 includes: Extract the voiceprint features and speech content features of the target user from the verification audio; Extract the facial features and lip movement features of the target user from the verification video; The voiceprint feature, the speech content feature, the facial feature, and the lip movement feature are identified as at least one verification feature corresponding to the common feature dimension.

[0098] In this embodiment of the specification, the authentication module 803 includes: The time alignment unit extracts the time sequence of the verification action from the voiceprint features and determines the time sequence as the time axis; it determines the original timestamps corresponding to the speech content features and the lip movement features respectively; based on the determined original timestamps, it maps the speech content features and the lip movement features onto the time axis to obtain a time-aligned multimodal feature sequence. The verification unit verifies the identity of the target user based on the multimodal feature sequence to obtain the identity verification result of the target user.

[0099] In the embodiments of this specification, the verification unit includes: Obtain the corresponding verification logic for each of the voiceprint features, speech content features, and lip movement features; Simultaneously, the verification logic corresponding to the voiceprint features, speech content features, and lip movement features is invoked to perform parallel verification of the voiceprint features, speech content features, and lip movement features at the same time point in the multimodal feature sequence. Based on the verification results of multiple verification features at the same time point, the comprehensive verification result at the same time point is determined. The comprehensive test results at multiple time points in the multimodal feature sequence are fused and decided to obtain the identity verification result of the target user.

[0100] In the embodiments of this specification, the verification unit includes: Extract the target user's identity data from a pre-set user database, the identity data including reserved voiceprint data; Extract text content from the dynamic verification instruction and determine the standard lip change trend corresponding to the text content; Simultaneously, it is verified whether the voiceprint features at the same time point are compatible with the reserved voiceprint data, whether the speech content features are consistent with the text content, and whether the lip movement features are consistent with the standard lip change trend. The user dimension score is determined based on the test results for the voiceprint features, the command dimension score is determined based on the test results for the speech content features, and the liveness dimension score is determined based on the test results for the lip movement features. The user dimension score, the instruction dimension score, and the liveness dimension score are weighted and fused to determine the comprehensive test result at the same time point.

[0101] In the embodiments described in this specification, the authentication device further includes an auxiliary authentication module, which includes: The habitual characteristics of the target user are obtained, which are summarized based on the target user's historical operational behavior or preset by the target user. Based on the actions performed by the target user in relation to the verification event, determine the behavioral characteristics of the target user; The compatibility between the behavioral features and the habit features is tested, and a hidden verification result for the target user is generated based on the test results. When both the hidden verification result and the identity verification result show that the verification is successful, it is determined that the target user has passed the identity verification.

[0102] In the embodiments of this specification, the auxiliary verification module includes: When the authentication result shows that the authentication is successful, the authentication result of the target user is transmitted to the hardware layer so that the hardware layer can sign the authentication result using a built-in key; Receive the signature result returned by the hardware layer, and generate a hardware verification result by verifying the correctness of the signature result; When the hardware verification result shows that the verification is successful, it is determined that the target user has passed the authentication.

[0103] For ease of description, the above devices are described by dividing them into various modules or units based on their functions. Of course, when implementing one or more embodiments of this specification, the functions of each module or unit can be implemented in one or more software and / or hardware components, or a module that performs the same function can be implemented by a combination of multiple sub-modules or sub-units, etc. The device embodiments described above are merely illustrative; the division of each module and unit is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or modules can be combined or integrated into another system, or some features can be ignored or not executed, etc.

[0104] This specification provides an identity verification device that, in response to an identity verification request triggered by a target user in response to a verification event, determines the risk level corresponding to the verification event and generates a dynamic verification instruction adapted to the risk level. It acquires verification video and / or verification audio generated when the target user performs a verification action guided by the dynamic verification instruction. Based on a preset mapping relationship between risk levels and feature dimensions, it determines the first feature dimension mapped to the risk level of the verification event. Then, based on the device information of the target user's device, it determines the feature dimension matching the device information as the second feature dimension corresponding to the verification event. Finally, based on the shared feature dimension between the first and second feature dimensions, it extracts at least one verification feature from the verification audio and / or verification video to verify the target user's identity using at least one verification feature, thus obtaining the target user's identity verification result. In this way, by identifying the verification event, determining its risk level, and adaptively determining the corresponding verification feature for the target user based on the determined risk level, it reduces verification errors caused by using general verification methods and effectively improves the accuracy of identity verification.

[0105] The above are the authentication devices provided in the embodiments of this specification. Based on the same idea, the embodiments of this specification also provide an authentication device, such as... Figure 9 As shown.

[0106] The authentication device can be a terminal device or a server, as described in the above embodiments.

[0107] Authentication devices can vary significantly in configuration and performance, and may include a communication interface 902, a user interface 904, a processor 906, and data storage 908. These components are interconnected and communicate with each other via a system bus, network, or other connection mechanism 910. The communication interface 902 enables the authentication device 900 to communicate with other devices, access networks, and transmission networks via analog or digital modulation. For example, the communication interface 902 may include a chipset and antenna for wireless communication with a radio access network or access point. Furthermore, the communication interface 902 can be a wired interface such as Ethernet, Token Ring, or a USB port, or a wireless interface such as Wi-Fi, Bluetooth, Global Positioning System (GPS), or a wide-area wireless interface (e.g., WiMAX or LTE). Of course, the communication interface 902 may also support other forms of physical layer interfaces and standard or proprietary communication protocols. The communication interface 902 may also include multiple physical communication interfaces, such as Wi-Fi, Bluetooth, and wide-area wireless interfaces.

[0108] User interface 904 includes receiving user input and providing output to the user. Therefore, user interface 904 may include input components such as a keypad, keyboard, touch-sensitive or presence-sensitive panel, computer mouse, trackball, joystick, microphone, still camera, and video camera, and output components such as a display screen (which may be combined with a touch-sensitive panel), CRT, LCD, LED, display using DLP technology, printer, and other similar devices known or developed in the future. User interface 904 may also generate auditory output via speakers, speaker jacks, audio output ports, audio output devices, headphones, and other similar devices known or developed in the future. In some embodiments, user interface 904 may include software, circuitry, or other forms of logic capable of transmitting and receiving data from external user input / output devices. Additionally or alternatively, authentication device 900 may support remote access from other devices via communication interface 902 or another physical interface (not shown). User interface 904 may be configured to receive user input, the position and movement of which may be indicated by indicators or cursors described herein. User interface 904 may also be configured as a display device for rendering or displaying text fragments.

[0109] Processor 906 may contain one or more general-purpose processors and / or special-purpose processors.

[0110] Data storage 908 may include one or more volatile and / or non-volatile storage components and may be integrated wholly or partially with processor 906. Data storage 908 may include removable and non-removable components.

[0111] Processor 906 is capable of executing program instructions 918 (e.g., compiled or uncompiled program logic and / or machine code) stored in data storage 908 to perform the various functions described herein. Data storage 908 may contain a non-transitory computer-readable medium on which program instructions are stored, which, when executed by authentication device 900, enable authentication device 900 to perform any methods, processes, or functions disclosed in this specification and / or the accompanying drawings. Execution of program instructions 918 by processor 906 may result in processor 906 using data 912.

[0112] For example, program instructions 918 may include an operating system 922 (e.g., an operating system kernel, device drivers, and / or other modules) installed on authentication device 900, and one or more applications 920 (e.g., a browser, social application, or game application). Similarly, data 912 may include operating system data 916 and application data 914. Operating system data 916 is primarily accessible to operating system 922, while application data 914 is primarily accessible to one or more applications 920. Application data 914 may reside in a file system visible or hidden from the user of authentication device 900.

[0113] Application 920 can communicate with operating system 912 through one or more application programming interfaces (APIs). These APIs help application 920 read and / or write application data 914, transmit or receive information via communication interface 902, receive or display information on user interface 904, etc.

[0114] In some terminology, application 920 may be simply referred to as "app". Furthermore, application 920 can be downloaded to authentication device 900 through one or more online app stores or app markets. However, applications can also be installed on authentication device 900 in other ways, such as through a web browser or a physical interface on authentication device 900 (e.g., a USB port).

[0115] Specifically, in this embodiment, the authentication device 900 includes a data storage 908 and one or more program instructions 918, wherein one or more program instructions 918 are stored in the data storage 908, and one or more program instructions 918 are configured to be executed by one or more processors. The one or more program instructions include computer-executable instructions for performing the following: In response to an authentication request triggered by a target user for an authentication event, the risk level corresponding to the authentication event is determined, and a dynamic authentication instruction adapted to the risk level is generated. Based on the preset mapping relationship between risk levels and feature dimensions, the first feature dimension mapped to the risk level corresponding to the verification event is determined. Based on the device information of the target user's device, the feature dimension that matches the device information is determined as the second feature dimension corresponding to the verification event; The system acquires the verification video and / or verification audio generated by the target user performing a verification action under the guidance of the dynamic verification instruction, and extracts at least one verification feature corresponding to the common feature dimension from the verification audio and / or the verification video based on the common feature dimension between the first feature dimension and the second feature dimension. The identity of the target user is verified using the at least one verification feature to obtain the identity verification result of the target user.

[0116] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on its differences from other embodiments. In particular, the authentication device embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0117] This specification provides an identity verification device that, in response to an identity verification request triggered by a target user in response to a verification event, determines the risk level corresponding to the verification event and generates a dynamic verification instruction adapted to the risk level. It acquires verification video and / or verification audio generated by the target user performing verification actions under the guidance of the dynamic verification instruction. Based on a preset mapping relationship between risk levels and feature dimensions, it determines the first feature dimension mapped to the risk level of the verification event. Then, based on the device information of the target user's device, it determines the feature dimension matching the device information as the second feature dimension corresponding to the verification event. Finally, based on the shared feature dimension between the first and second feature dimensions, it extracts at least one verification feature from the verification audio and / or verification video to verify the target user's identity using at least one verification feature, thus obtaining the target user's identity verification result. In this way, by identifying the verification event, determining its risk level, and adaptively determining the corresponding verification feature for the target user based on the determined risk level, it reduces verification errors caused by using general verification methods and effectively improves the accuracy of identity verification.

[0118] Furthermore, based on the above Figures 1 to 7 This specification also provides a storage medium for storing computer-executable instruction information in one or more embodiments. In one specific embodiment, the storage medium may be a USB flash drive, optical disc, hard disk, etc. When the computer-executable instruction information stored in the storage medium is executed by a processor, it can realize the following process: In response to an authentication request triggered by a target user for an authentication event, the risk level corresponding to the authentication event is determined, and a dynamic authentication instruction adapted to the risk level is generated. Based on the preset mapping relationship between risk levels and feature dimensions, the first feature dimension mapped to the risk level corresponding to the verification event is determined. Based on the device information of the target user's device, the feature dimension that matches the device information is determined as the second feature dimension corresponding to the verification event; The system acquires the verification video and / or verification audio generated by the target user performing a verification action under the guidance of the dynamic verification instruction, and extracts at least one verification feature corresponding to the common feature dimension from the verification audio and / or the verification video based on the common feature dimension between the first feature dimension and the second feature dimension. The identity of the target user is verified using the at least one verification feature to obtain the identity verification result of the target user.

[0119] The various embodiments in this specification 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 above-described storage medium embodiment is basically similar to the method embodiment, so the description is relatively simple; relevant parts can be referred to the description of the method embodiment.

[0120] This specification provides a storage medium that, in response to an authentication request triggered by a target user in response to a verification event, determines the risk level corresponding to the verification event and generates a dynamic verification instruction adapted to the risk level. It acquires verification video and / or verification audio generated by the target user performing verification actions under the guidance of the dynamic verification instruction. Based on a preset mapping relationship between risk levels and feature dimensions, it determines the first feature dimension mapped to the risk level of the verification event. Then, based on the device information of the target user's device, it determines the feature dimension matching the device information as the second feature dimension corresponding to the verification event. Finally, based on the shared feature dimension between the first and second feature dimensions, it extracts at least one verification feature from the verification audio and / or verification video to verify the target user's identity using at least one verification feature, thus obtaining the target user's authentication result. In this way, by identifying the verification event, determining its risk level, and adaptively determining the corresponding verification feature for the target user based on the determined risk level, it reduces verification errors caused by using general verification methods and effectively improves the accuracy of authentication.

[0121] Furthermore, based on the above Figures 1 to 7 This specification also provides one or more embodiments of a computer program product, including a computer program, which, when executed by a processor, can perform the following processes: In response to an authentication request triggered by a target user for an authentication event, the risk level corresponding to the authentication event is determined, and a dynamic authentication instruction adapted to the risk level is generated. Based on the preset mapping relationship between risk levels and feature dimensions, the first feature dimension mapped to the risk level corresponding to the verification event is determined. Based on the device information of the target user's device, the feature dimension that matches the device information is determined as the second feature dimension corresponding to the verification event; The system acquires the verification video and / or verification audio generated by the target user performing a verification action under the guidance of the dynamic verification instruction, and extracts at least one verification feature corresponding to the common feature dimension from the verification audio and / or the verification video based on the common feature dimension between the first feature dimension and the second feature dimension. The identity of the target user is verified using the at least one verification feature to obtain the identity verification result of the target user.

[0122] The various embodiments in this specification 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 above-described embodiment of a computer program product is relatively simple in description because it is fundamentally similar to the method embodiment; relevant parts can be referred to the description of the method embodiment.

[0123] This specification provides a computer program product that, in response to an authentication request triggered by a target user in response to a verification event, determines the risk level corresponding to the verification event and generates a dynamic verification instruction adapted to the risk level. It acquires verification video and / or verification audio generated when the target user performs a verification action guided by the dynamic verification instruction. Based on a preset mapping relationship between risk levels and feature dimensions, it determines the first feature dimension mapped to the risk level of the verification event. Then, based on the device information of the target user's device, it determines the feature dimension matching the device information as the second feature dimension corresponding to the verification event. Next, based on the shared feature dimension between the first and second feature dimensions, it extracts at least one verification feature from the verification audio and / or verification video to verify the target user's identity using at least one verification feature, thus obtaining the target user's authentication result. In this way, by identifying the verification event, determining its risk level, and adaptively determining the corresponding verification feature for the target user based on the determined risk level, it reduces verification errors caused by using general verification methods and effectively improves the accuracy of authentication.

[0124] The foregoing has described specific embodiments of this specification. 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 still achieve the desired results. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are possible or may be advantageous. Moreover, although one or more embodiments of this specification provide method steps as described in the embodiments or flowcharts, it is understood that the order of steps listed in the embodiments or flowcharts is merely one possible execution order among many steps and does not represent the only execution order. Therefore, when method steps are involved in the claims, adjustments to the order of those steps, or parallelism between steps, are also within the scope of protection of the claims.

[0125] 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 and "integrate" a digital system onto a PLD themselves, without needing chip manufacturers to design and manufacture dedicated integrated circuit chips. Furthermore, nowadays, instead of manually manufacturing integrated circuit chips, this programming is mostly implemented using "logic compiler" software. Similar to the software compiler used in program development, the original code before compilation must also be written in a specific programming language, called a Hardware Description Language (HDL). There are many HDLs, 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, and RHDL (Ruby Hardware Description Language). 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 one of these hardware description languages ​​and programming it into an integrated circuit, the hardware circuit implementing the logical method flow can be easily obtained.

[0126] 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: ARC 625D, 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, application-specific integrated circuits, 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.

[0127] 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.

[0128] For ease of description, the above apparatus is described by dividing it into various functional units. Of course, when implementing one or more embodiments of this specification, the functions of each unit can be implemented in one or more software and / or hardware.

[0129] Those skilled in the art will understand that the embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, one or more embodiments of this specification may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, one or more embodiments of this specification may take the form of a computer program product 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.

[0130] Embodiments in this specification are described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this specification. 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 parallel device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable parallel device, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0131] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable fraud device to operate 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.

[0132] These computer program instructions can also be loaded onto a computer or other programmable device, causing a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable device 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.

[0133] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0134] 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 RAM. Memory is an example of computer-readable media.

[0135] Computer-readable media include both permanent and non-permanent, removable and non-removable media that can store information using 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, 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.

[0136] 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 limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical or equivalent elements in the process, method, article, or apparatus that includes said element. Furthermore, "a," "an," and "the" are not specifically singular and may include plural forms. Ordinal numbers such as "first," "second," etc., do not necessarily indicate order; they are often used to distinguish objects. For example, "first server" and "second server" usually refer to two servers, described as "first server" and "second server" to differentiate them; however, sometimes these two servers may be the same server. Moreover, in this specification, unless explicitly stated otherwise, "receiving and sending data" does not necessarily mean direct receiving and sending; it can be indirect receiving and sending (i.e., receiving and sending indirectly through one or more entities). Similarly, in this specification, unless otherwise stated, the relationships between structures can be direct or indirect.

[0137] Furthermore, the specific terms used in this specification to describe embodiments, such as "an embodiment," "one embodiment," or "some embodiments," refer to a particular feature, structure, or characteristic related to at least one embodiment of this specification. Therefore, it should be emphasized and noted that "an embodiment," "one embodiment," or "an alternative embodiment" mentioned twice or more in different locations in this specification do not necessarily refer to the same embodiment. Moreover, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples, without contradiction.

[0138] Those skilled in the art will understand that the embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, one or more embodiments of this specification may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, one or more embodiments of this specification may take the form of a computer program product 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.

[0139] One or more embodiments of this specification 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. One or more embodiments of this specification can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0140] The various embodiments in this specification 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 in the method embodiments.

[0141] The above description is merely an embodiment of this specification and is not intended to limit this document. Various modifications and variations can be made to this specification by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this specification should be included within the scope of the claims in this document.

Claims

1. An authentication method, the method comprising: In response to an authentication request triggered by a target user for an authentication event, the risk level corresponding to the authentication event is determined, and a dynamic authentication instruction adapted to the risk level is generated. Based on the preset mapping relationship between risk levels and feature dimensions, the first feature dimension mapped to the risk level corresponding to the verification event is determined. Based on the device information of the target user's device, the feature dimension that matches the device information is determined as the second feature dimension corresponding to the verification event; The system acquires the verification video and / or verification audio generated by the target user performing a verification action under the guidance of the dynamic verification instruction, and extracts at least one verification feature corresponding to the common feature dimension from the verification audio and / or the verification video based on the common feature dimension between the first feature dimension and the second feature dimension. The identity of the target user is verified using the at least one verification feature to obtain the identity verification result of the target user.

2. The method according to claim 1, wherein determining the risk level corresponding to the verification event in response to an authentication request triggered by a target user for a verification event includes: In response to an authentication request triggered by a target user for a verification event, at least one risk assessment dimension adapted to the verification event is determined; Calculate the risk score of the verification event under each risk assessment dimension, and determine the risk level corresponding to the verification event based on the risk scores corresponding to multiple risk assessment dimensions.

3. The method according to claim 2, wherein calculating the risk score of the verification event under each risk assessment dimension, and determining the risk level corresponding to the verification event based on the risk scores corresponding to multiple risk assessment dimensions, includes: The event type of the verification event is determined, and the event risk score of the verification event is determined based on the risk assessment rules corresponding to the event type. Determine the device used by the target user when triggering the verification event, assess the security of the device, and determine the device risk score of the verification event based on the assessment results; The risky operational behaviors performed by the target user within a preset historical period are obtained, and the behavioral risk score of the verification event is determined based on the risk level and frequency of the risky operational behaviors. The risk level corresponding to the verification event is determined based on the event hazard score, the equipment hazard score, and the behavior risk score.

4. The method according to claim 1, wherein the common feature dimension includes a voiceprint feature dimension, a speech content feature dimension, a facial feature dimension, and a lip movement feature dimension, and the step of extracting at least one verification feature corresponding to the common feature dimension from the verification audio and / or the verification video includes: Extract the voiceprint features and speech content features of the target user from the verification audio; Extract the facial features and lip movement features of the target user from the verification video; The voiceprint feature, the speech content feature, the facial feature, and the lip movement feature are identified as at least one verification feature corresponding to the common feature dimension.

5. The method according to claim 4, wherein verifying the identity of the target user using the at least one verification feature to obtain the identity verification result of the target user includes: Extract the time sequence of the verification action from the voiceprint features, and determine the time sequence as the time axis; Determine the original timestamps corresponding to the speech content features and the lip movement features respectively; Based on the determined original timestamp, the speech content features and the lip movement features are mapped onto the time axis to obtain a time-aligned multimodal feature sequence; The identity of the target user is verified based on the multimodal feature sequence to obtain the identity verification result of the target user.

6. The method according to claim 5, wherein verifying the identity of the target user based on the multimodal feature sequence to obtain the identity verification result of the target user includes: Obtain the corresponding verification logic for each of the voiceprint features, speech content features, and lip movement features; Simultaneously, the verification logic corresponding to the voiceprint features, speech content features, and lip movement features is invoked to perform parallel verification of the voiceprint features, speech content features, and lip movement features at the same time point in the multimodal feature sequence. Based on the verification results of multiple verification features at the same time point, the comprehensive verification result at the same time point is determined. The comprehensive test results at multiple time points in the multimodal feature sequence are fused and decided to obtain the identity verification result of the target user.

7. The method according to claim 6, wherein the parallel verification of the voiceprint features, speech content features, and lip movement features at the same time point in the multimodal feature sequence, and the determination of the comprehensive verification result at the same time point based on the verification results of multiple verification features at the same time point, includes: Extract the target user's identity data from a pre-set user database, the identity data including reserved voiceprint data; Extract text content from the dynamic verification instruction and determine the standard lip change trend corresponding to the text content; Simultaneously, it is verified whether the voiceprint features at the same time point are compatible with the reserved voiceprint data, whether the speech content features are consistent with the text content, and whether the lip movement features are consistent with the standard lip change trend. The user dimension score is determined based on the test results for the voiceprint features, the command dimension score is determined based on the test results for the speech content features, and the liveness dimension score is determined based on the test results for the lip movement features. The user dimension score, the instruction dimension score, and the liveness dimension score are weighted and fused to determine the comprehensive test result at the same time point.

8. The method according to claim 1, further comprising: The habitual characteristics of the target user are obtained, which are summarized based on the target user's historical operational behavior or preset by the target user. Based on the actions performed by the target user in relation to the verification event, determine the behavioral characteristics of the target user; The compatibility between the behavioral features and the habit features is tested, and a hidden verification result for the target user is generated based on the test results. When both the hidden verification result and the identity verification result show that the verification is successful, it is determined that the target user has passed the identity verification.

9. The method according to claim 1, further comprising: When the authentication result shows that the authentication is successful, the authentication result of the target user is transmitted to the hardware layer so that the hardware layer can sign the authentication result using a built-in key; Receive the signature result returned by the hardware layer, and generate a hardware verification result by verifying the correctness of the signature result; When the hardware verification result shows that the verification is successful, it is determined that the target user has passed the authentication.

10. An identity verification device, the device comprising: The request parsing module, in response to the identity verification request triggered by the target user for the verification event, determines the risk level corresponding to the verification event and generates a dynamic verification instruction that is adapted to the risk level. The feature processing module determines the first feature dimension mapped to the risk level corresponding to the verification event based on the preset mapping relationship between risk level and feature dimension. Based on the device information of the target user's device, the feature dimension that matches the device information is determined as the second feature dimension corresponding to the verification event; the verification video and / or verification audio generated by the user performing a verification action under the guidance of the dynamic verification instruction are obtained, and based on the common feature dimension between the first feature dimension and the second feature dimension, at least one verification feature corresponding to the common feature dimension is extracted from the verification audio and / or the verification video. The identity verification module uses the at least one verification feature to verify the identity of the target user and obtain the identity verification result of the target user.

11. An identity verification device, the device comprising: processor; as well as A memory configured to store computer-executable instructions, which, when executed, cause the processor to: In response to an authentication request triggered by a target user for an authentication event, the risk level corresponding to the authentication event is determined, and a dynamic authentication instruction adapted to the risk level is generated. Based on the preset mapping relationship between risk levels and feature dimensions, the first feature dimension mapped to the risk level corresponding to the verification event is determined. Based on the device information of the target user's device, the feature dimension that matches the device information is determined as the second feature dimension corresponding to the verification event; The system acquires the verification video and / or verification audio generated by the target user performing a verification action under the guidance of the dynamic verification instruction, and extracts at least one verification feature corresponding to the common feature dimension from the verification audio and / or the verification video based on the common feature dimension between the first feature dimension and the second feature dimension. The identity of the target user is verified using the at least one verification feature to obtain the identity verification result of the target user.

Images