CN110516416B

CN110516416B - Identity authentication method, authentication end and client

Info

Publication number: CN110516416B
Application number: CN201910723081.0A
Authority: CN
Inventors: 谢韬
Original assignee: China Mobile Communications Group Co Ltd; MIGU Culture Technology Co Ltd
Current assignee: China Mobile Communications Group Co Ltd; MIGU Culture Technology Co Ltd
Priority date: 2019-08-06
Filing date: 2019-08-06
Publication date: 2021-08-06
Anticipated expiration: 2039-08-06
Also published as: CN110516416A

Abstract

The embodiment of the invention provides an identity authentication method, an authentication end and a client, wherein the method comprises the following steps: if a verification request sent by a client is received, acquiring an identified text which completes emotional tendency identification; sending the recognized text to the client; receiving a client emotional tendency recognition result for the recognized text returned by the client; and determining a verification result based on the marked emotional tendency recognition result of the recognized text and the client emotional tendency recognition result. According to the method, the verification end and the client provided by the embodiment of the invention, the emotional tendency recognition result of the client of the recognized text returned by the client is received for verification, the limitation is small, the user recognition difficulty is low, and the emotional tendency recognition is used as a verification mode due to the high research and development cost of the emotional tendency recognition, so that the verification threshold of the machine script can be improved, and the safety of identity verification is further improved.

Description

Identity authentication method, authentication end and client

Technical Field

The invention relates to the technical field of internet security, in particular to an identity authentication method, an authentication end and a client.

Background

With the development of computer technology, more and more scripts are applied to situations such as ticket swiping, ticket robbing, batch registration, brute force cracking and the like, so that the server burden is increased, the product experience is damaged, and a serious safety problem is brought.

In order to prevent the machine script operation problem, the identity verification method such as verification codes is applied to high-security level operations such as website, APP login and registration. The identity authentication method can provide the problem that the machine script cannot process but the real user can answer, so that whether the operation is human or machine is distinguished, and the machine is prevented from accessing the system. The existing identity authentication method comprises a short message authentication code, a graphic authentication code, a selection graphic authentication and the like, wherein the short message authentication code can generate expenses and increase the operation cost, and is strongly dependent on a mobile phone, so that the method has limitations; the safety of the graphic verification code is not high, and the safety is improved by adding interference pixels, and meanwhile, the difficulty is brought to the identification of a real user; the gallery updating of the selected graph check depends on manual work, the number of the pictures is not large, the pictures are easy to traverse and completely identified, the safety is low, and the user identification is difficult.

Therefore, how to improve the security of the authentication while reducing the limitations and the difficulty of user identification remains a problem to be solved urgently by those skilled in the art.

Disclosure of Invention

The embodiment of the invention provides an identity authentication method, an authentication end and a client, which are used for solving the problems of poor safety, strong limitation and difficult user identification of the existing identity authentication method.

In a first aspect, an embodiment of the present invention provides an identity authentication method, including:

if a verification request sent by a client is received, acquiring an identified text which completes emotional tendency identification;

sending the recognized text to the client;

receiving a client emotional tendency recognition result for the recognized text returned by the client;

and determining a verification result based on the marked emotional tendency recognition result of the recognized text and the client emotional tendency recognition result.

Preferably, the method further comprises:

if a verification request sent by a client is received, acquiring a text to be recognized, which is not subjected to emotion tendency recognition;

sending the text to be recognized to the client;

receiving a client emotional tendency recognition result of the text to be recognized returned by the client;

if the verification result is that the verification is passed, storing the client emotional tendency recognition result of the text to be recognized into the client emotional tendency set of the text to be recognized;

and if the client emotional tendency set meets the preset identification condition, determining a calibration emotional tendency identification result of the text to be identified based on the client emotional tendency set, and updating the text to be identified into the identified text.

Preferably, the determining a verification result based on the calibration emotional tendency recognition result of the recognized text and the client emotional tendency recognition result specifically includes:

calculating the recognition accuracy rate based on the marked emotional tendency recognition result of the recognized text and the client emotional tendency recognition result;

if the identification accuracy is greater than or equal to a preset accuracy threshold, determining that the verification result is that the verification is passed; otherwise, determining that the verification result is verification failure.

Preferably, if a verification request sent by a client is received, acquiring a recognized text for which emotion tendency recognition is completed, the method further includes:

acquiring an initial text;

inputting the initial text into an emotional tendency recognition model, and acquiring a model emotional tendency recognition result and a confidence coefficient output by the emotional tendency recognition model; the emotional tendency recognition model is obtained by training based on a sample text and a sample emotional tendency recognition result;

if the confidence degree is greater than or equal to a preset confidence threshold value, marking the initial text as the recognized text, and taking the model emotional tendency recognition result as a calibrated emotional tendency recognition result of the recognized text;

otherwise, the initial text is marked as the text to be recognized.

Preferably, the receiving a client emotional tendency recognition result of the recognized text returned by the client further includes:

counting the recognition error rate of the recognized text;

and if the recognition error rate is greater than a preset recognition error threshold value, correcting the marked emotional tendency recognition result of the recognized text.

Preferably, the nominal emotional tendency recognition result comprises at least one of the following:

positive, neutral, negative.

In a second aspect, an embodiment of the present invention provides an identity authentication method, including:

sending a verification request to a verification end;

receiving the recognized text which is sent by the verifying end in response to the verification request and completes emotional tendency recognition;

acquiring a client emotional tendency recognition result corresponding to the input recognized text;

and returning the client emotional tendency recognition result to the verification end to trigger the verification end to determine the verification result based on the calibration emotional tendency recognition result of the recognized text and the client emotional tendency recognition result.

Preferably, the sending the authentication request to the authentication end further includes:

receiving a text to be identified which is sent by the verification end in response to the verification request and does not finish emotion tendency identification, and acquiring a client emotion tendency identification result corresponding to the input text to be identified;

returning the client emotional tendency recognition result corresponding to the text to be recognized to the verification end to trigger the verification end to execute the following operations:

In a third aspect, an embodiment of the present invention provides a verification end, including:

the text acquisition unit is used for acquiring the recognized text which finishes emotion tendency recognition if a verification request sent by the client is received;

a text sending unit, configured to send the recognized text to the client;

the tendency receiving unit is used for receiving a client emotion tendency recognition result for the recognized text returned by the client;

and the identity verification unit is used for determining a verification result based on the calibration emotional tendency recognition result of the recognized text and the client emotional tendency recognition result.

In a fourth aspect, an embodiment of the present invention provides a client, including:

the request sending unit is used for sending a verification request to the verification end;

a text receiving unit, configured to receive a recognized text that has been subjected to emotion tendency recognition and is sent by the verification end in response to the verification request;

the tendency acquisition unit is used for acquiring a client emotion tendency identification result corresponding to the input identified text;

and the tendency sending unit is used for returning the client emotional tendency recognition result to the verification end so as to trigger the verification end to determine the verification result based on the calibration emotional tendency recognition result of the recognized text and the client emotional tendency recognition result.

In a fifth aspect, an embodiment of the present invention provides an electronic device, including a processor, a communication interface, a memory, and a bus, where the processor and the communication interface, the memory complete communication with each other through the bus, and the processor may call logic instructions in the memory to perform the steps of the method as provided in the first aspect or the second aspect.

In a sixth aspect, embodiments of the present invention provide a non-transitory computer readable storage medium, on which a computer program is stored, which when executed by a processor, implements the steps of the method as provided in the first or second aspect.

According to the identity verification method provided by the embodiment of the invention, the client emotion tendency recognition result of the recognized text returned by the client is received for verification, the limitation is small, the user recognition difficulty is low, and because the research and development cost of emotion tendency recognition is high, the emotion tendency recognition is used as a verification mode, so that the verification threshold of a machine script can be improved, and the security of identity verification is further improved.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, a brief description will be given below of the drawings required for the embodiments or the technical solutions in the prior art, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.

Fig. 1 is a schematic flow chart of an identity authentication method according to an embodiment of the present invention;

fig. 2 is a schematic structural diagram of an authentication system according to an embodiment of the present invention;

fig. 3 is a schematic flow chart of an authentication method according to another embodiment of the present invention;

fig. 4 is a schematic structural diagram of a verification end according to an embodiment of the present invention;

fig. 5 is a schematic structural diagram of a client according to an embodiment of the present invention;

fig. 6 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be obtained by a person skilled in the art without inventive effort based on the embodiments of the present invention, are within the scope of the present invention.

The existing identity authentication method generally has the problems of strong limitation, low safety and high user identification difficulty. Accordingly, the embodiment of the invention provides an identity authentication method. Fig. 1 is a schematic flow chart of an identity authentication method according to an embodiment of the present invention, and as shown in fig. 1, an execution subject of the method is an authentication end, and the method includes:

step 110, if a verification request sent by the client is received, acquiring the recognized text of which the emotional tendency recognition is completed.

Specifically, the recognized text refers to a text whose emotional tendency is recognized and corresponding emotional tendency recognition result (for the convenience of distinguishing and description, the emotional tendency recognition result is referred to as a nominal emotional tendency recognition result) is obtained, that is, the nominal emotional tendency recognition result of the recognized text is known in advance.

And calibrating the emotional tendency recognition result for comparing with a recognition result (called a client emotional tendency recognition result for convenience of distinguishing and description) obtained by performing emotional tendency recognition on the recognized text received by the client by the user.

The client emotion recognition result can be input into the client by a user and then provided to the verification end by the client.

Aiming at the same recognized text, if the comparison result is that the corresponding marked emotional tendency recognition result is the same as the client emotional tendency recognition result, the verification result that the identity verification is passed can be obtained; and if the comparison result is that the corresponding marked emotional tendency identification result is different from the client emotional tendency identification result, the verification result that the identity verification fails can be obtained.

The nominal emotional tendency recognition result may be one of a plurality of predetermined emotional types, such as one of the emotional types "positive", "negative", and "neutral", and further, such as one of the more complicated emotional types "anger", "frustration", "calm", "joy", and "happy". When the client needs to perform authentication, an authentication request may be sent to the authentication end to request the authentication end to perform authentication for a client operating the client. After receiving a verification request sent by the client, the recognized text can be selected from a large number of pre-stored recognized texts. Here, the recognized text selected may be one or more.

Step 120, the recognized text is sent to the client.

Specifically, after the recognized text is acquired, the acquired recognized text is sent to the client side for the client side to perform emotional tendency recognition.

And step 130, receiving a client emotional tendency recognition result for the recognized text returned by the client.

Specifically, after the client receives a plurality of recognized texts, the client operating the client performs emotional tendency recognition on the recognized texts, inputs a result of the emotional tendency recognition on the recognized texts, namely a client emotional tendency recognition result of the recognized texts, to the client, and returns the client emotional tendency recognition result of the recognized texts to the verifying end by the client, so that the verifying end performs verification. And then, the verifying end receives the client emotion tendency recognition result of the recognized text returned by the client. Here, the client emotional tendency recognition result may be one of a plurality of emotion types set in advance.

And step 140, determining a verification result based on the calibration emotional tendency identification result of the identified text and the client emotional tendency identification result.

Specifically, the labeled emotional tendency recognition result of the recognized text may be obtained by performing emotional tendency recognition by a worker in advance, or may be output through a pre-trained emotional tendency recognition model, which is not specifically limited in this embodiment of the present invention. The client emotional tendency recognition result of the recognized text is obtained through the step 130, and the method provided by the embodiment of the invention is executed on the basis of the assumption that the marked emotional tendency recognition result of the recognized text is the actual emotional type of the recognized text.

Therefore, whether the emotional tendency recognition of the client aiming at the recognized text is correct or not can be known based on the calibration emotional tendency recognition result of the recognized text and the client emotional tendency recognition result. And obtaining a verification result after knowing whether the client side correctly identifies the emotional tendency of the identified text.

Here, the verification result may be verification pass or verification fail. The verification result may be obtained in a variety of manners, for example, the verification result is confirmed based on the number of recognized texts that are correctly recognized by the client emotional tendency, or the verification result is confirmed based on the ratio of the number of recognized texts that are correctly recognized by the client emotional tendency to the number of all recognized texts.

According to the method provided by the embodiment of the invention, the client emotion tendency recognition result of the recognized text returned by the client is received for verification, the limitation is small, the user recognition difficulty is low, and the emotion tendency recognition is used as a verification mode due to the high research and development cost of the emotion tendency recognition, so that the verification threshold of the machine script can be improved, and the safety of identity verification is further improved.

With the development of artificial intelligence technology, the technology for recognizing text emotional tendency comes up. However, artificial intelligence needs strong technical capability support, personnel requirements are high, the research and development period is long, and medium-sized and small-sized enterprises are difficult to develop artificial intelligence emotional tendency recognition schemes customized in specific fields. Moreover, human language semantics are complex, artificial intelligent recognition often has inaccurate conditions, and especially in the conditions of wrongly written characters, skip writing, harmonic sounds and the like which may occur in a Chinese text, the text emotional tendency recognition result is inaccurate. Based on the embodiment, the identity authentication method provided by the embodiment of the invention realizes low-cost text emotional tendency recognition by means of an identity authentication process on the basis of the client emotional tendency recognition result of each recognized text returned by the authentication client. The method also comprises the following steps:

step 111, if a verification request sent by a client is received, acquiring a text to be recognized, which does not finish emotion tendency recognition; and step 121, sending the text to be recognized to the client.

Specifically, the text to be recognized refers to a text which is not subjected to emotion tendency recognition and therefore does not obtain a corresponding calibrated emotion tendency recognition result, where the text to be recognized may be a text which cannot be recognized by a pre-trained emotion tendency recognition model, or a text which is output by an emotion tendency recognition model and has a low confidence of a model emotion tendency recognition result, or a text which is not input to an emotion tendency recognition model for recognition, or a text which is not subjected to emotion tendency recognition by a worker.

After receiving a verification request sent by a client, the acquired recognized text is sent to the client, and the acquired text to be recognized is also sent to the client for the client to recognize emotional tendency. In the embodiment of the present invention, step 111 is executed synchronously with step 110, and step 121 is executed synchronously with step 120.

And step 131, receiving a client emotional tendency recognition result of the text to be recognized returned by the client.

Specifically, after the client receives a plurality of recognized texts and a plurality of texts to be recognized, the client operating the client performs emotion tendency recognition on the recognized texts and the texts to be recognized respectively, and returns the client emotion tendency recognition result of the recognized texts and the client emotion tendency recognition result of the texts to be recognized to the verification end. In the embodiment of the present invention, step 131 and step 130 are executed synchronously.

And step 151, if the verification result is that the verification is passed, storing the client emotional tendency recognition result of the text to be recognized into the client emotional tendency set of the text to be recognized.

Specifically, step 151 is performed after determining the verification result in step 140. And the client emotional tendency set of the text to be recognized is used for storing a client emotional tendency recognition result of the text to be recognized returned by the client.

And respectively storing the client emotional tendency recognition results of the text to be recognized returned by the client into the corresponding client emotional tendency sets of the text to be recognized only when the verification result is that the verification is passed. If the verification result is that the verification is not passed, the client side is indicated to be suspected of machine verification, the client emotion tendency recognition result of the text to be recognized returned by the client side is most likely to be a malicious attack from the script, and if the client emotion tendency recognition result of the text to be recognized is stored in the client emotion tendency set, the emotion tendency recognition of the text to be recognized may be interfered, so the client emotion tendency recognition result of the text to be recognized is not stored in the client emotion tendency set.

And step 160, if the client emotional tendency set meets the preset identification condition, determining a calibration emotional tendency identification result of the text to be identified based on the client emotional tendency set, and updating the text to be identified into the identified text.

Specifically, a plurality of client emotional tendency recognition results for the text to be recognized are stored in the client emotional tendency set of the text to be recognized. The several client emotional tendency recognition results may be client emotional tendency recognition results returned by different clients in different authentication processes.

Here, the preset recognition condition is a preset recognition condition. The preset identification condition may be that the number of the client emotional tendency identification results in the client emotional tendency set is greater than a preset number threshold, or that the number of the client emotional tendency identification results in the client emotional tendency set is greater than a preset number threshold and the proportion of any emotion type of the client emotional tendency identification results is greater than a preset type proportion threshold, which is not specifically limited in the embodiment of the present invention.

If the client emotional tendency set meets the preset identification condition, determining a calibration emotional tendency identification result of the corresponding text to be identified based on the client emotional tendency set, obtaining the calibration emotional tendency identification result to represent that the text to be identified is already identified, and re-marking the text to be identified as the identified text. There are various methods for determining the calibrated emotional tendency recognition result of the text to be recognized, for example, directly taking the emotional type with the highest number of the client emotional tendency recognition results in the client emotional tendency set as any emotional type as the calibrated emotional tendency recognition result.

According to the method provided by the embodiment of the invention, when the identity verification is completed, the emotional tendency of the text to be recognized is recognized by the client by means of the identity verification process, and the emotional tendency recognition result of the client of the text to be recognized is obtained, so that the emotional tendency recognition of the text to be recognized is realized, the text emotional tendency recognition cost is reduced, and the text emotional tendency recognition accuracy is improved.

Based on any of the above embodiments, in the method, step 140 specifically includes: acquiring the recognition accuracy based on the marked emotional tendency recognition result of the recognized text and the client emotional tendency recognition result; if the identification accuracy is greater than or equal to a preset accuracy threshold, determining that the verification result is that the verification is passed; otherwise, determining that the verification result is verification failure.

Specifically, whether the emotional tendency recognition of the client side for the recognized text is correct can be obtained by judging whether the calibration emotional tendency recognition result of the recognized text is consistent with the client emotional tendency recognition result. After learning whether the emotional tendency recognition of the client for each recognized text is correct, the recognition accuracy, namely the ratio of the number of recognized texts with correct emotional tendency recognition of the client to the total number of recognized texts, can be obtained. And after the identification accuracy is obtained, determining a verification result by comparing the identification accuracy with a preset accuracy threshold. Here, the preset accuracy threshold is a minimum value of the accuracy of the verification pass that is set in advance. If the identification accuracy is greater than or equal to a preset accuracy threshold, determining that the verification result is that the verification is passed; and if the identification accuracy is smaller than the preset accuracy threshold, determining that the verification result is verification failure.

According to any of the above embodiments, the method further includes, before step 110:

step 101, obtaining an initial text.

Specifically, before sending the recognized text and the text to be recognized to the client, the recognized text and the text to be recognized need to be acquired respectively. Further, the initial text needs to be acquired first. Here, the initial text refers to a text acquired from a text source, and for example, the initial text may be captured from the text source by using a web crawler technology for a preset topic.

Step 102, inputting the initial text into an emotional tendency recognition model, and acquiring a model emotional tendency recognition result and a confidence coefficient output by the emotional tendency recognition model; and the emotional tendency recognition model is obtained by training based on the sample text and the sample emotional tendency recognition result.

Specifically, after the initial text is obtained, the initial text may be input into the emotion tendency recognition model, the emotion tendency recognition model performs emotion tendency recognition on the input initial text, and a model emotion tendency recognition result and confidence are generated and output. Here, the emotion tendency recognition model is a model trained in advance for emotion tendency recognition of the input initial text. The model emotional tendency recognition result is an emotional type generated and output by the emotional tendency recognition model based on the input initial text, and the confidence coefficient is used for indicating the credibility of the model emotional tendency recognition result.

In addition, before step 102 is executed, the emotion tendency recognition model may be obtained by training in advance, and specifically, the emotion tendency recognition model may be obtained by training in the following manner: firstly, collecting a large amount of sample texts and sample emotional tendency recognition results thereof; the sample text is a text which is subjected to emotional tendency recognition and has obtained a corresponding emotional tendency recognition result, the sample emotional tendency recognition result is an emotional tendency recognition result of the sample text, and the sample emotional tendency recognition result can be obtained by performing emotional tendency recognition by a worker in advance. And the sample text corresponds to the sample emotional tendency recognition result one by one. And training the initial model based on the sample text and the sample emotional tendency recognition result, thereby obtaining an emotional tendency recognition model. The initial model may be a single neural network model or a combination of a plurality of neural network models, and the embodiment of the present invention does not specifically limit the type and structure of the initial model.

Step 103, if the confidence degree is greater than or equal to a preset confidence threshold value, updating the initial text into a recognized text, and taking the model emotional tendency recognition result as a calibration emotional tendency recognition result of the recognized text; otherwise, updating the initial text into the text to be recognized.

Specifically, after a model emotional tendency recognition result and a confidence coefficient of any initial text are obtained, the initial text is judged to be a recognized text or a text to be recognized by comparing the confidence coefficient with a preset confidence threshold value. Here, the preset confidence threshold is a minimum value of the confidence corresponding to the recognized text that is set in advance.

If the confidence degree is greater than or equal to a preset confidence threshold value, namely the credibility of the model emotional tendency recognition result of the initial text meets the requirement, confirming that the initial text is the recognized text, and taking the model emotional tendency recognition result as a calibration emotional tendency recognition result of the recognized text;

if the confidence degree is smaller than the preset confidence threshold value, the credibility of the model emotional tendency recognition result of the initial text does not meet the requirement, the model emotional tendency recognition result cannot be used as a calibration emotional tendency recognition result of the initial text, and the initial text is confirmed to be the text to be recognized.

According to the method provided by the embodiment of the invention, the recognized text and the text to be recognized are obtained by setting the emotional tendency recognition model, so that a large amount of available materials are provided for verification.

According to any of the above embodiments, in the method, step 160 further includes: and if the client emotional tendency set of the text to be recognized meets the preset recognition condition, training an emotional tendency recognition model based on the text to be recognized and the calibration emotional tendency recognition result of the text to be recognized.

Specifically, if the client emotional tendency set of the text to be recognized meets the preset recognition condition, the calibrated emotional tendency recognition result of the text to be recognized is determined based on the client emotional tendency recognition result in the client emotional tendency set. After the calibrated emotional tendency recognition result of the text to be recognized is obtained, the text to be recognized and the calibrated emotional tendency recognition result of the text to be recognized can be respectively used as a sample text and a sample emotional tendency recognition result to be input into an emotional tendency recognition model so as to train the emotional tendency recognition model, and the accuracy of the emotional tendency recognition result of the output model of the emotional tendency recognition model is further improved.

Based on any of the above embodiments, in the method, step 160 specifically includes: if the number of the client emotional tendency recognition results is greater than or equal to the preset number of recognition results in the client emotional tendency set, and the proportion of any emotional type of the client emotional tendency recognition results is greater than the preset type proportion threshold value, taking the emotional type as a calibration emotional tendency recognition result of the text to be recognized, and updating the text to be recognized into a recognized text

Specifically, the preset identification condition, that is, the number of the client emotional tendency identification results in the client emotional tendency set is greater than or equal to the preset identification result number, and the proportion of any emotional type of the client emotional tendency identification results is greater than the preset type proportion threshold, and when the preset identification condition is met, the calibrated emotional tendency identification result can be determined and the text to be identified is marked as the identified text again.

Here, the client emotional tendency recognition result is one of a plurality of emotional types, the number of the client emotional tendency recognition results as the emotional types is counted for any emotional type, and the ratio of the number of the client emotional tendency recognition results as the emotional types to the total number of the client emotional tendency recognition results in the client emotional tendency set is used as the percentage of the emotional types.

The preset number of recognition results is a preset threshold value of the number of client emotional tendency recognition results. The preset type ratio threshold is a preset ratio threshold of the emotion types.

And when the number of the client emotional tendency recognition results is greater than or equal to the number of the preset recognition results in the client emotional tendency set, and the client emotional tendency recognition result is that the ratio of any emotional type is greater than a preset type ratio threshold value, confirming that the client emotional tendency set of the text to be recognized meets the preset recognition condition, taking the emotional type of which the ratio of the emotional type is greater than the preset type ratio threshold value as a calibration emotional tendency recognition result of the text to be recognized, and re-marking the text to be recognized with the calibration emotional tendency recognition result as the recognized text.

For example, the preset number of recognition results is 10, and the preset type fraction threshold is 70%. Aiming at a text A to be recognized, 10 client emotional tendency recognition results are stored in a client emotional tendency set and are equal to the number of preset recognition results; the emotion types of 8 client emotional tendency recognition results are negative, the emotion types of 1 client emotional tendency recognition result are neutral, the emotion types of 1 client emotional tendency recognition result are positive, the proportion of the emotion types is 80%, and the proportion is larger than a preset type proportion threshold value, so that the text A to be recognized is updated to be a recognized text, and negative is used as a calibration emotional tendency recognition result.

Based on any embodiment, the method further includes, after the step 130: counting the recognition error rate of the recognized text; and if the recognition error rate is greater than a preset recognition error threshold value, correcting the marked emotional tendency recognition result of the recognized text.

Specifically, after the client emotional tendency recognition result of the recognized text is obtained, whether the client side correctly recognizes the emotional tendency of the recognized text can be known by comparing the client emotional tendency recognition result of the recognized text with the calibrated emotional tendency recognition result, and then the recognition error rate of the recognized text is counted. The recognition error rate is a ratio of the number of recognition errors to the total number of recognition times, and the recognition error rate may be obtained based on the client emotional tendency recognition results of the recognized text received within a preset time, or may be obtained after receiving a preset number of client emotional tendency recognition results of the recognized text, which is not limited in this embodiment of the present invention. For example, the result of the calibrated emotional tendency recognition of the recognized text B is "neutral"; within 1 month, the recognized text B is recognized by the client 120 times in total, wherein 100 received client emotional tendency recognition results are "positive", and 20 received client emotional tendency recognition results are "neutral", that is, the recognition error rate is 83.33%.

And after the identification error rate is obtained, judging whether the identification error rate is greater than a preset identification error threshold value. Here, the preset recognition error threshold is a preset maximum value of the recognition error rate. If the recognition error rate is greater than the preset recognition error threshold, the calibrated emotional tendency recognition result of the recognized text may be wrong, and at this time, the emotional tendency recognition needs to be performed on the recognized text again to obtain a new calibrated emotional tendency recognition result, so as to correct the calibrated emotional tendency recognition result.

According to any embodiment, the method comprises at least one of positive, neutral and negative calibration emotional tendency recognition results. Correspondingly, the client emotional tendency recognition result, the model emotional tendency recognition result and the sample emotional tendency recognition result all comprise at least one of positive, neutral and negative.

Based on any embodiment, the identity authentication method comprises the following steps:

firstly, capturing initial texts from a text source by a web crawler technology aiming at a preset theme, respectively inputting the initial texts into emotion tendency recognition models obtained by pre-training, and generating and outputting model emotion tendency recognition results and confidence degrees of the initial texts by performing semantic analysis on the initial texts. And classifying the initial text according to the confidence coefficient of the initial text, and judging the initial text to be the recognized text or the text to be recognized. Here, the recognized text corresponds to the calibrated emotional tendency recognition result, i.e., the model emotional tendency recognition result corresponding to the higher confidence level. The emotional tendency recognition result is calibrated to be positive, neutral or negative.

When the client needs to perform authentication, an authentication request may be sent to the authentication end to request the authentication end to perform authentication on the client. When receiving a verification request sent by a client, the verification end needs to return a plurality of texts to the client, wherein the plurality of texts comprise N recognized texts and M texts to be recognized.

After the client receives the plurality of texts, the client operating the client identifies the emotional tendency of each text, inputs the emotional tendency identification result of the client aiming at each text, and returns the emotional tendency identification result of the client of each text to the verifying end so that the verifying end can verify the emotional tendency.

And then, the verifying end receives the client emotional tendency recognition result of each text returned by the client, selects the client emotional tendency recognition result of the recognized text from the client emotional tendency recognition results, and acquires the calibration emotional tendency recognition result of the recognized text stored in advance. And judging whether the marked emotional tendency recognition result of the recognized text is consistent with the client emotional tendency recognition result, so that whether the client correctly recognizes the emotional tendency of the recognized text. After learning whether the client side correctly identifies the emotional tendency of the identified text, the identification accuracy can be obtained. Here, assuming that N is 5, and of the client emotion tendency recognition results corresponding to 5 recognized texts, 4 recognized texts are correct, and one recognized text is incorrect, the recognition accuracy rate is 80%.

And after the identification accuracy is obtained, determining a verification result by comparing the identification accuracy with a preset accuracy threshold, and returning the verification result to the client. If the preset accuracy threshold is 70%, the verification result is that the verification is passed when the identification accuracy is 80%, and the verification result is that the verification is failed when the identification accuracy is 60%.

And when the verification result is that the verification is passed, respectively storing the client emotional tendency recognition results of the M texts to be recognized, which are returned by the client, into client emotional tendency sets corresponding to different texts to be recognized.

And then, judging whether the customer emotional tendency set of the text to be recognized meets a preset recognition condition, namely whether the quantity of the customer emotional tendency recognition results in the customer emotional tendency set is greater than or equal to the quantity of the preset recognition results, and whether the proportion of any emotional type of the customer emotional tendency recognition results is greater than a preset type proportion threshold value. If the preset identification condition is met, the emotion type with the occupation ratio larger than the preset type occupation ratio threshold value is used as a marked emotion tendency identification result of the text to be identified, and the text to be identified is marked as the identified text again. For example, the preset number of recognition results is 10, and the preset type ratio threshold is 70%. Aiming at a text A to be recognized, 10 client emotional tendency recognition results are stored in a client emotional tendency set and are equal to the number of preset recognition results; the emotion types of 8 client emotion tendency recognition results are negative, the emotion types of 1 client emotion tendency recognition result are neutral, the emotion types of 1 client emotion tendency recognition result are positive, the proportion of the emotion types is 80%, and the proportion is larger than a preset type proportion threshold value, so that the text A to be recognized is updated to be a recognized text A, and negative is used as a calibration emotion tendency recognition result. And for the text B to be recognized, 8 client emotional tendency recognition results are stored in the client emotional tendency set, the number of the client emotional tendency recognition results is less than the number of the preset recognition results, and the preset recognition conditions are not met. Aiming at a text C to be recognized, 12 client emotional tendency recognition results are stored in a client emotional tendency set, wherein the emotional type of 6 client emotional tendency recognition results is negative, the emotional type of 3 client emotional tendency recognition results is neutral, the emotional type of 3 client emotional tendency recognition results is positive, the occupation ratio of the emotional types of negative, neutral and positive is respectively 50%, 25% and 25%, the occupation ratio is smaller than a preset type occupation ratio threshold, and the preset recognition condition is not met.

In addition, the recognized text A is obtained by updating the text A to be recognized, the marked emotional tendency recognition result of the recognized text A is obtained based on 10 client emotional tendency recognition results, and the model emotional tendency recognition result output by the emotional tendency recognition model is not used, so that the recognized text A and the emotional tendency recognition result thereof are used as the sample text and the sample emotional tendency recognition result to train the emotional tendency recognition model.

According to the method provided by the embodiment of the invention, the client emotion tendency recognition result of the recognized text returned by the client is received for verification, the limitation is small, the user recognition difficulty is low, and the emotion tendency recognition is used as a verification mode due to the high research and development cost of the emotion tendency recognition, so that the verification threshold of the machine script can be improved, and the verification safety is further improved. Meanwhile, the emotional tendency recognition result of the text to be recognized can be obtained based on the client emotional tendency recognition result of the text to be recognized returned by the client, so that the text emotional tendency recognition is realized, the text emotional tendency recognition cost is reduced, and the accuracy of the text emotional tendency recognition is improved.

Based on any of the above embodiments, fig. 2 is a schematic structural diagram of an identity verification system provided in an embodiment of the present invention, as shown in fig. 2, the identity verification system includes a collection interface 210, a verification interface 220, a request interface 230, a message middleware 240, an analysis processing module 250, a recognized library 260, an unidentified library 270, and an intermediate library 280.

The collection interface 210 is configured to collect an initial text and transmit the initial text to the analysis processing module 250 through the message middleware 240. The analysis processing module 250 includes an emotional tendency recognition model obtained through pre-training, the analysis processing module 250 inputs the initial text into the emotional tendency recognition model, obtains a model emotional tendency recognition result and a confidence coefficient output by the emotional tendency recognition model, and further divides the initial text into a recognized text and a text to be recognized by comparing the confidence coefficient with a preset confidence threshold value, and stores the recognized text and the text to be recognized in the recognized library 260 and the unrecognized library 270 respectively.

The request interface 230 is configured to receive an authentication request sent by a client, and transmit the authentication request to the analysis processing module 250 through the message middleware 240. After receiving the verification request, the analysis processing module 250 extracts a number of recognized texts and a number of texts to be recognized from the recognized library 260 and the unrecognized library 270, respectively, and returns the number of recognized texts and the number of texts to be recognized to the client via the message middleware 240 and the request interface 230.

After receiving the plurality of recognized texts and the plurality of texts to be recognized, the client performs emotion tendency recognition on the recognized texts and the plurality of texts to be recognized respectively, and returns the emotion tendency recognition result of the client to the identity verification system through the verification interface 220.

The verification interface 220 is configured to receive the client emotional tendency recognition result of the recognized text and the client emotional tendency recognition result of the text to be recognized returned by the client, and transmit the calibrated client emotional tendency recognition result of the recognized text and the client emotional tendency recognition result of the text to be recognized to the analysis processing module 250 through the message middleware 240. The analysis processing module 250 determines a verification result according to the client emotional tendency recognition result and the calibration emotional tendency recognition result of the recognized text, and returns the verification result to the client through the message middleware 240 and the verification interface 220.

In addition, when the verification result is that the verification is passed, the analysis processing module 250 also stores the received client emotional tendency recognition result of the text to be recognized in the client emotional tendency set corresponding to the text to be recognized in the intermediate repository 280. If the analysis processing module 250 judges that the client emotional tendency set meets the preset identification condition, the calibration emotional tendency identification result of the corresponding text to be identified is determined based on the client emotional tendency identification result in the client emotional tendency set, and the text to be identified is updated to be the identified text and is transferred to the identified library 260.

Based on any of the above embodiments, fig. 3 is a schematic flow chart of an authentication method according to another embodiment of the present invention, as shown in fig. 3, an execution subject of the method is a client, and the method includes:

step 310, sending an authentication request to the authentication end.

When the client needs to perform identity authentication, an authentication request is sent to the authentication end to request the authentication end to perform identity authentication on a client operating the client.

After receiving a verification request sent by the client, the verification end can select a recognized text from a large number of pre-stored recognized texts and return the recognized text to the client. Here, the recognized text selected may be one or more.

And step 320, receiving the recognized text which is sent by the verifying end in response to the verification request and completes the emotional tendency recognition.

And step 330, acquiring a client emotional tendency recognition result corresponding to the input recognized text.

Specifically, after the recognized text is received, a client operating the client performs emotional tendency recognition on the recognized text, and inputs a result of performing emotional tendency recognition on the recognized text, that is, a client emotional tendency recognition result of the recognized text, to the client, that is, a client emotional tendency recognition result corresponding to the recognized text is obtained. Here, the client emotional tendency recognition result may be one of a plurality of emotion types set in advance.

And 340, returning the client emotional tendency identification result to the verification end to trigger the verification end to determine the verification result based on the calibration emotional tendency identification result of the identified text and the client emotional tendency identification result.

Specifically, after the client emotional tendency recognition result is obtained, the client emotional tendency recognition result of the recognized text is returned to the verification end, so that the verification end can verify the client emotional tendency recognition result.

Then, the verifying terminal receives the returned client emotional tendency recognition result of the recognized text. After the verifying end obtains the client emotional tendency recognition result, whether the client side correctly recognizes the emotional tendency of the recognized text can be known based on the marked emotional tendency recognition result of the recognized text and the client emotional tendency recognition result. And obtaining a verification result after knowing whether the client side correctly identifies the emotional tendency of the identified text.

According to the method provided by the embodiment of the invention, when identity authentication is required, the identified text sent by the authentication end is received, and the returned client emotional tendency identification result corresponding to the identified text is used for the authentication end to perform identity authentication, so that the limitation is small, the user identification difficulty is low, and because the research and development cost of emotional tendency identification is high, the emotional tendency identification is used as an authentication mode, the authentication threshold of a machine script can be improved, and the security of the identity authentication is further improved.

Based on any of the above embodiments, the method further comprises: receiving a to-be-identified text which is sent by a verification end in response to a verification request and does not finish emotion tendency identification, and acquiring a client emotion tendency identification result corresponding to the input to-be-identified text; returning a client emotion tendency identification result corresponding to the text to be identified to the verification end to trigger the verification end to execute the following operations: if the verification result is that the verification is passed, storing the client emotional tendency recognition result of the text to be recognized into the client emotional tendency set of the text to be recognized; and if the client emotional tendency set meets the preset identification condition, determining a calibration emotional tendency identification result of the text to be identified based on the client emotional tendency set, and updating the text to be identified into the identified text.

Specifically, the verification end sends the recognized text to the client and also sends the text to be recognized to the client. After receiving the text to be recognized, a client operating the client performs emotional tendency recognition on the text to be recognized, and inputs a result of the emotional tendency recognition on the text to be recognized, namely a client emotional tendency recognition result of the text to be recognized, to the client, namely a client emotional tendency recognition result corresponding to the text to be recognized is obtained.

And after the client emotional tendency recognition results of the recognized text and the text to be recognized are obtained, returning the client emotional tendency recognition result of the recognized text and the client emotional tendency recognition result of the text to be recognized to the verification end. And after the verification end determines the verification result based on the client emotional tendency recognition result of the recognized text and the calibration emotional tendency recognition result, if the verification result is that the verification is passed, the client emotional tendency recognition results of the text to be recognized returned by the client are respectively stored into the corresponding client emotional tendency sets of the text to be recognized. If the client emotional tendency set meets the preset identification condition, the verification end determines a calibration emotional tendency identification result of the corresponding text to be identified based on the client emotional tendency set, obtains the calibration emotional tendency identification result to represent that the text to be identified is already identified, and re-marks the text to be identified as the identified text. There are various methods for determining the calibrated emotional tendency recognition result of the text to be recognized, for example, directly taking the emotional type with the highest number of the client emotional tendency recognition results in the client emotional tendency set as any emotional type as the calibrated emotional tendency recognition result.

Based on any of the above embodiments, fig. 4 is a schematic structural diagram of an authentication end provided in an embodiment of the present invention, as shown in fig. 4, the authentication end includes a text obtaining unit 410, a text sending unit 420, a trend receiving unit 430, and an identity authentication unit 440;

the text acquiring unit 410 is configured to acquire an identified text that has completed emotion tendency identification if a verification request sent by a client is received;

the text sending unit 420 is configured to send the recognized text to the client;

the tendency receiving unit 430 is configured to receive a client emotional tendency recognition result for the recognized text returned by the client;

the identity verification unit 440 is configured to determine a verification result based on the calibrated emotional tendency recognition result of the recognized text and the client emotional tendency recognition result.

The verification end provided by the embodiment of the invention verifies the client emotion tendency recognition result of the recognized text returned by the receiving client, so that the limitation is small, the user recognition difficulty is low, and the emotion tendency recognition is used as a verification mode to improve the verification threshold of the machine script and further improve the safety of identity verification due to high research and development cost of the emotion tendency recognition.

Based on any of the above embodiments, in the verification end, the text obtaining unit 410 is further configured to obtain a text to be recognized that does not complete emotional tendency recognition if a verification request sent by the client is received;

the text sending unit 420 is further configured to send the text to be recognized to the client;

the tendency receiving unit 430 is further configured to receive a client emotion tendency recognition result of the text to be recognized, which is returned by the client;

the verification end further comprises:

the set storage unit is used for storing the client emotional tendency recognition result of the text to be recognized into the client emotional tendency set of the text to be recognized if the verification result is that the verification is passed;

and the client identification unit is used for determining a calibration emotional tendency identification result of the text to be identified based on the client emotional tendency set if the client emotional tendency set meets a preset identification condition, and updating the text to be identified into an identified text.

Based on any of the above embodiments, in the verification end, the identity verification unit 440 is specifically configured to:

Based on any embodiment, the verification end further comprises an initial marking unit; the initial marking unit is used for:

acquiring an initial text;

otherwise, the initial text is marked as the text to be recognized.

Based on any of the above embodiments, the verification end further comprises a model optimization unit; the model optimization unit is configured to:

and if the client emotional tendency set meets the preset identification condition, training the emotional tendency identification model based on the text to be identified and the calibration emotional tendency identification result of the text to be identified.

Based on any of the above embodiments, in the verification end, the client identification unit is specifically configured to:

and if the number of the customer emotional tendency recognition results in the customer emotional tendency set is greater than or equal to the number of preset recognition results, and the proportion of any emotional type of the customer emotional tendency recognition results is greater than a preset type proportion threshold value, taking any emotional type as a calibration emotional tendency recognition result of the text to be recognized, and updating the text to be recognized into a recognized text.

Based on any embodiment, the verification end further comprises a calibration correction unit; the calibration correction unit is used for:

counting the recognition error rate of the recognized text;

Based on any embodiment, in the verifying end, the calibrated emotional tendency recognition result includes at least one of the following: positive, neutral, negative.

Based on any of the above embodiments, fig. 5 is a schematic structural diagram of a client according to an embodiment of the present invention, and as shown in fig. 5, the client includes a request sending unit 510, a text receiving unit 520, a tendency obtaining unit 530, and a tendency sending unit 540;

the request sending unit 510 is configured to send an authentication request to an authentication end;

the text receiving unit 520 is used for receiving the recognized text which is sent by the verifying terminal in response to the verification request and completes emotional tendency recognition;

the tendency acquiring unit 530 is used for acquiring a client emotion tendency recognition result corresponding to the input recognized text;

the tendency sending unit 540 is configured to return the client emotional tendency recognition result to the verification end, so as to trigger the verification end to determine a verification result based on the calibration emotional tendency recognition result of the recognized text and the client emotional tendency recognition result.

When identity authentication is required, the client side provided by the embodiment of the invention receives the recognized text sent by the authentication end, and returns the client emotion tendency recognition result corresponding to the recognized text for the authentication end to perform identity authentication, so that the limitation is small, the user recognition difficulty is low, and because the research and development cost of emotion tendency recognition is high, the emotion tendency recognition is used as an authentication mode, the authentication threshold of a machine script can be improved, and the security of identity authentication is further improved.

Based on any embodiment, in the client, the text receiving unit 520 is further configured to receive the text to be recognized, which is sent by the verifying end in response to the verification request and does not complete emotion tendency recognition;

the tendency acquiring unit 530 is further configured to acquire a client emotion tendency recognition result corresponding to the input text to be recognized;

the tendency sending unit 540 is further configured to return the client emotion tendency identification result corresponding to the text to be identified to the verifying end, so as to trigger the verifying end to perform the following operations:

Fig. 6 is a schematic entity structure diagram of an electronic device according to an embodiment of the present invention, and as shown in fig. 6, the electronic device may include: a processor (processor)601, a communication Interface (Communications Interface)602, a memory (memory)603 and a communication bus 604, wherein the processor 601, the communication Interface 602 and the memory 603 complete communication with each other through the communication bus 604. The processor 601 may call a computer program stored on the memory 603 and executable on the processor 601 to perform the authentication method provided by the above embodiments, for example, including: if a verification request sent by a client is received, acquiring an identified text which completes emotional tendency identification; sending the recognized text to the client; receiving a client emotional tendency recognition result for the recognized text returned by the client; and determining a verification result based on the marked emotional tendency recognition result of the recognized text and the client emotional tendency recognition result.

The processor 601 may also call a computer program stored on the memory 603 and executable on the processor 601 to perform the authentication method provided by the above embodiments, for example, including: sending a verification request to a verification end; receiving the recognized text which is sent by the verifying end in response to the verification request and completes emotional tendency recognition; acquiring a client emotional tendency recognition result corresponding to the input recognized text; and returning the client emotional tendency recognition result to the verification end to trigger the verification end to determine the verification result based on the calibration emotional tendency recognition result of the recognized text and the client emotional tendency recognition result.

In addition, the logic instructions in the memory 603 may be implemented in the form of software functional units and stored in a computer readable storage medium when the logic instructions are sold or used as independent products. Based on such understanding, the technical solutions of the embodiments of the present invention may be essentially implemented or make a contribution to the prior art, or may be implemented in the form of a software product stored in a storage medium and including instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the methods described in the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.

Embodiments of the present invention further provide a non-transitory computer-readable storage medium, on which a computer program is stored, where the computer program is implemented to perform the identity verification method provided in the foregoing embodiments when executed by a processor, for example, the method includes: if a verification request sent by a client is received, acquiring an identified text which completes emotional tendency identification; sending the recognized text to the client; receiving a client emotional tendency recognition result for the recognized text returned by the client; and determining a verification result based on the marked emotional tendency recognition result of the recognized text and the client emotional tendency recognition result.

Embodiments of the present invention further provide a non-transitory computer-readable storage medium, on which a computer program is stored, where the computer program is implemented to perform the identity verification method provided in the foregoing embodiments when executed by a processor, for example, the method includes: sending a verification request to a verification end; receiving the recognized text which is sent by the verifying end in response to the verification request and completes emotional tendency recognition; acquiring a client emotional tendency recognition result corresponding to the input recognized text; and returning the client emotional tendency recognition result to the verification end to trigger the verification end to determine the verification result based on the calibration emotional tendency recognition result of the recognized text and the client emotional tendency recognition result.

The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.

Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.

Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.

Claims

1. An identity verification method, comprising:

sending the recognized text to the client;

receiving a client emotional tendency recognition result for the recognized text returned by the client; the client emotional tendency recognition result is an emotional tendency recognition result which is input to a client side by a client operating the client side to recognize the emotional tendency of the recognized text;

2. The method of identity verification according to claim 1, the method further comprising:

sending the text to be recognized to the client;

3. The identity verification method according to claim 1, wherein the determining a verification result based on the calibration emotional tendency recognition result of the recognized text and the client emotional tendency recognition result specifically comprises:

4. The identity authentication method according to claim 2, wherein if receiving an authentication request sent by a client, acquiring the recognized text of which emotional tendency recognition is completed, the method further comprises:

acquiring an initial text;

otherwise, the initial text is marked as the text to be recognized.

5. The identity verification method according to any one of claims 1 to 4, wherein the receiving of the client emotion tendency recognition result of the recognized text returned by the client further comprises:

counting the recognition error rate of the recognized text;

6. The identity verification method of any one of claims 1 to 4, wherein the nominal emotional tendency recognition result comprises at least one of:

positive, neutral, negative.

7. An identity verification method, comprising:

sending a verification request to a verification end;

acquiring a client emotional tendency recognition result corresponding to the input recognized text; the client emotional tendency recognition result is an emotional tendency recognition result which is input to a client side by a client operating the client side to recognize the emotional tendency of the recognized text;

8. The identity authentication method according to claim 7, wherein the sending of the authentication request to the authentication end further comprises:

9. An authentication peer, comprising:

a text sending unit, configured to send the recognized text to the client;

the tendency receiving unit is used for receiving a client emotion tendency recognition result for the recognized text returned by the client; the client emotional tendency recognition result is an emotional tendency recognition result which is input to a client side by a client operating the client side to recognize the emotional tendency of the recognized text;

10. A client, comprising:

the tendency acquisition unit is used for acquiring a client emotion tendency identification result corresponding to the input identified text; the client emotional tendency recognition result is an emotional tendency recognition result which is input to a client side by a client operating the client side to recognize the emotional tendency of the recognized text;