CN109015690B

CN109015690B - Active interactive dialogue robot system and method

Info

Publication number: CN109015690B
Application number: CN201811011300.4A
Authority: CN
Inventors: 韩冰
Original assignee: Icsoc Beijing Communication Technology Co ltd
Current assignee: Icsoc Beijing Communication Technology Co ltd
Priority date: 2018-08-31
Filing date: 2018-08-31
Publication date: 2022-01-28
Anticipated expiration: 2038-08-31
Also published as: CN109015690A

Abstract

The invention provides an active interactive dialogue robot system and a method, comprising a pre-sale consultation module and a middle-sale handling module, wherein the pre-sale consultation module comprises a call management module, a response module, a satisfaction survey module and a user module; the system comprises the modules, so that the user can be verified before the user transacts the business, whether the user transacts for the user is identified, the safety is improved, and unnecessary troubles are avoided for the user.

Description

Active interactive dialogue robot system and method

Technical Field

The invention relates to the technical field of human-computer interaction, in particular to an active interactive dialogue robot system and method.

Background

The existing customer service robot can communicate with a user in a mode of asking questions by the user, and is characterized in that a triggering condition depends on the active question of the user, the robot passively answers based on a knowledge base and a support service type and cannot provide active service for the user, Chinese patent CN107053208A provides an active interactive dialogue robot system and an active dialogue method of the system, the system comprises a dialogistic management module, a question generation module, an answer acquisition module, an answer verification module and a command execution module, the robot system has active dialogue capability, the dialogies in a scene are extracted and configured in the dialogue robot system in the actual dialogue process, and the dialogue robot acquires and verifies answers based on the dialogistic configuration, so that the aim of saving a large amount of labor is fulfilled; although the robot has active conversation capability, the robot transacts business for users in a question-answering mode, cannot distinguish whether users transact business personally, is lack of safety, has the risk that other people falsely transact business by using user identity information, and is easy to bring unnecessary trouble to users.

Disclosure of Invention

In order to solve the technical problem, the invention provides an active interactive dialogue robot system and method.

The specific technical scheme of the invention is as follows:

an active interactive dialogue robot system comprises a pre-sale consultation module and a mid-sale transaction module;

the pre-sale consultation module comprises a conversation management module, an answer module, a satisfaction survey module and a user module;

a dialogistic management module: the system is used for creating an industrial speech and skill library and customizing a dialogue automatic question and answer corpus which accords with the business;

an answer module: the system comprises a user module, a database server and a database server, wherein the user module is used for acquiring questions sent by the user module, matching corresponding answers in an industrial and tactical library and sending the answers to the user module;

satisfaction survey module: the system is used for investigating the satisfaction degree of the user on the consultation when or after the consultation is finished;

a user module: the answer module is used for providing questions for users, answering the questions and receiving answers sent by the answer module;

the in-sale transaction module comprises a verification module, a dialogue module and a command execution module;

a verification module: the system is used for verifying whether the user transacts for oneself or not;

a dialogue module: the system comprises a verification module, a command execution module and a data processing module, wherein the verification module is used for receiving a command sent by the verification module, proposing a question to a user according to an application scene, collecting an answer of the user, verifying the legality of the answer of the user and sending the command to the command execution module according to a verification result;

a command execution module: for executing the corresponding service command or exiting;

the verification module comprises a voice verification module;

a voice verification module: the voice recognition module is used for sending voice verification information to the user, judging whether the user transacts the user or not, if the user transacts the user, sending an instruction to the dialogue module, and if the user transacts the user not, sending an instruction to the command execution module, wherein the instruction content is quitting.

The corresponding business command for executing can be business such as ordering and purchasing goods for the customer or purchasing train tickets for the customer; voice authentication is a prior art, and a user can make a voice call or other modes capable of achieving the purpose of authentication, but the voice authentication is not limited to the mode; the validity of the answer of the user is checked by conventional techniques, including all methods capable of achieving the purpose, such as an interactive dialogue robot system and an active dialogue method of the system disclosed in chinese patent CN107053208A, where the method describes dimensions of checking and judging in checking the validity of the answer of the user by an answer checking module, including: whether the format of the answer file is text, picture and sound, whether the answer content comprises specific text and specific enumerated values, and the method for verifying the legality of the answer by the user in the invention comprises the following steps, but is not limited to the form; the voice verification module comprises all devices capable of realizing voice verification, such as a mobile phone payment system disclosed in chinese patent CN104134140A, the voice verification module in the system comprises a microphone, a speaker and a voice recognition program, and the voice verification module of the present invention comprises the above devices but is not limited to this form; the speech management module in the invention includes all methods capable of implementing the function, for example, chinese patent CN107053208A provides an active interactive dialogue robot system and an active dialogue method of the system, the speech management module in the dialogue robot system is mainly used for managing all problems in the active dialogue process, and manages the speech state in the active dialogue process by itself, and the speech management module in the invention can be implemented by the same method, but is not limited to this method.

Further, the dialogue robot system further includes a fault elimination module, and the fault elimination module includes:

a fault monitoring module: the working state of the conversation robot system is collected constantly, whether the working state of the conversation robot is abnormal or not is judged, if the working state of the conversation robot is abnormal, an instruction is sent to a fault processing module, and if the working state of the conversation robot is normal, no processing is carried out;

a fault processing module: for calling the relevant staff, who will handle the fault at the first time.

Furthermore, the fault monitoring module comprises an answer speed judging module, a satisfaction survey judging module, a voice verification judging module and an execution judging module;

an answer speed judgment module: the time T1 for judging the question sending time of the user module and the time T2 for answering the question by the dialogue robot system, the sizes of a time difference threshold Tn and a time difference T2-T1 are judged, if the time T2-T1 is more than or equal to the time Tn, an instruction is sent to the fault processing module, and if the time T2-T1 is less than the time Tn, an instruction is sent to the satisfaction survey judging module;

satisfaction survey judge module: the system is used for judging whether to send a satisfaction survey short message or a voice telephone to the user after the consultation of the user is finished, if so, sending an instruction to the voice verification judging module, and if not, sending an instruction to the fault processing module;

the voice verification judgment module: the voice verification module is used for judging whether to execute voice verification to a user before executing related services, if so, sending an instruction to the execution judgment module, and if not, sending an instruction to the fault processing module;

an execution judgment module: the interactive robot is used for judging whether the interactive robot transacts related services for the user after the user selects to purchase, if so, the interactive robot does not process, and if not, the interactive robot sends an instruction to the fault processing module.

Further, the verification module further comprises an image verification module, the image verification module comprising:

a first facial feature acquisition module: the face feature comparison module is used for acquiring and identifying a plurality of face features of a user and sending the identified face features to the first face feature comparison module;

an identity information acquisition module: the identity card reader is used for scanning an identity card of a user, identifying a card photo on the identity card and sending the card photo to the first surface characteristic comparison module;

a first facial feature comparison module: the system comprises a first face feature acquisition module, a second face feature acquisition module, a third face feature acquisition module, a fourth face feature acquisition module, a fifth face feature acquisition module, a sixth face feature acquisition module, a fifth face feature acquisition module and a sixth face feature acquisition module.

The first facial feature acquisition module comprises all devices capable of acquiring facial features of a user, such as a facial feature control image acquisition access control system disclosed in chinese patent CN107749110A, wherein the image acquisition device in the patent comprises a camera and an image processing unit, the camera is connected with the image processing unit and is used for acquiring images in a monitored area, and the facial recognition module of the present invention comprises the image acquisition device in the patent, but is not limited to this form; comparing the facial features of the user with the certificate photo is a conventional technology, including all the ways that this effect can be achieved, for example, chinese patent CN107742106A discloses a facial matching method and device based on an unmanned vehicle, the facial matching method is: the method comprises the steps of obtaining a face image of a passenger taking an unmanned vehicle, determining the similarity between the face image of the passenger and face images of known loved persons based on a pre-trained face recognition model, determining whether the passenger is the loved person according to the similarity, and as disclosed in Chinese patent CN103927747A, the method comprises the following steps: the method of combining structured light and binocular stereo vision in a patient space obtains three-dimensional point cloud of the surface of the patient space, identifies nose tips and eyes on a two-dimensional image by an ASM method based on statistical information, carries out surface reconstruction on a medical image acquired before an operation to obtain a human head three-dimensional model, extracts facial features on the three-dimensional model by carrying out principal component analysis on a local field, further clusters and identifies corresponding nose tips and eyes on the features, obtains corresponding points of the nose tips and the eyes, calculates an initial change matrix between the human head model and the three-dimensional point cloud, and then completes point cloud registration of a model space of the human head three-dimensional model and the patient space by utilizing a closest point algorithm (ICP). But is not limited to this form;

further, the verification module still includes double verification module, and double verification module includes:

a double face information identifier module: the system comprises a double-person facial feature comparison module, a double-person facial feature comparison module and a user identification module, wherein the double-person facial feature comparison module is used for simultaneously acquiring two groups of facial features of two users, each group of facial features comprises a plurality of facial features of one user, and the two groups of facial features are sent to the double-person facial feature comparison module;

double identity information acquisition module: the system comprises a double face feature comparison module, a double face feature comparison module and a double face feature comparison module, wherein the double face feature comparison module is used for scanning identity documents of two users and acquiring document photos on the two identity documents;

double facial feature comparison module: the certificate photo recognition module is used for receiving the two groups of facial features sent by the double-person facial information recognition sub-module and the certificate photo sent by the double-person identity information acquisition module, respectively comparing the two groups of facial features with the corresponding certificate photos, calculating the similarity Tm, comparing the two maximum similarities Tm1 and Tm2 with a similarity threshold value T3, when Tm1 is larger than or equal to T3 and Tm2 is larger than or equal to T3, successfully verifying, sending an instruction to the dialogue module, and when Tm1 is smaller than T3 or Tm2 is smaller than T3, sending an instruction to the instruction execution module, wherein the instruction content is exit.

Further, the dialog module includes:

a problem storage module: the system comprises a question generation module, a question storage module, a question generation module and a question storage module, wherein the question generation module is used for storing all questions to be asked in a certain scene, arranging all the questions to be asked in sequence, and sending all the questions arranged in sequence to the question generation module when receiving an instruction sent by the question generation module;

a problem generation module: the system comprises a question storage module, a user module and an answer acquisition module, wherein the question storage module is used for storing questions to be answered and questions to be answered;

an answer acquisition module: the answer verification module is used for acquiring answer content of the user module to the question and sending the answer content to the answer verification module;

an answer checking module: the answer management module is used for receiving the answer content sent by the answer acquisition module, checking the legality of the answer content and sending an instruction to the command execution module or the repeated question-asking module according to a checking result;

a repeated question asking module: the system is used for clearing answers input by a user and repeatedly sending questions with illegal verification to the user module;

the answer verification module comprises:

the digital answer judging module: the system comprises a character type answer judging module, a character type answer judging module and a display module, wherein the character type answer judging module is used for converting the answer content of a user into a text and judging whether the answer of the user is a number or not;

the digital length judging module: the answer processing module is used for receiving the number Xn of the numbers sent by the answer judging module of the number classes, judging the size of the Xn and the threshold value X1 of the number of the numbers, if the Xn is X1, sending an instruction to the question generating module, sending the next question to the user module, and if the Xn is not X1, sending an instruction to the repeated questioning module;

the character answer judging module: the system is used for judging whether the answer of the user is a character answer or not, if so, identifying the character content in the text and sending the character content to the same character judgment module, and if not, sending an instruction to the scanning judgment module;

the same character judging module: the system comprises a question generation module, a question answer judgment module, a question repeating module and a question repeating module, wherein the question generation module is used for receiving text contents sent by the text answer judgment module, judging whether words which are the same as preset answers exist in the text contents, if yes, sending an instruction to the question generation module, sending the next question to the user module, and if not, sending an instruction to the question repeating module;

a scanning judgment module: the certificate question issuing module is used for scanning certificates provided by users and judging whether the scanned certificates are qualified or not, if so, sending an instruction to the question generating module to issue the next question, and if not, sending an instruction to the repeated question module.

Further, the scanning judgment module comprises:

a content matching module: the certificate matching module is used for judging whether the text content on the certificate provided by the user is consistent with the question or not, if so, sending an instruction to the certificate matching module, and if not, sending an instruction to the repeated question module;

the certificate matching module: the system is used for receiving the instruction sent by the content matching module, identifying and judging whether the content stamped on the certificate conforms to the text content, if so, sending the instruction to the question generation module to ask the next question, if not, judging that the certificate is unqualified, and sending the instruction to the repeated question asking module.

Further, the question generation module includes:

the number judging module: the answer acquisition module is used for judging the number P of the users who choose to transact the certificates after the questions are sent to the user module, if P is 1, sending an instruction to the answer acquisition module, and if P is more than 1, sending an instruction to the answer mode module;

an answer mode module: the system is used for judging whether answers of a plurality of users for answering a certain question are the same or not, if yes, sending an instruction to a primary answer acquisition module, and if not, sending an instruction to a multi-answer acquisition module;

the answer acquisition module comprises:

a primary answer acquisition module: the answer verification module is used for acquiring answer content of the user to the question at one time, sending the answer content to the answer verification module, and sending the next question to the user module after the answer verification is successful;

a multiple answer acquisition module: the system is used for collecting answer contents of the user to the questions for multiple times, the collecting times are the same as the number of the user, all the answer contents are sent to the answer verification module according to the collecting sequence, and the next question is sent to the user module after all the answer contents are verified successfully.

Further, the dialogue robot system also comprises a rechecking module, and the rechecking module comprises:

a list generation module: the system is used for generating a list of all questions and corresponding answers answered by the user and sending the list to the user module after the user answers all questions and verifies successfully;

and (4) rechecking the submodule: the system comprises a secondary identity verification module, a problem selection module and a user module, wherein the secondary identity verification module is used for receiving an instruction sent by the user module and identifying the content of the instruction, if the content of the instruction does not need to be modified, the instruction is sent to the secondary identity verification module, and if the content of the instruction needs to be modified, the instruction is sent to the problem selection module;

a question selection module: the system comprises a rechecking submodule, a user module and a control module, wherein the rechecking submodule is used for receiving an instruction sent by the rechecking submodule, calling a problem needing to be modified and sending the problem to the user module;

the secondary identity authentication module: the system comprises a command execution module, a user interface module and a command execution module, wherein the command execution module is used for executing identity authentication again after a user sends an instruction which does not need to be modified, and sending the instruction to the command execution module according to an authentication result;

the secondary identity verification module comprises:

a second facial feature acquisition module: the facial feature comparison module is used for acquiring facial features of the user again after the user sends an instruction which does not need to be modified, and sending the facial features to the second facial feature comparison module;

a second facial feature comparison module: and the face characteristic acquisition module is used for comparing the face characteristic acquired by the second face characteristic acquisition module with the face characteristic acquired by the first face characteristic acquisition module, judging whether the face characteristics are matched, if so, sending an instruction to the command execution module, wherein the instruction content is to execute a corresponding service, and if not, sending an instruction to the command execution module, and the instruction content is to exit.

Further, the satisfaction survey module comprises:

a mouth expression recognition module: the method is used for collecting facial features of a plurality of clients during communication with the clients and identifying mouth expressions of the clients, and specifically comprises the following steps: the number of puckered mouths F1, the number of smiles F2, the number of beep mouths F3, the number of outmoded smiles F4;

eye expression recognition module: the method is used for collecting facial features of a plurality of clients during communication with the clients and identifying eye expressions of the clients, and specifically comprises the following steps: the number of times of glaring G1 and the number of times of white eyes G2;

eyebrow expression recognition module: the method is used for collecting facial features of a plurality of clients during communication with the clients and identifying eye expressions of the clients, and specifically comprises the following steps: the number of eyebrow plucking times H1 and the number of eyebrow crumpling times H2;

a head action recognition module: the method is used for collecting facial features of a plurality of clients during communication with the clients and identifying eye expressions of the clients, and specifically comprises the following steps: the number of shaking heads J1, the number of nodding heads J2;

a score calculation module: the method is used for collecting expressions identified by the mouth expression identification module, the eye expression identification module, the eyebrow expression identification module and the head action identification module and calculating the satisfaction degree Q, wherein the Q is calculated by the following method; Q-F1 × F1+ F21 × F2-F3 × F3+ F4 × F4-G1 × G1-G2 × G2-H1 × H1-H2 × H2-J1 × J1+ J2 × J2, wherein F1+ F2+ F3+ F4+ G1+ G2+ H1+ H2+ J1+ J2 equals 100%.

According to the active interactive dialogue robot system, the face recognition module, the verification module, the dialogue module, the command execution module and the user module are arranged, whether a user transacts for the user can be recognized before the user transacts the certificate, the safety is improved, the risk that other people maliciously falsely use the user identity information to transact business is reduced, and unnecessary troubles are avoided for the user.

Drawings

FIG. 1 is a block diagram showing the structure of an active interactive dialogue robot system according to embodiment 1;

FIG. 2 is a block diagram showing a structure of a failure elimination module according to embodiment 2;

FIG. 3 is a block diagram showing a structure of a failure monitoring module according to embodiment 3;

FIG. 4 is a block diagram showing the structure of an image verification module according to embodiment 4;

FIG. 5 is a block diagram showing the structure of a double-person authentication module according to embodiment 5;

FIG. 6 is a block diagram showing the structure of a dialogue module according to embodiment 6;

FIG. 7 is a block diagram showing a structure of a scanning judgment module according to embodiment 7;

FIG. 8 is a block diagram showing the structure of a problem generation module according to embodiment 8;

FIG. 9 is a block diagram showing the structure of a review module according to embodiment 9;

FIG. 10 is a block diagram showing the structure of a secondary authentication module according to embodiment 9;

FIG. 11 is a block diagram showing the structure of a satisfaction survey module according to example 10.

Detailed Description

The invention is further described with reference to the following figures and examples, which are provided for the purpose of illustrating the general inventive concept and are not intended to limit the scope of the invention.

Example 1

The invention provides an active interactive dialogue robot system, which comprises a pre-sale consultation module 1 and a mid-sale transaction module 2, wherein the pre-sale consultation module is connected with the pre-sale consultation module;

the pre-sale consultation module 1 comprises a conversation storage module 11, an answer module 12, a satisfaction survey module 13 and a user module 14;

the dialect storage module 11: the system is used for creating an industrial speech and skill library and customizing a dialogue automatic question and answer corpus which accords with the business;

the answer module 12: the answer matching system is used for collecting the questions sent by the user module 14, matching corresponding answers in an industry and technology library and sending the answers to the user module 14;

satisfaction survey module 13: the system is used for investigating the satisfaction degree of the user on the consultation when or after the consultation is finished;

the user module 14: for the user to ask a question, answer the question and receive the answer sent by the answer module 12;

the in-sale transaction module 2 comprises a verification module 21, a dialogue module 22 and a command execution module 23;

the verification module 21: the system is used for verifying whether the user transacts for oneself or not;

the dialogue module 22: the instruction receiving module is used for receiving the instruction sent by the verification module 21, proposing a question to a user according to an application scene, collecting an answer of the user, verifying the legality of the answer of the user, and sending the instruction to the instruction execution module 23 according to a verification result;

the command execution module 23: for executing the corresponding service command or exiting;

the authentication module 21 includes a voice authentication module 211;

the voice verification module 211: the voice authentication module is used for sending voice authentication information to a user, judging whether the user transacts the user, if the user transacts the user, sending an instruction to the dialogue module 22, and if the user transacts the user, sending an instruction to the command execution module 23, wherein the instruction content is quitting.

The active interaction dialogue robot system provided by the embodiment can provide consultation and purchase service for a user mainly through the above setting, the identity of the user is verified before the user purchases commodities, whether the user is transacted for the user is identified, if the verification fails, the transaction is quitted, if the verification succeeds, the problem is sent to the user, the risk that other people maliciously falsely use the user account to purchase the commodities is reduced, and unnecessary trouble is avoided being brought to the user.

Example 2

The active interactive dialogue robot system of the present embodiment is different from embodiment 1 in that, as shown in fig. 2, the dialogue robot system further includes a troubleshooting module 3, where the troubleshooting module 3 includes:

the fault monitoring module 31: constantly collecting the working state of the conversation robot system, judging whether the working state of the conversation robot is abnormal or not, if so, sending an instruction to the fault processing module 32, and if not, not processing;

the failure processing module 32: for calling the relevant staff, who will handle the fault at the first time.

The active interactive dialogue robot system provided by the embodiment can monitor the working state of the dialogue robot at any time by arranging the fault elimination module, and once a fault occurs, related workers can be called to process the fault at the first time, so that the user experience of a user is indirectly improved.

Example 3

The fault monitoring module is further limited in this embodiment, as shown in fig. 3, the fault monitoring module 31 includes an answer speed determining module 311, a satisfaction survey determining module 312, a voice verification determining module 313, and an execution determining module 314;

answer speed determination module 311: a time T1 for judging the time when the user module 14 sends the question and a time T2 for answering the question by the dialogue robot system, the sizes of a time difference threshold Tn and a time difference T2-T1 are judged, if the time T2-T1 is more than or equal to Tn, an instruction is sent to the fault processing module 32, and if the time T2-T1 is less than the time Tn, an instruction is sent to the satisfaction survey judging module 312;

satisfaction survey and determination module 312: the voice recognition module is used for judging whether to send a satisfaction survey short message or a voice call to the user after the consultation of the user is finished, if so, sending an instruction to the voice verification judging module 313, and if not, sending an instruction to the fault processing module 32;

the voice verification judging module 313: the voice authentication module is used for judging whether to execute voice authentication to a user before executing related services, if so, sending an instruction to the execution judging module 314, and if not, sending an instruction to the fault processing module 32;

the execution determination module 314: the interactive robot is used for judging whether the interactive robot transacts related services for the user after the user selects to purchase, if so, the interactive robot does not process the services, and if not, the interactive robot sends an instruction to the fault processing module 32.

According to the active interactive dialogue robot system provided by the embodiment, whether the robot system fails or not is judged in a multi-dimensional mode through specifically limiting the fault monitoring module, so that the accuracy of fault detection is improved.

Example 4

The present embodiment further defines the verification module, as shown in fig. 4, the verification module 21 further includes an image verification module 212, and the image verification module 212 includes:

first facial feature acquisition module 2121: the facial recognition module is used for acquiring and recognizing a plurality of facial features of the user and sending the recognized facial features to the first facial feature comparison module 2123;

the identity information obtaining module 2122: the system is used for scanning the identity document of the user, identifying a document photo on the identity document and sending the document photo to the first surface characteristic comparison module 2123;

first facial feature comparison module 2123: the system is used for receiving the facial features sent by the first facial feature acquisition module 2121 and the certificate photo sent by the identity information acquisition module 2122, comparing all the facial features with the certificate photo, calculating the similarity Tq of the facial features and the certificate photo, comparing the maximum similarity Tq with a similarity threshold Tp, sending an instruction to the dialogue module 22 if the verification is successful if the Tq is greater than or equal to Tp, and sending an instruction to the command execution module 23 if the verification is failed if the Tq is less than Tp, wherein the instruction content is exit.

The method mainly aims at services needing to be transacted by identity documents, such as buying train tickets, transacting passports, visas and the like; the active interactive dialogue robot system provided by the embodiment mainly obtains the facial features of the user and the certificate photos on the identity document of the user respectively, compares the facial features of the user and the certificate photos on the identity document of the user, and judges that the user transacts the active interactive dialogue robot system when the similarity is higher than a set threshold value, so that the problem is sent to the user.

Example 5

In this embodiment, the problem generation module is further limited, as shown in fig. 5, the verification module 21 further includes a double verification module 213, and the double verification module 213 includes:

double face information identifier module 2131: the system comprises a double facial feature comparison module 2133, a user facial feature comparison module 2133 and a user facial feature comparison module 2133, wherein the double facial features are used for comparing the two users;

double identity information acquisition module 2132: the system comprises a double face feature comparison module 2133, a double face feature comparison module 2133 and a double face feature comparison module 2133, wherein the double face feature comparison module 2133 is used for scanning identity documents of two users and obtaining two identity document photos;

double facial feature comparison module 2133: the certificate photo recognition module is used for receiving the two groups of facial features sent by the double-person facial information recognition sub-module 2131 and the certificate photo sent by the double-person identity information acquisition module 2132, respectively comparing the two groups of facial features with the corresponding certificate photos, calculating the similarity Tm of the two groups of facial features, comparing the two maximum similarities Tm1 and Tm2 with a similarity threshold value T3, sending an instruction to the dialogue module 22 when Tm1 is greater than or equal to T3 and Tm2 is greater than or equal to T3, and sending an instruction to the command execution module 23 when Tm1 is less than T3 or Tm2 is less than T3, wherein the instruction is quitting.

The active interactive dialogue robot system provided by the embodiment can realize the function of assisting handling by arranging the double-person verification module, particularly, the old people or the people who are not good at using electronic products can be entrusted with others to help handling, when the active interactive dialogue robot system is used, facial features of users and assistant handling personnel are collected and compared with certificate photos on corresponding identity documents, if the similarity of the two is higher than the similarity threshold value, one of the facial features is proved to be the assistant handling personnel, namely, a problem can be sent to the users, if the similarity of one or two of the facial features is lower than the similarity threshold value, one of the identity documents is proved to be unused by one person, handling is quitted at the moment, the legal interest of the users is protected, and the users are prevented from being cheated.

Example 6

The present embodiment further defines the dialog module, and as shown in fig. 6, the dialog module 22 includes:

the question storage module 221: the question generating module 222 is used for storing all questions to be asked in a certain scene, arranging all the questions to be asked in sequence, and sending all the questions arranged in sequence to the question generating module 222 when receiving an instruction sent by the question generating module 222;

the question generation module 222: the system is used for receiving the instruction sent by the verification module 2, sending the instruction to the question storage module 221, sequentially sending the received questions in sequence to the user module 14, and sending the instruction to the answer acquisition module 223;

answer collection module 223: for collecting the answer content of the user module 14 to the question and sending the answer content to the answer checking module 224;

answer verification module 224: the answer acquisition module 223 is used for receiving the answer content sent by the answer acquisition module 223, checking the validity of the answer content, and sending an instruction to the command execution module 23 or the repeated question-asking module 225 according to the check result;

repeat questioning module 225: for clearing the answers entered by the user and repeatedly sending questions to the user module 14 that are not validated;

the answer verification module 224 includes:

digital answer judging module 2241: the answer processing module is used for converting the answer content of the user into a text and judging whether the answer of the user is a number, if so, identifying the number Xn of the number and sending the Xn to the number length judging module 2242, and if not, sending an instruction to the character answer judging module 2243;

number length determination module 2242: the answer receiving module 2241 is configured to receive the number Xn of the digits sent by the answer judging module 2241, judge the size of the Xn and the threshold X1 of the number of the digits, send an instruction to the question generating module 222 and send the next question to the user module 14 if Xn is equal to X1, and send an instruction to the repeat questioning module 225 if Xn is equal to X1;

text answer judging module 2243: the answer judging module is used for judging whether the answer of the user is a character answer or not, if so, identifying the character content in the text and sending the character content to the same character judging module 2244, and if not, sending an instruction to the scanning judging module 2245;

the same character determination module 2244: the answer receiving module 2243 is configured to receive text content sent by the text answer determining module 2243, determine whether there is a word in the text content that is the same as a preset answer, if yes, send an instruction to the question generating module 222, send the next question to the user module 14, and if not, send an instruction to the repeated question module 225;

scanning judgment module 2245: for scanning the certificate provided by the user and judging whether the scanned certificate is qualified, if so, sending an instruction to the question generation module 222 to ask the next question, and if not, sending an instruction to the repeated question module 225.

According to the active interactive dialogue robot system provided by the embodiment, the accuracy of the verification result is improved by specifically defining the answer checking module and judging the validity of the answer of the user in multiple dimensions.

Example 7

In this embodiment, the scan determination module is further limited, and as shown in fig. 7, the scan determination module 2245 includes:

the content matching module 22451: the certificate matching module 22452 is used for judging whether the text content on the certificate provided by the user is consistent with the question, if so, sending an instruction to the certificate matching module 22452, and if not, sending an instruction to the repeated questioning module 225;

credential matching module 22452: the question generation module 22451 is used for receiving the instruction sent by the content matching module 22451, identifying and judging whether the content stamped on the certificate is consistent with the text content, if so, sending an instruction to the question generation module 222 to ask the next question, and if not, judging that the certificate is unqualified and sending an instruction to the repeated question module 225.

The scanning judgment module provided by the embodiment can scan the data provided by the user and judge whether the data is qualified or not, firstly, whether the text content on the certificate provided by the user is matched with the problem or not is judged, if the certificate provided by the user is not matched with the certificate required to be provided by the problem, the question is asked again at the moment, the user can provide the certificate matched with the problem, if the certificate provided by the user is matched and described to be correct, whether the stamping content on the certificate is matched with the text content or not is judged again, if the stamping content is matched, the stamping is proved to be correct, the next problem is provided, if the stamping content is not matched, the stamping is proved to be wrong, the file with the correct stamping needs to be provided again, the problem is asked again at the moment, and the user can provide the file with the correct stamping.

Example 8

In this embodiment, the problem generation module is further limited, and as shown in fig. 8, the number of people determination module 2221: after the question is sent to the user module 14, it is determined that the number P of people the user selects to transact the certificate, if P is equal to 1, an instruction is sent to the primary answer acquisition module 2231, and if P is greater than 1, an instruction is sent to the answer mode module 2222;

answer mode module 2222: the answer acquisition module 2231 is configured to determine whether answers of multiple users to a certain question are the same, and if yes, send an instruction to the answer acquisition module 2232, and if not, send an instruction to the answer acquisition module 2232;

the answer collecting module 223 includes:

primary answer acquisition module 2231: the answer checking module 224 is used for collecting the answer content of the user to the question at one time, sending the answer content to the answer checking module 224, and sending the next question to the user module 14 after the answer checking is successful;

multiple answer collection module 2232: the system is used for collecting the answer content of the user to the question for multiple times, wherein the collection times are the same as the number of the users, all the answer content is sent to the answer verification module 224 according to the collection sequence, and the next question is sent to the user module 14 after all the answer content is verified successfully.

This embodiment carries out further injecing to the problem generation module, through the number of handling of judging the user, it handles or many people are handled to discern one person, be 1 man-hour when the number, then handle according to conventional procedure, when the number is many people, then can carry out many people and handle simultaneously, judge this moment whether the answer of a certain question is the same, if the same, then only gather user's answer can carry out the check-up, if inequality, then need gather user's answer many times, and check-up the legitimacy of answering many times respectively, for example: a plurality of students handle the passport together based on seeking study, when the question of asking is the name, the judged result is that the user is different to the answer of this question, gathers user's answer content many times this moment, when the question of asking is the source of leaving the country, the judged result is that the user is the same to the answer of this question, only gather user's answer this moment can, save user's time and energy, improve certificate and handle efficiency.

Example 9

The present embodiment provides an active interactive dialogue robot system, as shown in fig. 9 and fig. 10, the dialogue robot system further includes a review module 4, and the review module 4 includes:

the list generation module 41: the system is used for generating a list of all the questions and the corresponding answers answered by the user and sending the list to the user module 14 after the user answers all the questions and verifies successfully;

review submodule 42: the system is used for receiving the instruction sent by the user module 14 and identifying the content of the instruction, if the content of the instruction does not need to be modified, the system sends the instruction to the secondary identity verification module 44, and if the content of the instruction needs to be modified, the system sends the instruction to the problem selection module 43;

question selection module 43: the system is used for receiving the instruction sent by the rechecking submodule 42, calling the problem to be modified and sending the problem to the user module 14;

the secondary authentication module 44: the command execution module 23 is used for performing identity authentication again after the user sends the command which does not need to be modified, and sending the command to the command execution module according to the authentication result;

the secondary authentication module 44 includes:

second facial feature acquisition module 441: the facial feature comparison module 442 is used for collecting facial features of the user again after the user sends an instruction which does not need to be modified;

second face feature comparison module 442: and the module is configured to compare the facial features acquired by the second facial feature acquisition module 441 with the facial features acquired by the first facial feature acquisition module 211, determine whether the facial features match each other, send an instruction to the command execution module 23 if the facial features match each other, where the instruction content is to execute a corresponding service, and send an instruction to the command execution module 23 if the facial features do not match each other, where the instruction content is to exit.

In the active interactive dialogue robot system provided by this embodiment, by setting the rechecking module, a list is sent to the user after all answers of the user are successfully checked, so that the user can determine whether all answer information is correct, if the user authorization does not need to be modified, the user can handle the certificate at this time, if the user authorization needs to be modified, a question needing to be modified is sent to the user, and at this time, the user can modify the question with an incorrect answer, thereby avoiding a certificate handling error caused by a mouth error or an entry error in the handling process of the user; the secondary identity authentication module is further limited, the user is authenticated again after the answer content is confirmed by the user without modification, if the facial features of the user are matched with the facial features collected before the user is handled, the user is proved to be still handled at the moment, corresponding certificates can be handled for the user, if the facial features of the user are not matched with the facial features collected before the user is handled, the fact that the handling personnel and the user before the user is handled are proved not to be the same person at the moment is proved, the user quits the handling, unnecessary loss caused to the user is avoided, and the certificate handling safety is further improved.

Example 10

In this embodiment, the secondary identity verification module is further limited, and as shown in fig. 11, the satisfaction investigation module 13 includes:

the mouth expression recognition module 131: the method is used for collecting facial features of a plurality of clients during communication with the clients and identifying mouth expressions of the clients, and specifically comprises the following steps: the number of puckered mouths F1, the number of smiles F2, the number of beep mouths F3, the number of outmoded smiles F4;

eye expression recognition module 132: the method is used for collecting facial features of a plurality of clients during communication with the clients and identifying eye expressions of the clients, and specifically comprises the following steps: the number of times of glaring G1 and the number of times of white eyes G2;

the eyebrow expression recognition module 133: the method is used for collecting facial features of a plurality of clients during communication with the clients and identifying eye expressions of the clients, and specifically comprises the following steps: the number of eyebrow plucking times H1 and the number of eyebrow crumpling times H2;

the head motion recognition module 134: the method is used for collecting facial features of a plurality of clients during communication with the clients and identifying eye expressions of the clients, and specifically comprises the following steps: the number of shaking heads J1, the number of nodding heads J2;

score calculation module 135: the facial expression recognition module 131, the eye expression recognition module 132, the eyebrow expression recognition module 133 and the head movement recognition module 134 are used for collecting the expressions recognized by the facial expression recognition module and calculating the satisfaction degree Q, and Q is calculated by the following method; Q-F1 × F1+ F21 × F2-F3 × F3+ F4 × F4-G1 × G1-G2 × G2-H1 × H1-H2 × H2-J1 × J1+ J2 × J2, wherein F1+ F2+ F3+ F4+ G1+ G2+ H1+ H2+ J1+ J2 equals 100%.

The satisfaction survey module is further limited by the embodiment, so that the calculated satisfaction is more accurate.

Claims

1. An active interactive dialogue robot system is characterized by comprising a pre-sale consultation module (1) and a mid-sale transaction module (2);

the pre-sale consultation module (1) comprises a conversation storage module (11), an answer module (12), a satisfaction survey module (13) and a user module (14);

the telephony storage module (11): the system is used for creating an industrial speech and skill library and customizing a dialogue automatic question and answer corpus which accords with the business;

the answer module (12): the system is used for collecting the questions sent by the user module (14), matching corresponding answers in the industrial and tactical library and sending the answers to the user module (14);

the satisfaction survey module (13): the system is used for investigating the satisfaction degree of the user on the consultation when or after the consultation is finished;

the user module (14): for the user to ask questions, answer questions and receive answers sent by the answer module (12);

the in-sale transaction module (2) comprises a verification module (21), a dialogue module (22) and a command execution module (23);

the verification module (21): the system is used for verifying whether the user transacts for oneself or not;

the dialog module (22): the system is used for receiving the instruction sent by the verification module (21), proposing a question to a user according to an application scene, collecting the answer of the user, checking the legality of the answer of the user, and sending the instruction to the command execution module (23) according to a checking result;

the command execution module (23): for executing the corresponding service command or exiting;

the authentication module (21) comprises a voice authentication module (211);

the voice verification module (211): the voice authentication module is used for sending voice authentication information to a user, judging whether the user transacts the user or not, if so, sending an instruction to the dialogue module (22), and if not, sending an instruction to the command execution module (23), wherein the instruction content is exit;

the authentication module (21) further comprises an image authentication module (212), the image authentication module (212) comprising:

a first facial feature acquisition module (2121): the face feature comparison module is used for acquiring and identifying a plurality of face features of a user and sending the identified face features to the first face feature comparison module (2123);

identity information acquisition module (2122): the system comprises a first face characteristic comparison module (2123) for scanning the identity document of the user, identifying a document photo on the identity document and sending the document photo to the first face characteristic comparison module;

a first facial feature comparison module (2123): the system is used for receiving the facial features sent by the first facial feature acquisition module (2121) and the certificate photos sent by the identity information acquisition module (2122), comparing all the facial features with the certificate photos, calculating the similarity Tq of the facial features and the certificate photos, comparing the maximum similarity Tq with a similarity threshold Tp, sending an instruction to the dialogue module (22) if the verification is successful if the Tq is larger than or equal to Tp, and sending an instruction to the command execution module (23) if the verification is failed if the Tq is smaller than Tp, wherein the instruction content is exit;

the authentication module (21) further comprises a double authentication module (213), the double authentication module (213) comprising:

double face information identification submodule (2131): the system comprises a double-person facial feature comparison module (2133) and a user facial feature comparison module (HIC) which are used for acquiring two groups of facial features of two users at the same time, wherein each group of facial features comprises a plurality of facial features of one user, and the two groups of facial features are sent to the double-person facial feature comparison module;

double identity information acquisition module (2132): the double-person facial feature comparison module (2133) is used for scanning the identity documents of the two users, acquiring the document photos on the two identity documents and sending the two acquired document photos to the double-person facial feature comparison module;

double facial feature comparison module (2133): the certificate photo recognition module is used for receiving two groups of facial features sent by the double-person facial information recognition sub-module (2131) and the certificate photo sent by the double-person identity information acquisition module (2132), respectively comparing the two groups of facial features with the corresponding certificate photo, calculating the similarity Tm of the two groups of facial features, comparing the two maximum similarities Tm1 and Tm2 with a similarity threshold value T3, when Tm1 is more than or equal to T3 and Tm2 is more than or equal to T3, successfully verifying, sending an instruction to the dialogue module (22), and when Tm1 is less than T3 or Tm2 is less than T3, sending an instruction to the command execution module (23), wherein the instruction content is exit.

2. The active interactive dialog robot system of claim 1 characterized in that the dialog robot system further comprises a troubleshooting module (3), the troubleshooting module (3) comprising:

fault monitoring module (31): the working state of the conversation robot system is collected constantly, whether the working state of the conversation robot is abnormal or not is judged, if the working state of the conversation robot is abnormal, an instruction is sent to a fault processing module (32), and if the working state of the conversation robot is normal, no processing is carried out;

fault handling module (32): for calling the relevant staff, who will handle the fault at the first time.

3. The active interactive dialogue robot system according to claim 2, wherein the fault monitoring module (31) comprises an answer speed judgment module (311), a satisfaction survey judgment module (312), a voice verification judgment module (313), an execution judgment module (314);

-said answer speed decision module (311): the time T1 for judging the question sending time of the user module (14) and the time T2 for answering the question by the dialogue robot system, the sizes of a time difference threshold Tn and a time difference T2-T1 are judged, if the time T2-T1 is more than or equal to the Tn, an instruction is sent to the fault processing module (32), and if the time T2-T1 is less than the Tn, an instruction is sent to the satisfaction survey judging module (312);

the satisfaction survey judging module (312): the voice verification and judgment module is used for judging whether to send a satisfaction survey short message or a voice call to the user after the consultation of the user is finished, if so, sending an instruction to the voice verification and judgment module (313), and if not, sending an instruction to the fault processing module (32);

the voice verification judgment module (313): the voice authentication system is used for judging whether to execute voice authentication to a user before executing related services, if so, sending an instruction to an execution judging module (314), and if not, sending an instruction to a fault processing module (32);

the execution determination module (314): the interactive robot is used for judging whether the interactive robot transacts related services for the user after the user selects to purchase, if so, the interactive robot does not process, and if not, the interactive robot sends an instruction to the fault processing module (32).

4. The active interactive dialog robot system of claim 1 wherein the dialog module (22) comprises:

question storage module (221): the question storage module is used for storing all questions to be asked in a certain scene, arranging all the questions to be asked in sequence, and sending all the questions arranged in sequence to the question generation module (222) when receiving an instruction sent by the question generation module (222);

problem generation module (222): the system is used for receiving the instruction sent by the verification module (2), sending the instruction to the question storage module (221), sequentially sending the received questions in sequence to the user module (14) and sending the instruction to the answer acquisition module (223);

answer acquisition module (223): the answer verification module is used for collecting answer content of the user module (14) to the question and sending the answer content to the answer verification module (224);

answer verification module (224): the answer collecting module (223) is used for receiving the answer content sent by the answer collecting module, checking the validity of the answer content, and sending an instruction to the command executing module (23) or the repeated questioning module (225) according to the checking result;

repeat questioning module (225): for emptying the answers entered by the user and repeatedly sending questions to the user module (14) that are not validated;

the answer verification module (224) comprises:

a digital answer judging module (2241): the system is used for converting the user answer content into a text and judging whether the answer of the user is a number, if so, identifying the number Xn of the number and sending the Xn to a number length judging module (2242), and if not, sending an instruction to a character answer judging module (2243);

a digital length determination module (2242): the answer-question processing module is used for receiving the number Xn of the numbers sent by the answer-question judging module (2241), judging the size of the Xn and a number threshold X1, if the Xn is X1, sending an instruction to the question generating module (222), sending the next question to the user module (14), and if the Xn is not X1, sending an instruction to the repeated question module (225);

a text answer judging module (2243): the answer judging module is used for judging whether the answer of the user is a character answer or not, if so, identifying the character content in the text and sending the character content to the same character judging module (2244), and if not, sending an instruction to the scanning judging module (2245);

the same character judgment module (2244): the question generation module (2243) is used for receiving the text content sent by the text answer judgment module (2243), judging whether words which are the same as preset answers exist in the text content, if so, sending an instruction to the question generation module (222), sending the next question to the user module (14), and if not, sending an instruction to the repeated question module (225);

a scan judgment module (2245): the certificate question issuing system is used for scanning certificates provided by users and judging whether the scanned certificates are qualified or not, if so, sending an instruction to the question generating module (222) to issue the next question, and if not, sending an instruction to the repeated question module (225).

5. The active interactive dialog robot system of claim 4 wherein the scan determination module (2245) comprises:

content matching module (22451): the certificate matching module is used for judging whether the text content on the certificate provided by the user is consistent with the question or not, if so, sending an instruction to the certificate matching module (22452), and if not, sending an instruction to the repeated questioning module (225);

credential matching module (22452): the question generation module (22451) is used for receiving the instruction sent by the content matching module (22451), identifying and judging whether the content stamped on the certificate is consistent with the text content, if so, sending the instruction to the question generation module (222), proposing the next question, if not, judging that the question is not qualified, and sending the instruction to the repeated question module (225).

6. The active interactive dialog robot system of claim 4 wherein the question generation module (222) comprises:

a number of people determination module (2221): the system is used for judging the number P of the users who choose to transact the certificates after the questions are sent to the user module (14), if P is equal to 1, an instruction is sent to the primary answer acquisition module (2231), and if P is larger than 1, an instruction is sent to the answer mode module (2222);

answer mode module (2222): the answer collecting module is used for judging whether answers of a plurality of users for answering a certain question are the same or not, if yes, an instruction is sent to the primary answer collecting module (2231), and if not, an instruction is sent to the multi-time answer collecting module (2232);

the answer collection module (223) comprises:

primary answer acquisition module (2231): the answer checking module is used for collecting answer contents of the user to the questions at one time, sending the answer contents to the answer checking module (224), and sending the next question to the user module (14) after the answer checking module successfully checks the answer contents;

multiple answer acquisition module (2232): the system is used for collecting the answer content of the user to the question for multiple times, the collection times are the same as the number of the user, all the answer content is sent to an answer checking module (224) according to the collection sequence, and the next question is sent to the user module (14) after all the answer content is checked successfully.

7. The active interactive dialog robot system of claim 1 characterized in that the dialog robot system further comprises a review module (4), the review module (4) comprising:

list generation module (41): the system is used for generating a list of all questions and corresponding answers answered by the user and sending the list to the user module (14) after the user answers all questions and verifies successfully;

review submodule (42): the system is used for receiving an instruction sent by a user module (14) and identifying the content of the instruction, if the content of the instruction does not need to be modified, the instruction is sent to a secondary identity verification module (44), and if the content of the instruction needs to be modified, the instruction is sent to a problem selection module (43);

question selection module (43): the system is used for receiving the instruction sent by the rechecking submodule (42), calling the problem to be modified and sending the problem to the user module (14);

the secondary authentication module (44): the system is used for carrying out identity authentication again after a user sends an instruction which does not need to be modified, and sending the instruction to the command execution module (23) according to an authentication result;

the secondary authentication module (44) comprises:

a second facial feature acquisition module (441): the facial feature comparison module is used for acquiring facial features of the user again after the user sends an instruction without modification and sending the facial features to the second facial feature comparison module (442);

a second facial feature comparison module (442): the face feature acquiring module is used for comparing the face features acquired by the second face feature acquiring module (441) with the face features acquired by the first face feature acquiring module (211), judging whether the face features are matched, if so, sending an instruction to the command executing module (23), wherein the instruction content is to execute corresponding business, and if not, sending an instruction to the command executing module (23), and the instruction content is exit.

8. The active interactive dialog robot system of claim 1 characterized in that the satisfaction survey module (13) comprises:

mouth expression recognition module (131): the method is used for collecting facial features of a plurality of clients during communication with the clients and identifying mouth expressions of the clients, and specifically comprises the following steps: the number of puckered mouths F1, the number of smiles F2, the number of beep mouths F3, the number of outmoded smiles F4;

an ocular expression recognition module (132): the method is used for collecting facial features of a plurality of clients during communication with the clients and identifying eye expressions of the clients, and specifically comprises the following steps: the number of times of glaring G1 and the number of times of white eyes G2;

an eyebrow expression recognition module (133): the method is used for collecting facial features of a plurality of clients during communication with the clients and identifying eye expressions of the clients, and specifically comprises the following steps: the number of eyebrow plucking times H1 and the number of eyebrow crumpling times H2;

head motion recognition module (134): the method is used for collecting facial features of a plurality of clients during communication with the clients and identifying eye expressions of the clients, and specifically comprises the following steps: the number of shaking heads J1, the number of nodding heads J2;

score calculation module (135): the facial expression recognition module (131), the eye expression recognition module (132), the eyebrow expression recognition module (133) and the head action recognition module (134) are used for collecting the expressions recognized by the mouth expression recognition module and calculating the satisfaction degree Q, and the Q is calculated by the following method; Q-F1 × F1+ F21 × F2-F3 × F3+ F4 × F4-G1 × G1-G2 × G2-H1 × H1-H2 × H2-J1 × J1+ J2 × J2, wherein F1+ F2+ F3+ F4+ G1+ G2+ H1+ H2+ J1+ J2 equals 100%.