WO2020073495A1 - Artificial intelligence-based reexamination method, apparatus, and device, and storage medium - Google Patents

Abstract

An artificial intelligence-based reexamination method, apparatus, and device, and a storage medium, relating to the technical field of artificial intelligence. The method comprises: receiving a reexamination request triggered by a user, and obtaining, according to the reexamination request, information to be reexamined submitted by the user and the service number of a service corresponding to the information to be reexamined (S10); searching a service approval progress sheet for past approval information of the service corresponding to the service number according to the service number (S20); determining a reexamination start node and a reexamination party according to the information to be reexamined and the past approval information (S30); and submitting the information to be reexamined to the reexamination start node, so that the reexamination party reexamines the information to be reexamined (S40).

Description

Review method, device, equipment and storage medium based on artificial intelligence The

This application requires the priority of the Chinese patent application filed on October 08, 2018 in the China Patent Office with the application number 201811167589.9 and the invention titled "Review Method, Device, Equipment and Storage Media Based on Artificial Intelligence" The reference is incorporated in the application.

Technical field

This application relates to the field of artificial intelligence technology, and in particular to a review method, device, equipment and storage medium based on artificial intelligence.

Background technique

With the in-depth popularization of electronic informatization, many things that used to be done with physical paper materials in the past have also gradually transitioned towards the development of electronic information processing. For example, in the past, when submitting application materials to enterprise organizations and business departments, paper application materials were used to complete the submission and follow-up review of the materials by mail. Now, you can directly log in to the website of enterprise organizations and business departments and provide the electronic The business process flow interface uploads and submits information about the relevant business to be reviewed.

However, in the existing electronic business audit method, if a certain process link of the business does not meet the requirements and is returned during the review process, the applicant will revise according to the reason for the return and submit it again for review, the current review The method can only be reviewed from the beginning, or it can only be returned to the reviewer who made the rejection operation for review. Either way, the unnecessary reviewer will participate in the review of the rejection process again during the review process, which not only results in wasted human resources, but also reduces the efficiency of the review.

Therefore, there is an urgent need to provide an intelligent review method to reduce unnecessary reviewers to participate in the review of the rejection process again.

Summary of the invention

The main purpose of this application is to provide a review method, device, equipment and storage medium based on artificial intelligence, aiming to solve the problem that in the prior art, due to the single review mode in the electronic business processing process, unnecessary reviewers repeatedly participate in the review process , Causing technical problems that waste human resources.

In order to achieve the above purpose, the present application provides a review method based on artificial intelligence. The method includes the following steps:

Receiving a review request triggered by the user, obtaining the service ID of the service submitted by the user and the service corresponding to the service review according to the review request, the service includes at least one approval process node, and each approval process node corresponds to at least one review square;

According to the business number, search the history approval information of the business corresponding to the business number in the business approval schedule;

According to the information to be reviewed and the historical approval information, determine the review start node and review party;

Submit the information to be reviewed to the review initiation node, so that the reviewing party can review the information to be reviewed.

In addition, in order to achieve the above object, the present application also proposes a review device based on artificial intelligence, which includes:

The obtaining module is used for receiving a review request triggered by a user, and obtaining the service ID of the service submitted by the user and the service corresponding to the service review according to the request, the service includes at least one approval process node, and each approval process The node corresponds to at least one reviewer;

A search module, configured to search the history approval information of the business corresponding to the business number in the business approval schedule according to the business number;

The determining module is used to determine the starting node and the reviewing party based on the information to be reviewed and the historical approval information;

The submitting module is used to submit the information to be reviewed to the review initiation node, so that the reviewing party can review the information to be reviewed.

In addition, in order to achieve the above object, the present application also proposes an artificial intelligence-based review device, the device includes: a memory, a processor, and an artificial intelligence-based review device stored on the memory and capable of running on the processor Review procedure, the artificial intelligence-based review procedure is configured to implement the steps of the artificial intelligence-based review method as described above.

In addition, in order to achieve the above object, the present application also proposes a storage medium on which an artificial intelligence-based review program is stored, and the artificial intelligence-based review program is implemented by the processor as described above based on The steps of the artificial intelligence review method.

The artificial intelligence-based review method, device, device, and storage medium of this embodiment obtain the information to be reviewed by the user and the service number of the business corresponding to the information to be reviewed according to the review request triggered by the user. The business number looks up the historical approval information of the corresponding business in the business approval schedule, and then uses artificial intelligence technology to determine the review initiation node and reviewer based on the pending review information and historical approval information, so that the user-submitted review information can be directly submitted to The review process node related to the information to be reviewed, and the reviewer who is specifically responsible for reviewing the content contained in the information to be reviewed, greatly simplifies the review process, and reduces the waste of human resources and improves the efficiency of the review. For a better experience.

BRIEF DESCRIPTION

1 is a schematic structural diagram of an artificial intelligence-based review device for a hardware operating environment involved in an embodiment of the present application;

2 is a schematic flowchart of a first embodiment of a review method based on artificial intelligence of this application;

3 is a schematic flowchart of a second embodiment of a review method based on artificial intelligence of this application;

4 is a structural block diagram of a first embodiment of an artificial intelligence-based review device of the present application.

The implementation, functional characteristics and advantages of the present application will be further described in conjunction with the embodiments and with reference to the drawings.

detailed description

It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

Referring to FIG. 1, FIG. 1 is a schematic structural diagram of an artificial intelligence-based review device for a hardware operating environment involved in an embodiment of the present application.

As shown in FIG. 1, the artificial intelligence-based review device may include: a processor 1001, such as a central processor (Central Processing Unit, CPU), communication bus 1002, user interface 1003, network interface 1004, memory 1005. Among them, the communication bus 1002 is used to implement connection communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (WIreless-FIdelity, WI-FI) interface). The memory 1005 may be a high-speed random access memory (Random Access Memory (RAM) memory can also be a stable non-volatile memory (Non-Volatile Memory, NVM), such as disk storage. The memory 1005 may optionally be a storage device independent of the foregoing processor 1001.

Those skilled in the art can understand that the structure shown in FIG. 1 does not constitute a limitation on the artificial intelligence-based review device, and may include more or fewer components than those illustrated, or combine certain components, or different components Layout.

As shown in FIG. 1, the memory 1005 as a storage medium may include an operating system, a data storage module, a network communication module, a user interface module, and a review program based on artificial intelligence.

In the artificial intelligence-based review device shown in FIG. 1, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processing in the artificial intelligence-based review device of this application The device 1001 and the memory 1005 may be provided in an artificial intelligence-based review device that calls the artificial intelligence-based review program stored in the memory 1005 through the processor 1001 and executes the The review method of artificial intelligence.

An embodiment of the present application provides an artificial intelligence-based review method. Referring to FIG. 2, FIG. 2 is a schematic flowchart of a first embodiment of an artificial intelligence-based review method of the present application.

In this embodiment, the review method based on artificial intelligence includes the following steps:

Step S10: Receive a review request triggered by the user, and obtain the information to be reviewed submitted by the user and the service number of the service corresponding to the information to be reviewed according to the review request.

Specifically, the execution subject in this embodiment is specifically a user terminal capable of interacting with the user.

It should be understood that in practical applications, the user terminal mentioned above may specifically be a smart terminal device that can access the network to log in to the websites of enterprise organizations and business departments, such as smartphones, tablets, personal computers, etc., here I won't list them one by one, and I don't make any restrictions on this.

In addition, it should be understood that the user-triggered review request mentioned above may specifically be a corresponding client application (Application, APP) installed on the user terminal to access the website of an enterprise institution or business department by clicking on the user terminal. Either the browser version or the PC version (the version running on the personal computer) is triggered when the submit review function button is pressed.

Correspondingly, under normal circumstances, the user-triggered function button for submitting a review is modified according to the reason for rejection and uploaded after the information to be reviewed is clicked, and the above operations are usually performed on the business corresponding to the information to be reviewed This is done under the interface, so when the review request is triggered, the information to be reviewed submitted by the user and the business number of the corresponding business will be extracted from the corresponding area according to the preset rules.

In addition, the information to be reviewed in this embodiment may specifically be text information, that is, text format, such as a mobile phone number, name, address, etc., or a picture format, such as a picture of a user's ID card, or a user Face images etc.

In addition, the above-mentioned business corresponding to the information to be reviewed is an electronic business submitted online by the user.

Correspondingly, the service number is the order number of the electronic service or other identification code that can identify its uniqueness.

In addition, it is worth mentioning that in practical applications, there may be multiple approval process nodes under a business. Taking credit business as an example, a credit business application submitted by a user online usually involves a number of approval process nodes, such as a user's basic data approval process node, a credit enhancement certification data approval process node, and a face-to-face approval process node. In addition, in order to ensure the efficiency of the review, different reviewers are usually arranged to be responsible for the approval of different approval process nodes. Therefore, the business mentioned in this embodiment includes at least one approval process node, and each approval process node may correspond to at least one reviewer.

It should be noted that the above is only an example, and does not constitute any limitation to the technical solution of the present application. In a specific implementation, a person skilled in the art can set page logic reasonably according to needs, and no limitation is made here.

Step S20: According to the service number, search the history approval information of the service corresponding to the service number in the service approval schedule.

Specifically, the above-mentioned business approval schedule is specifically a reviewer responsible for reviewing the service applied by the user, and records the progress made according to the processing progress when processing the approval process node that he is responsible for.

Specifically, in this embodiment, the historical approval information mentioned above mainly includes the review status made by each reviewer corresponding to each approval process node on the current approval process node.

In addition, in the present embodiment, the above-mentioned audit status is roughly classified into several categories of processed, in-process, and unprocessed.

Further, the processed audit status can also be subdivided into rejected status and approved status.

For ease of understanding, the following examples illustrate:

For example, use "0" for unprocessed, that is, every step in the approval process has not been reviewed; use "1" for processed, that is, all steps in the approval process have been reviewed; use "2" The representative is processing, that is, the approval process is being reviewed, and some steps have been reviewed.

Further, when the audit status is "1", other agreed identification information can also be used to identify which steps have been approved and which step is rejected; when the audit status is "2", other The agreed identification information identifies which steps have been reviewed and which are unaudited.

In addition, in practical applications, the reviewing parties of each approval process link and each step under each approval process node may be different. Even if the same reviewer takes into account multiple approval process nodes or multiple steps, its review work is also different. Therefore, it is also possible to set a specific number, etc., for the reviewer corresponding to each sub-step in advance, so that after the dismissal operation is made, it is convenient for who makes subsequent queries.

In addition, in specific applications, the historical approval information found may also include the reason for rejection.

It should be noted that the above is only an example, and does not constitute any limitation on the technical solution of the present application. In a specific implementation, a person skilled in the art can set a more reasonable classification according to the need for the reviewer to choose. There are no restrictions here.

Step S30: According to the information to be reviewed and the historical approval information, determine a review start node and a review party.

Specifically, since the historical approval information mentioned in this embodiment mainly includes the review status made by each reviewer corresponding to each approval process node on the current approval process node, when determining the review start node, the specific Each review party corresponding to each review process node in the historical review information and the review approval process reviews the review status of the current review process node, and selects one review process node as the review start node.

In order to better understand the operation of the review start node, the following specific instructions:

First, according to the review status of each review party corresponding to each review process node on the current review process node, the review process node whose review status is processed is selected to obtain a candidate set of review start nodes, and the processed review process node is The approval status is the approval process node of the rejected status or approved status.

Then, based on the rejection information corresponding to the approval review node and the approval process node whose approval status is the rejection status, select an approval process node from the candidate set of review activation nodes as the review initiation node.

By using the above-mentioned methods to determine the review start node, without changing the user's habits and increasing their operating burden, with the help of artificial intelligence technology, through the use of machine learning method for pre-learning training, which can be based on the user to submit review information The content and historical approval information quickly locate the approval process node that is specifically responsible for reviewing the content in the information to be reviewed as the review initiation node, which realizes that the information to be reviewed only needs to be submitted to the related approval process node for subsequent review work, effectively speeding up Review efficiency.

It should be noted that the above is only a specific implementation manner for determining a review start node, and does not constitute any limitation on the technical solution of the present application.

In addition, regarding the above-mentioned method of determining the reviewing party based on the information to be reviewed and the historical approval information, it may specifically be the review work undertaken by each reviewer corresponding to the review-to-review information and the review initiation node Description information, select at least one reviewer from each reviewer corresponding to the review start node as the reviewer.

For example, through keyword extraction technology, the first keyword is extracted from the information to be reviewed separately, the second keyword is extracted from the description information of the audit work of each reviewer, and then the first keyword and the second keyword are compared For the similarity between them, the reviewer corresponding to the second keyword with the highest similarity is selected as the reviewer, or the reviewer corresponding to each of the second keywords whose similarity is greater than the preset threshold is used as the reviewer.

Step S40: Submit the information to be reviewed to the review initiation node, so that the reviewing party can review the information to be reviewed.

Specifically, when submitting the information to be reviewed to the review initiation node, in order to ensure the security of the information to be reviewed submitted by the user, the review information can be encrypted first, and then the encrypted information to be reviewed can be sent To the server of the enterprise organization or business department accessed by the user, and then delivered to the terminal device of the reviewing party by the server.

It should be noted that the above is only an example, and does not constitute any limitation to the technical solution of the present application. In a specific implementation, a person skilled in the art may set it as needed, and there is no limitation here.

From the above description, it is not difficult to find that the artificial intelligence-based review method provided in this embodiment obtains the service number of the service submitted by the user and the service corresponding to the service to be reviewed according to the request triggered by the user The searched business number is searched for in the business approval schedule to find the corresponding historical approval information of the business, and then use artificial intelligence technology to determine the review start node and reviewer based on the pending review information and historical approval information, so that the user can submit the pending review information directly Submitted to the approval process node related to the information to be reviewed, and reviewed by the reviewer who is specifically responsible for reviewing the content contained in the information to be reviewed, which greatly simplifies the review process and reduces the waste of human resources and improves the efficiency of the review. Brings a better experience.

Referring to FIG. 3, FIG. 3 is a schematic flowchart of a second embodiment of an artificial intelligence-based review method of the present application.

Based on the foregoing first embodiment, the review method based on artificial intelligence in this embodiment before step S30 further includes:

Step S00, it is determined that the information to be reviewed meets the review requirements.

The operation described in step S00 to determine that the information to be reviewed meets the review conditions can be roughly achieved through four sub-steps, and the four sub-steps will be specifically described below.

Step S01: Acquire the reason for rejection corresponding to the service according to the service number.

Specifically, the reason for rejection may be stored in the above-mentioned business approval schedule or may be stored separately.

In addition, in order to facilitate searching, a mapping relationship table can be established, and the mapping relationship table is used to store the relationship between the service number of each service and the reason for rejection of each service.

In addition, in order to ensure the accuracy of the rejection reasons found according to the mapping relationship table, each time the reviewer makes a rejection operation for the service corresponding to the service number, the corresponding rejection reason may be recorded.

In step S02, according to the reason for rejection, an audit rule corresponding to the reason for rejection is obtained.

Specifically, the aforementioned audit rules are pre-set according to the problems that may be encountered during the audit process.

For ease of understanding, the following will still take the application of the electronic business as a credit business for example.

For example, when the reason for rejection is that the photo of the credit business applicant has not been uploaded, the corresponding review rule may be: the pixel required by the photo, the background color, and the facial features need to be in the middle of the screen.

For another example, when the residence address of the credit business applicant is not filled in for reasons of rejection, the corresponding review rule may be: the address needs to include the country, province, urban area, street, community, building number, etc.

It should be noted that the above is only an example, and does not constitute any limitation on the technical solution of the present application. In a specific implementation, a person skilled in the art may set it as needed, and no limitation is made here.

Step S03, judging whether the information to be reviewed meets the requirements of the review rules, and if the information to be reviewed meets the requirements of the review rules, it is determined that the information to be reviewed meets the review conditions.

In addition, when it is determined that the information to be reviewed does not meet the requirements of the review rules, that is, the information to be reviewed does not meet the review conditions, in order to assist the user to make appropriate modifications to the currently submitted information to be reviewed, to ensure that the submitted The review information can successfully pass the subsequent review, and the review rule can be reported to the user, so that the user can modify the information to be reviewed according to the review rule.

In addition, it should be noted that, in practical applications, the obtained information to be reviewed may be text information, or may be information in various formats such as the user's face image, so in order to facilitate understanding of the basis mentioned in step S03 The audited rules determine whether the information to be reviewed meets the review conditions. The following briefly describes the above two formats.

(1) When the information to be reviewed is text information, determining whether the information to be reviewed meets the review conditions according to the audited rules includes:

First, use the keyword extraction method to extract keywords from the text information.

Specifically, in practical applications, in order to ensure the reference value of the extracted keywords, before extracting keywords from the text information using the keyword extraction method, the text information may be pre-processed, for example, To stop words, that is, to remove words that have no actual meaning in the current disease information, such as: ,,,,, etc., for example, to remove invalid special characters, such as emoticons, various punctuation marks, etc., no longer here List them one by one, and do not limit them.

Then, according to the review rules, keywords that contain the content specified in the review rules are selected.

For example, when the information to be reviewed is the applicant's residential address, the content specified in the review rules may specifically be the words province, city, county, town, district, street, community, building, etc.

Finally, it is determined whether the proportion of keywords containing the content specified in the review rule is not lower than the first threshold specified in the review rule. If the proportion of keywords containing the content specified in the review rule is not lower than the first threshold, the information to be reviewed meets the requirements of the review rule.

(2) When the information to be reviewed is a user's face image, determining whether the information to be reviewed meets the conditions for review according to the reviewed rules includes:

First, the user's facial features (eyebrows, glasses, nose, mouth, ears) contour information is extracted from the face image, and the image matrix corresponding to the facial features contour information is determined according to the facial features contour information.

Specifically, the operation of extracting the user's facial features contour information from the face image may be as follows: performing grayscale processing on the face image to obtain a grayscale image; and sequentially performing binary value on the grayscale image Processing and smooth denoising to obtain a binary image with interference information removed; using the edge detection method, the user's facial features contour information is extracted from the binary image.

Then, the image matrix is compared with the reference template matrix specified in the review rule to determine the similarity value between the image matrix and the reference template matrix.

It should be understood that the above-mentioned reference template matrix is also obtained based on the facial features contour information, but the facial feature contour information corresponding to the reference template matrix is in accordance with the requirements of the audit rules (such as the facial features need to be symmetrical and appear completely on the screen) Information of preset facial features.

Finally, determine whether the similarity value is not lower than the second threshold specified in the review rule, and if the similarity value is not lower than the second threshold, the information to be reviewed meets the requirements of the review rule .

Specifically, in order to ensure the accuracy of the judgment result, the two matrices to be compared need to be two matrices with the same pixel. That is, if the reference template matrix is 20 * 20 pixels in size, the generated image matrix should also be 20 * 20 pixels in size.

It should be noted that the above is only a specific implementation method for judging whether the information to be reviewed meets the conditions for review according to the review rules given for the information to be reviewed in text format and image format, which is not to the technical solution of the present application. The composition is limited. In practical applications, the format of the information to be reviewed may also be a voice format. For the judgment of the information to be reviewed in the voice format, operations such as filtering and interference removal may be performed first, and then the voice information is clearly To make judgments, specific implementations can be set by those skilled in the art according to needs, which will not be repeated or limited here.

From the above description, it is not difficult to find that the artificial intelligence-based review method provided in this embodiment determines whether the information to be reviewed meets the review conditions before determining the review start node. If the review conditions are met, the operation to determine the review start node is performed , That is, before submitting the information to be reviewed to the review initiation node, the reviewer will conduct a rough review of the review information before the review, so as to ensure that the information to be reviewed by the reviewer is as effective as possible Information to make the follow-up audit smoother, avoiding rejection again, and affecting the efficiency of the audit.

It should be noted that those of ordinary skill in the art may understand that all or part of the steps to implement the above embodiments may be completed by hardware, or may be completed by a program instructing related hardware. The program may be stored in a computer-readable In the storage medium, the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk.

Referring to FIG. 4, FIG. 4 is a structural block diagram of a first embodiment of an artificial intelligence-based review device of the present application.

As shown in FIG. 4, the artificial intelligence-based review apparatus provided in the embodiment of the present application includes: an acquisition module 4001, a search module 4002, a determination module 4003, and a submission module 4004.

Wherein, the obtaining module 4001 is configured to receive a review request triggered by the user, and obtain the information to be reviewed submitted by the user and the service number of the service corresponding to the information to be reviewed according to the review request.

The searching module 4002 is configured to search the historical approval information of the business corresponding to the business number in the business approval schedule according to the business number.

The determination module 4003 is configured to determine a review start node and a review party based on the information to be reviewed and the historical approval information.

The submitting module 4004 is configured to submit the information to be reviewed to the review initiation node, so that the reviewing party can review the information to be reviewed.

In addition, it should be noted that the business described in this embodiment includes at least one approval process node, and each approval process node corresponds to at least one reviewer; the historical approval information includes each reviewer corresponding to each approval process node to the current approval The audit status made by the process node.

Further, in order to facilitate determination of the review start node and review party, in a specific implementation, the determination module 4003 may be refined into a review start node determination sub-module and a review party determination sub-module.

Specifically, the review start node determination submodule is used to filter out the review process nodes whose processing status is processed according to the review status of each review process node corresponding to each review process node to obtain the review start node Select the set, and select an approval process node from the candidate set of review start nodes as the review start node according to the rejection reason corresponding to the review process information and the approval process node whose review status is in the rejected state.

The reviewing party determination sub-module is configured to select at least one of each reviewing party corresponding to the review starting node based on the information to be reviewed and the description information of the reviewing work of each reviewing party corresponding to the review starting node The reviewer serves as the reviewer.

In addition, the above-mentioned processed approval process node is an approval process node whose approval state is a rejected state or an approved state.

It should be noted that the above is only a specific implementation manner, and does not constitute any limitation to the technical solution of the present application.

It is not difficult to find from the above description that the artificial intelligence-based review device provided in this embodiment, after obtaining the user-to-be-reviewed information submitted by the user and the service number of the business corresponding to the information to be reviewed according to the review request triggered by the user, obtain The searched business number is searched for in the business approval schedule to find the corresponding historical approval information of the business, and then use artificial intelligence technology to determine the review start node and reviewer based on the pending review information and historical approval information, so that the user can submit the pending review information directly Submitted to the approval process node related to the information to be reviewed, and reviewed by the reviewer who is specifically responsible for reviewing the content contained in the information to be reviewed, which greatly simplifies the review process and reduces the waste of human resources and improves the efficiency of the review. Brings a better experience.

It should be noted that the workflow described above is only schematic and does not limit the scope of protection of this application. In practical applications, those skilled in the art can select some or all of them according to actual needs. The purpose of the solution of this embodiment is not limited here.

In addition, for technical details that are not described in detail in this embodiment, reference may be made to the review method based on artificial intelligence provided in any embodiment of this application, which will not be repeated here.

Based on the first embodiment of the artificial intelligence-based review device described above, the second embodiment of the artificial intelligence-based review device of the present application is proposed.

In this embodiment, the artificial intelligence-based review device further includes a judgment module. The judgment module is used for judging whether the information to be reviewed meets the conditions for review, and determines that the information to be reviewed meets the conditions for review.

Specifically, in order to realize the above operation, in the specific implementation, the judgment module may be specifically refined into a rejection reason acquisition submodule, an audit rule acquisition submodule, and a to-be-reviewed information judgment submodule.

Wherein, the reason for refusal obtaining sub-module is used to obtain the reason for refusal corresponding to the service according to the service number.

The audit rule acquisition sub-module is configured to acquire an audit rule corresponding to the reason for rejection based on the reason for rejection.

The information to be reviewed sub-module is used to determine whether the information to be reviewed meets the requirements of the review rules, and when the information to be reviewed meets the requirements of the review rules, it is determined that the information to be reviewed meets the review condition.

In addition, when it is determined that the information to be reviewed does not meet the requirements of the review rules, that is, the information to be reviewed does not meet the review conditions, in order to assist the user to make appropriate modifications to the currently submitted information to be reviewed, to ensure that the submitted The review information can successfully pass the subsequent review. The artificial intelligence review device may further include an auxiliary modification sub-module. The auxiliary modification submodule is used to report the audit rule to the user, so that the user can modify the information to be reviewed according to the audit rule.

In addition, it should be noted that, in practical applications, the obtained information to be reviewed may be text information, or may be information of a user's face image and other formats, so the information to be reviewed sub-module may specifically It is a text information judgment submodule or an image information judgment submodule.

Specifically, when the information to be reviewed is text information, the text information judgment sub-module needs to use a keyword extraction method to extract keywords from the text information, and then screen out the content containing the text according to the review rules Keywords of the content specified in the audit rules, and finally determine whether the proportion of keywords containing the content specified in the audit rules is not lower than the first threshold specified in the audit rules, if the content contains the content specified in the audit rules The proportion of keywords is not lower than the first threshold, then the information to be reviewed meets the requirements of the review rules.

When the information to be reviewed is a user's face image, the image information judgment sub-module needs to extract the user's facial features contour information from the face image, and determine the corresponding facial features contour information according to the facial features contour information Image matrix, then compare the image matrix with the reference template matrix specified in the review rules, determine the similarity value between the image matrix and the reference template matrix, and finally determine whether the similarity value Not lower than the second threshold specified in the review rule, and if the similarity value is not lower than the second threshold, the information to be reviewed meets the requirements of the review rule.

Further, when extracting the user's facial features contour information from the face image, the image information judgment sub-module specifically needs to perform gray-scale processing on the face image to obtain a gray-scale image, and then The grayscale image is sequentially subjected to binarization processing and smooth denoising processing to obtain a binary image with interference information removed, and finally, edge detection method is used to extract user facial features contour information from the binary image.

It should be noted that the above is only a specific implementation method for judging whether the information to be reviewed meets the conditions for review according to the review rules given for the information to be reviewed in text format and image format, which is not to the technical solution of the present application. The composition is limited. In practical applications, the format of the information to be reviewed may also be a voice format. For the judgment of the information to be reviewed in the voice format, operations such as filtering and interference removal may be performed first, and then the voice information is clearly To make judgments, specific implementations can be set by those skilled in the art according to needs, which will not be repeated or limited here.

It is not difficult to find from the above description that the artificial intelligence-based review device provided in this embodiment determines whether the information to be reviewed meets the review conditions before determining the review start node. , That is, before submitting the information to be reviewed to the review initiation node, the reviewer will conduct a rough review of the review information before the review, so as to ensure that the information to be reviewed by the reviewer is as effective as possible Information to make the follow-up audit smoother, avoiding rejection again, and affecting the efficiency of the audit.

It should be noted that the workflow described above is only schematic and does not limit the scope of protection of this application. In practical applications, those skilled in the art can select some or all of them according to actual needs. The purpose of the solution of this embodiment is not limited here.

In addition, for technical details that are not described in detail in this embodiment, reference may be made to the review method based on artificial intelligence provided in any embodiment of this application, which will not be repeated here.

In addition, it should be noted that in this article, the terms "include", "include" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system that includes a series of elements includes not only those Elements, but also include other elements that are not explicitly listed, or include elements inherent to this process, method, article, or system. Without more restrictions, the element defined by the sentence "include one ..." does not exclude that there are other identical elements in the process, method, article or system that includes the element.

The sequence numbers of the above embodiments of the present application are for description only, and do not represent the advantages and disadvantages of the embodiments.

Through the description of the above embodiments, those skilled in the art can clearly understand the above The method of the embodiment can be implemented by means of software plus the necessary general hardware platform, and of course it can also be implemented by hardware, but in many cases the former is the better implementation. Based on this understanding, the The technical solution can be embodied in the form of a software product in essence or part of the contribution to the existing technology. The computer software product is stored in a storage medium (such as read-only memory (Read Only) Memory, ROM) / RAM, disk, optical Disk), including several instructions to enable a terminal device (which may be a mobile phone, a computer, a server, or a network device, etc.) to execute the method described in each embodiment of the present application.

The above are only the preferred embodiments of the present application, and do not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made by the description and drawings of this application, or directly or indirectly used in other related technical fields , The same reason is included in the scope of patent protection of this application.

Claims (20)

A review method based on artificial intelligence, characterized in that the method includes: Receiving a review request triggered by the user, obtaining the service ID of the service submitted by the user and the service corresponding to the service review according to the review request, the service includes at least one approval process node, and each approval process node corresponds to at least one review square; According to the business number, search the history approval information of the business corresponding to the business number in the business approval schedule; According to the information to be reviewed and the historical approval information, determine the review start node and review party; Submit the information to be reviewed to the review initiation node, so that the reviewing party can review the information to be reviewed. The method according to claim 1, wherein the historical approval information includes the review status made by each reviewer corresponding to each approval process node on the current approval process node; The step of determining the review start node and reviewer based on the information to be reviewed and the historical approval information includes: According to the review status of each review party corresponding to each review process node to the current review process node, the review status is processed review process node, and a candidate set of review start nodes is obtained. The processed review process node is review state Approval process node for rejected status or approved status; Select an approval process node from the candidate set of review start nodes as the review start node according to the rejection information corresponding to the review process and the approval process node whose review status is the rejection state; Based on the information to be reviewed and the description information of the audit work responsible for each reviewer corresponding to the review start node, at least one reviewer is selected as the reviewer from each reviewer corresponding to the review start node. The method according to claim 1, wherein before the step of determining a review start node and a review party based on the information to be reviewed and the historical approval information, the method further comprises: Determine that the information to be reviewed meets the review requirements; Wherein, the step of determining that the information to be reviewed meets the review conditions includes: Obtain the reason for rejection corresponding to the business according to the business number; Obtain the audit rules corresponding to the reasons for rejection based on the reasons for rejection; It is determined whether the information to be reviewed meets the requirements of the review rules, and if the information to be reviewed meets the requirements of the review rules, it is determined that the information to be reviewed meets the review conditions. The method of claim 3, wherein the information to be reviewed is text information; The step of determining whether the information to be reviewed satisfies the provisions of the review rules includes: Use keyword extraction to extract keywords from the text information; According to the review rules, select keywords containing the content specified in the review rules; Determine whether the proportion of keywords containing the content specified in the review rule is not lower than the first threshold specified in the review rule; If the proportion of keywords containing the content specified in the review rule is not lower than the first threshold, the information to be reviewed meets the requirements of the review rule. The method according to claim 3, wherein the information to be reviewed is a user's face image; The step of determining whether the information to be reviewed satisfies the provisions of the review rules includes: Extracting the facial feature contour information of the user from the face image, and determining the image matrix corresponding to the facial feature contour information according to the facial feature contour information; Comparing the image matrix with the reference template matrix specified in the review rules to determine the similarity value between the image matrix and the reference template matrix; Determine whether the similarity value is not lower than the second threshold specified in the review rule; If the similarity value is not lower than the second threshold, the information to be reviewed meets the requirements of the review rules. The method according to claim 5, wherein the step of extracting the user's facial features contour information from the face image includes: Performing grayscale processing on the face image to obtain a grayscale image; Perform binary processing and smooth denoising on the grayscale image in sequence to obtain a binary image with interference information removed; Using the edge detection method, the user's facial features contour information is extracted from the binary image. The method of claim 3, wherein after the step of determining whether the information to be reviewed meets the requirements of the review rule, the method further comprises: If the information to be reviewed does not meet the requirements of the review rules, the review rules are reported to the user so that the user can modify the information to be reviewed according to the review rules. A review device based on artificial intelligence, characterized in that the device includes: The obtaining module is used for receiving a review request triggered by a user, and obtaining the service ID of the service submitted by the user and the service corresponding to the service review according to the request, the service includes at least one approval process node, and each approval process The node corresponds to at least one reviewer; A search module, configured to search the history approval information of the business corresponding to the business number in the business approval schedule according to the business number; The determining module is used to determine the starting node and the reviewing party based on the information to be reviewed and the historical approval information; The submitting module is used to submit the information to be reviewed to the review initiation node, so that the reviewing party can review the information to be reviewed. An artificial intelligence-based review device, characterized in that the device includes: a memory, a processor, and an artificial intelligence-based review program stored on the memory and runable on the processor, the manual-based review program The intelligent review program is configured to implement the following steps: Receiving a review request triggered by the user, obtaining the service ID of the service submitted by the user and the service corresponding to the service review according to the review request, the service includes at least one approval process node, and each approval process node corresponds to at least one review square; According to the business number, search the history approval information of the business corresponding to the business number in the business approval schedule; According to the information to be reviewed and the historical approval information, determine the review start node and review party; Submit the information to be reviewed to the review initiation node, so that the reviewing party can review the information to be reviewed. The artificial intelligence-based review device according to claim 9, wherein the historical approval information includes the review status of each review party corresponding to each approval process node on the current approval process node; The step of determining the review start node and reviewer based on the information to be reviewed and the historical approval information includes: According to the review status of each review party corresponding to each review process node to the current review process node, the review status is processed review process node, and a candidate set of review start nodes is obtained, and the processed review process node is review state Approval process node for rejected status or approved status; Select an approval process node from the candidate set of review start nodes as the review start node according to the rejection information corresponding to the review process and the approval process node whose review status is the rejection state; Based on the information to be reviewed and the description information of the audit work responsible for each reviewer corresponding to the review start node, at least one reviewer is selected as the reviewer from each reviewer corresponding to the review start node. The artificial intelligence-based review device according to claim 9, wherein before the step of determining a review start node and a review party based on the information to be reviewed and the historical approval information, the method further includes: Determine that the information to be reviewed meets the review requirements; Wherein, the step of determining that the information to be reviewed meets the review conditions includes: Obtain the reason for rejection corresponding to the business according to the business number; Obtain the audit rules corresponding to the reasons for rejection based on the reasons for rejection; It is determined whether the information to be reviewed meets the requirements of the review rules, and if the information to be reviewed meets the requirements of the review rules, it is determined that the information to be reviewed meets the review conditions. The artificial intelligence-based review device according to claim 11, wherein the information to be reviewed is text information; The step of determining whether the information to be reviewed satisfies the provisions of the review rules includes: Use keyword extraction to extract keywords from the text information; According to the review rules, select keywords containing the content specified in the review rules; Determine whether the proportion of keywords containing the content specified in the review rule is not lower than the first threshold specified in the review rule; If the proportion of keywords containing the content specified in the review rule is not lower than the first threshold, the information to be reviewed meets the requirements of the review rule. The artificial intelligence-based review device according to claim 11, wherein the information to be reviewed is a user's face image; The step of determining whether the information to be reviewed satisfies the provisions of the review rules includes: Extracting the facial feature contour information of the user from the face image, and determining the image matrix corresponding to the facial feature contour information according to the facial feature contour information; Comparing the image matrix with the reference template matrix specified in the review rules to determine the similarity value between the image matrix and the reference template matrix; Determine whether the similarity value is not lower than the second threshold specified in the review rule; If the similarity value is not lower than the second threshold, the information to be reviewed meets the requirements of the review rules. The artificial intelligence-based review device according to claim 13, wherein the step of extracting the user's facial features contour information from the face image includes: Performing grayscale processing on the face image to obtain a grayscale image; Perform binary processing and smooth denoising on the grayscale image in sequence to obtain a binary image with interference information removed; Using the edge detection method, the user's facial features contour information is extracted from the binary image. The artificial intelligence-based review device according to claim 11, wherein after the step of determining whether the information to be reviewed meets the requirements of the review rule, the method further includes: If the information to be reviewed does not meet the requirements of the review rules, the review rules are reported to the user so that the user can modify the information to be reviewed according to the review rules. A storage medium, characterized in that a review program based on artificial intelligence is stored on the storage medium, and when the review program based on artificial intelligence is executed by a processor, the following steps are realized: Receiving a review request triggered by the user, obtaining the service ID of the service submitted by the user and the service corresponding to the service review according to the review request, the service includes at least one approval process node, and each approval process node corresponds to at least one review square; According to the business number, search the history approval information of the business corresponding to the business number in the business approval schedule; According to the information to be reviewed and the historical approval information, determine the review start node and review party; Submit the information to be reviewed to the review initiation node, so that the reviewing party can review the information to be reviewed. The storage medium according to claim 16, wherein the historical approval information includes the review status made by each reviewer corresponding to each approval process node on the current approval process node; The step of determining the review start node and reviewer based on the information to be reviewed and the historical approval information includes: According to the review status of each review party corresponding to each review process node to the current review process node, the review status is processed review process node, and a candidate set of review start nodes is obtained, and the processed review process node is review state Approval process node for rejected status or approved status; Select an approval process node from the candidate set of review start nodes as the review start node according to the rejection information corresponding to the review process and the approval process node whose review status is the rejection state; Based on the information to be reviewed and the description information of the audit work responsible for each reviewer corresponding to the review start node, at least one reviewer is selected as the reviewer from each reviewer corresponding to the review start node. The storage medium according to claim 16, wherein before the step of determining a review start node and a review party based on the information to be reviewed and the historical approval information, the method further includes: Determine that the information to be reviewed meets the review requirements; Wherein, the step of determining that the information to be reviewed meets the review conditions includes: Obtain the reason for rejection corresponding to the business according to the business number; Obtain the audit rules corresponding to the reasons for rejection based on the reasons for rejection; It is determined whether the information to be reviewed meets the requirements of the review rules, and if the information to be reviewed meets the requirements of the review rules, it is determined that the information to be reviewed meets the review conditions. The storage medium according to claim 18, wherein the information to be reviewed is text information; The step of determining whether the information to be reviewed satisfies the provisions of the review rules includes: Use keyword extraction to extract keywords from the text information; According to the review rules, select keywords containing the content specified in the review rules; Determine whether the proportion of keywords containing the content specified in the review rule is not lower than the first threshold specified in the review rule; If the proportion of keywords containing the content specified in the review rule is not lower than the first threshold, the information to be reviewed meets the requirements of the review rule. The storage medium according to claim 18, wherein the information to be reviewed is a user's face image; The step of determining whether the information to be reviewed satisfies the provisions of the review rules includes: Extracting the facial feature contour information of the user from the face image, and determining the image matrix corresponding to the facial feature contour information according to the facial feature contour information; Comparing the image matrix with the reference template matrix specified in the review rules to determine the similarity value between the image matrix and the reference template matrix; Determine whether the similarity value is not lower than the second threshold specified in the review rule; If the similarity value is not lower than the second threshold, the information to be reviewed meets the requirements of the review rules.