CN114332886A

CN114332886A - Object detection device, method, electronic device, storage medium, and program product

Info

Publication number: CN114332886A
Application number: CN202111635905.2A
Authority: CN
Inventors: 刘昊骋; 宫健; 岳洪达; 韩光耀; 冯博豪; 王红玉
Original assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Current assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Priority date: 2021-12-29
Filing date: 2021-12-29
Publication date: 2022-04-12

Abstract

The present disclosure provides an object detection device, an object detection method, an electronic device, a storage medium, and a program product, and relates to the field of computer technologies, in particular to the field of artificial intelligence and deep learning technologies. The object detection equipment comprises a classification module, an identification module, a receipt numerical value detection result determination module, a receipt effectiveness determination module and a receipt detection result determination module, wherein the classification module is used for classifying a plurality of receipt objects to obtain a receipt category; the identification module is used for respectively identifying the numerical value information of a plurality of bill objects to obtain bill numerical values; the receipt value detection result determining module is used for obtaining a receipt value detection result according to the receipt value incidence relations of the plurality of receipt objects; the document effectiveness determining module is used for detecting the effectiveness of a plurality of document objects to obtain the document effectiveness; and the document detection result determining module is used for determining a document detection result according to the document value detection result and the document effectiveness.

Description

Object detection device, method, electronic device, storage medium, and program product

Technical Field

The present disclosure relates to the field of computer technologies, and in particular, to an object detection device, an object detection method, an electronic device, a storage medium, and a program product.

Background

Document objects are of a wide variety, and detecting document objects is therefore a cumbersome task. How to improve the detection efficiency of the document object is an urgent problem to be solved.

Disclosure of Invention

The present disclosure provides an object detection apparatus, method, electronic apparatus, storage medium, and program product.

According to an aspect of the present disclosure, there is provided an object detecting apparatus including: the system comprises a classification module, an identification module, a document numerical value detection result determination module, a document effectiveness determination module and a document detection result determination module, wherein the classification module is used for classifying a plurality of document objects to obtain document types; the identification module is used for respectively identifying the numerical value information of the plurality of bill objects to obtain bill numerical values; the bill numerical value detection result determining module is used for obtaining a bill numerical value detection result according to the bill numerical value incidence relations of the plurality of bill objects; the document validity determining module is used for detecting the validity of the plurality of document objects to obtain the validity of the documents; and the bill detection result determining module is used for determining a bill detection result according to the bill value detection result and the bill validity.

According to another aspect of the present disclosure, there is provided an object detection method including: classifying the plurality of document objects to obtain document categories; respectively identifying the numerical value information of the plurality of bill objects to obtain bill numerical values; obtaining a receipt value detection result according to the receipt value incidence relations of the plurality of receipt objects; carrying out validity detection on the plurality of document objects to obtain document validity; and determining a bill detection result according to the bill value detection result and the bill validity.

According to another aspect of the present disclosure, there is provided an electronic device including: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of the embodiments of the present disclosure.

According to another aspect of the present disclosure, there is provided a non-transitory computer readable storage medium having stored thereon computer instructions for causing the computer to perform the method according to the embodiments of the present disclosure.

According to another aspect of the present disclosure, a computer program product is provided, comprising a computer program which, when executed by a processor, implements a method according to embodiments of the present disclosure.

It should be understood that the statements in this section do not necessarily identify key or critical features of the embodiments of the present disclosure, nor do they limit the scope of the present disclosure. Other features of the present disclosure will become apparent from the following description.

Drawings

The drawings are included to provide a better understanding of the present solution and are not to be construed as limiting the present disclosure. Wherein:

fig. 1 schematically shows a system architecture diagram of an object detection method, device according to an embodiment of the present disclosure;

FIG. 2 schematically shows a flow diagram of an object detection method according to an embodiment of the present disclosure;

FIG. 3 schematically shows a schematic diagram of an object detection method according to an embodiment of the present disclosure;

fig. 4 schematically shows a schematic diagram of an object detection method according to another embodiment of the present disclosure;

fig. 5 schematically shows a schematic diagram of an object detection method according to a further embodiment of the present disclosure;

fig. 6 schematically shows a schematic diagram of an object detection method according to a further embodiment of the present disclosure;

fig. 7 schematically shows a schematic diagram of an object detection method according to a further embodiment of the present disclosure;

FIG. 8 is a schematic diagram illustrating an application of the object detection method to a disease document object claim scene according to an embodiment of the disclosure;

fig. 9 schematically shows a schematic view of an object detection device according to an embodiment of the present disclosure;

fig. 10 schematically shows a block diagram of an electronic device that may implement the object detection method of an embodiment of the present disclosure.

Detailed Description

Exemplary embodiments of the present disclosure are described below with reference to the accompanying drawings, in which various details of the embodiments of the disclosure are included to assist understanding, and which are to be considered as merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

Fig. 1 schematically shows a system architecture of an object detection method and apparatus according to an embodiment of the present disclosure. It should be noted that fig. 1 is only an example of a system architecture to which the embodiments of the present disclosure may be applied to help those skilled in the art understand the technical content of the present disclosure, and does not mean that the embodiments of the present disclosure may not be applied to other devices, systems, environments or scenarios.

As shown in fig. 1, a system architecture 100 according to this embodiment may include

clients

101A, 101B, 101C, a network 102, and a server 103. Network 102 is the medium used to provide communication links between

clients

101A, 101B, 101C and server 103. Network 102 may include various connection types, such as wired, wireless communication links, or fiber optic cables, to name a few.

A user may use a

client

101A, 101B, 101C to interact with a server 103 over a network 102 to receive or send messages, etc. Various messaging client applications, such as navigation-type applications, web browser applications, search-type applications, instant messaging tools, mailbox clients, social platform software, etc. (examples only) may be installed on the

clients

101A, 101B, 101C.

The

clients

101A, 101B, 101C may be various electronic devices having display screens and supporting web browsing, including but not limited to smartphones, tablets, laptop and desktop computers, and the like. The

clients

101A, 101B, 101C of the disclosed embodiments may run applications, for example.

The server 103 may be a server that provides various services, such as a background management server (for example only) that provides support for websites browsed by users using the

clients

101A, 101B, 101C. The background management server may analyze and perform other processing on the received data such as the user request, and feed back a processing result (e.g., a webpage, information, or data obtained or generated according to the user request) to the client. In addition, the server 103 may also be a cloud server, that is, the server 103 has a cloud computing function.

It should be noted that the object detection method provided by the embodiment of the present disclosure may be executed by the server 103. Accordingly, the object detection device provided by the embodiment of the present disclosure may be disposed in the server 103. The object detection method provided by the embodiments of the present disclosure may also be performed by a server or a server cluster that is different from the server 103 and is capable of communicating with the

clients

101A, 101B, 101C and/or the server 103. Accordingly, the object detection device provided by the embodiment of the present disclosure may also be disposed in a server or a server cluster different from the server 103 and capable of communicating with the

clients

101A, 101B, 101C and/or the server 103.

In one example, the server 103 may obtain document objects from the

clients

101A, 101B, 101C over the network 102.

It should be understood that the number of clients, networks, and servers in FIG. 1 is merely illustrative. There may be any number of clients, networks, and servers, as desired for an implementation.

It should be noted that in the technical solution of the present disclosure, the processes of collecting, storing, using, processing, transmitting, providing, disclosing and the like of the personal information of the user are all in accordance with the regulations of the relevant laws and regulations, and do not violate the customs of the public order.

Fig. 2 illustrates a flow chart of an object detection method according to an embodiment of the present disclosure.

As shown in fig. 2, the object detection method 200 according to an embodiment of the present disclosure includes operations S210 to S250.

In operation S210, a plurality of document objects are classified to obtain a document class.

The object detection method of the embodiment of the present disclosure will be described below by taking the disease insurance claim as an example. According to the requirements of insurance claims, the document objects required to be provided by the insurance claims comprise: the medical record, the prescription, the payment bill, the check bill, the invoice and the like can be used as different bill categories.

For example, the image information of the document object may be obtained in a scanning manner, and the image information of the document object is input into the document classification model to obtain the document classification. The document classification models may include a VGG model, a ResNet model, and a DenseNet model. VGG, Visual Geometry Group, among others. ResNet, i.e., a deep residual network. DenseNet, i.e., a densely connected network.

In operation S220, the numerical information of the plurality of document objects is respectively recognized to obtain document numerical values.

In the process of insurance claim settlement, numerical information of a document object needs to be detected, and the numerical information can be understood as information representing money. For example, the payment amount of the bill object, which is a payment bill, is the bill value.

It should be understood that there are some document objects for the document category from which the document value cannot be derived. For example, since a document object of the document type of a medical record does not include the amount information, a document value indicating the amount information cannot be obtained from the medical record.

Illustratively, the value information of the document object can be recognized through an Optical Character Recognition technology, i.e. Optical Character Recognition, abbreviated as OCR, to obtain the document value. Specifically, the worker can mark the position of the numerical value information of the document object in the image information of the document object, and then recognize the text information of the position through an optical character recognition technology to obtain the document numerical value.

In operation S230, a document value detection result is obtained according to the document value association relationship of the plurality of document objects.

In the insurance claim settlement process, mutual authentication relationship exists among the bill objects, for example, a plurality of bill objects have a bill value association relationship, and the bill value of each bill object can be authenticated according to the bill value association relationship to obtain a bill value detection result.

In operation S240, validity detection is performed on the plurality of document objects to obtain document validity.

Document validity may be understood as the document object having validity for the current insurance claim. For example, for the disease insurance claim for hypertension, the related document object of car accident operation does not have document validity.

In operation S250, a document detection result is determined according to the document value detection result and the document validity.

The object detection method of the embodiment of the disclosure can classify the bill objects and determine the bill types, for example, bill objects of different bill types can be used for verifying whether the bill objects of insurance claims are complete. And identifying and detecting the document value, wherein the obtained document value detection result can indicate whether the document value of the document object is correct or not. In addition, validity detection is carried out on a plurality of document objects, and the object detection result is accurately determined by integrating three aspects of whether the document objects are complete, whether the numerical values of the document objects are correct and whether the document objects are valid, so that the document objects can be automatically detected, manpower resources are saved, and the object detection efficiency is improved.

It should be noted that the operation labels S210 to S250 only represent different operations, and do not represent the operation sequence, and the document detection result of the embodiment of the present disclosure can also be obtained by adjusting the operation sequence.

Fig. 3 schematically shows a schematic diagram of an object detection method according to an embodiment of the present disclosure.

Operations S310 to S350 of the object detection method 300 according to the embodiment of the present disclosure correspond to the above-described operations S210 to S250, respectively. That is, in operation S310, a plurality of document objects 301 are classified to obtain a document class 302. In operation S320, the value information of the plurality of document objects 301 is respectively identified to obtain the document values 303. In operation S330, a document value detection result 304 is obtained according to the document value association relationship of the plurality of document objects. In operation S340, validity detection is performed on the plurality of document objects 301 to obtain document validity 305. In operation S350, a document detection result 306 is determined according to the document value detection result 304 and the document validity 305.

Fig. 4 schematically shows a schematic diagram of an object detection method according to another embodiment of the present disclosure.

As shown in fig. 4, according to an object detection method 400 of another embodiment of the present disclosure, a document object may include a total document object and a sub-document object, and numerical information of the total document object includes a total document value, and the object detection method 400 may further include operations S460 to S470.

In operation S460, the total document value 403 of the total document object 401 is verified, and a total document value verification result 407 is obtained.

In operation S470, the total document target detection result 408 is determined according to the document detection result 406 of the total document object 401 and the document total value verification result 407.

The document detection result 406 of the total document object 401 is obtained through operations S420 to S450. Operations S420 to S450 correspond to operations S220 to S250 of the above-described embodiment, respectively. It should be understood that the sheet value detection result 404 obtained in operation S430 is a detection result of the total sheet value 403, the sheet validity 405 obtained in operation S440 is a validity of the total sheet object 401, and the sheet detection result 406 obtained in operation S450 is a sheet detection result of the total sheet object 401.

The total document object comprises a total document value, and the document value of the sub-document object can have an inclusion and contained relationship with the total document value. For example, in the case of disease insurance settlement, the total amount invoice includes amount information of each check sheet, payment sheet, and the like. The total invoice can be understood as a total document object, the total amount of the total invoice can be understood as a total document value, each check sheet and each payment sheet can be understood as a sub-document object, and the total document value comprises the document value of the sub-document object.

The total document value of the total document object can be used as the reference data of the insurance claim amount, and the object detection method of the embodiment of the disclosure can improve the object detection accuracy by checking the total document value of the total document object, and avoid the occurrence of the conditions of object detection errors and the like caused by unreasonable total document value of the total document object or counterfeit document object.

Fig. 5 schematically shows a schematic diagram of validity detection of a plurality of document objects in an object detection method according to yet another embodiment of the present disclosure.

According to another embodiment of the present disclosure, the following embodiments may be used to implement validity detection on a plurality of document objects in an object detection method, so as to obtain a specific example of document validity.

As shown in fig. 5, in operation S541, medical information of the plurality of sheet objects 501 is identified, and sheet object knowledge map elements 502 of the plurality of sheet objects 501 are obtained.

For example, in the case of disease insurance claims, the document object includes medical information related to the disease. The medical information may include: disease characteristics, disease symptoms, treatment methods, treatment medications, precautions, and the like. Document object knowledge map elements may include disease characteristics, disease symptoms, treatment methods, treatment medications, notes, and the like.

In operation S542, form object knowledge map element coverage 504 is determined from the form object knowledge map elements 502 and the medical knowledge map 503.

A medical knowledge-map may be understood as a continuously updated, medically relevant, complete knowledge-map. Document object knowledge graph element coverage may be expressed, for example, as: a ratio of a number of document object knowledge map elements appearing in the medical knowledge graph to a number of document object knowledge map elements. The document object knowledge map element coverage may represent the correlation of the document object with the medical knowledge map.

In operation S543, the document object knowledge graph element coverage 504 is compared with the validity threshold 505 to obtain the document validity 506.

Illustratively, the validity threshold may be a threshold range greater than 70%.

According to the object detection method, the knowledge graph is introduced, the effectiveness of the document is evaluated according to the comparison result of the coverage rate of the knowledge graph of the document object and the effectiveness threshold value, and the accurate document effectiveness can be still obtained under the condition that the document effectiveness is evaluated without manual intervention.

Fig. 6 schematically shows a schematic diagram of checking a total document value of a total document object in an object detection method according to another embodiment of the present disclosure.

According to another embodiment of the present disclosure, the following embodiments may be used to verify the total value of the total document object in the object detection method, so as to obtain a specific example of the total value verification result of the document. The document object knowledge map elements may include document object entities.

As shown in fig. 6, in operation S661, a sub-sheet object entity of the sub-sheet object is determined.

Still taking the above-mentioned total document object includes the total amount invoice, and the sub-document object includes the checking order and the payment order as an example, the sub-document object entity may include: disease characteristics, therapeutic drugs, and methods of treatment.

In operation S662, an initial word vector of the sub-document object entity is determined.

A word vector may be understood as a vector representation of a word in a natural language processing task. The names of the bill entities expressed in the word vector mode can represent the similar relation among different sub-bill object entities, and can also contain more information, and each dimension of the initial word vector has a specific meaning, so that the initial word vector can better express the sub-bill object entities.

In operation S663, each sub-document object is independently pre-trained to obtain a target word vector of each sub-document object.

Pre-training can be understood as: the process of using as much training data as possible to extract as many common features as possible from the training data, thereby reducing the learning burden of the model on a specific task.

The target word vector may be understood as a vector representation of a sub-document object entity that incorporates the text semantics of the current sub-document object.

In operation S664, a document standard total value 601 of the total document object is determined according to the target word vector.

In operation S665, a total document value check result 607 is determined according to the total document standard value 601 and the total document value 603.

For example, the total document value verification result can be determined according to the ratio of the standard total document value, the absolute value of the difference between the total document value and the standard total document value to the total document value. For example, when the ratio of the absolute value of the difference between the standard total value and the total value of the document to the total value of the document is less than 20%, the total value of the document is determined to be a pass.

According to the object detection method, the single-data object and the knowledge graph are combined, the sub-document object entity can represent medical information of the sub-document object, the medical information has specific meanings in the medical field, and the specific meanings need to be determined by combining text information of the current sub-document object, so that the target word vector obtained through pre-training can more accurately represent the name of the sub-document object entity fusing the specific meanings in the medical field. Meanwhile, the semantics expressed by the text information of different sub-document objects are different, so that the object detection method of the embodiment of the disclosure can not mix the semantics expressed by the text information of different sub-document objects by independently pre-training each sub-document object, so that the target word vector determined by different sub-document objects is more accurate, and the accuracy of the document total value verification result and the object detection efficiency are improved.

FIG. 6 is a diagram illustrating an exemplary determination of a total document value verification result based on a plurality of sub-document objects. The sub-document object entity 11 and the sub-document object entity 1N are determined according to the sub-document object 1, and the sub-document object entity N1 and the sub-document object entity Nn are determined according to the sub-document object N. And determining a corresponding initial word vector 111, an initial word vector 1nn, an initial word vector N11 and an initial word vector Nnn according to each sub-document object entity. And each sub-document object is independently pre-trained to obtain a corresponding target word vector 1v and a corresponding target word vector Nv.

Fig. 7 schematically shows a schematic diagram of obtaining a target word vector of each sub-document object in an object detection method according to yet another embodiment of the present disclosure.

According to another embodiment of the present disclosure, the following embodiments may be used to implement independent pre-training on each sub-document object in the object detection method, so as to obtain a specific example of the target word vector of each sub-document object.

As shown in FIG. 7, in operation S731, a disease characteristic 702 of the document object 701 is determined.

In operation S732, the target pre-training model 704 is determined according to the mapping relationship between the disease features 702 and the pre-training model 703.

In operation S733, the initial word vector 705 of the sub-document object is pre-trained using the target pre-training model 704 to obtain a target word vector 706.

The disease feature can be understood as the name of a disease, the initial word vector is a vector representation form of a sub-document object entity, and the sub-document object entity reflects a certain aspect of the disease, for example, a certain sub-document object entity is a disease symptom. There may be: different disease characteristics show the same disease symptoms, while different disease characteristics have different treatment methods and treatment medicines, and the text information of the corresponding sub-document objects has great difference. Therefore, according to the object detection method of the embodiment of the disclosure, when the corresponding target word vector is determined according to the initial sub-vector, the target pre-training model is determined according to the mapping relation between the disease characteristics and the pre-training model, and the initial word vector of the sub-document object is pre-trained by using the target pre-training model, so that a more accurate target word vector can be obtained, and the object detection efficiency is improved.

According to the object detection method in the embodiment of the present disclosure, in the above embodiment, determining the standard total number of documents of the total document object according to the target word vector may include: and inputting the target word vector into a document numerical value prediction model to obtain a standard total numerical value of the document.

The document numerical value prediction model can be understood as a prediction model which is trained by a large number of samples in advance, the input of the model is a target word vector, and the output of the model is a document standard total numerical value. Illustratively, the document value prediction model may include a logistic regression model.

The object detection method of the embodiment of the invention comprises the steps that when a plurality of sub-document objects are included, target word vectors of the sub-document objects can be spliced, and the word vectors obtained after the target word vectors are spliced are input into a document numerical value prediction model to obtain a standard total numerical value of the document.

The object detection method of the embodiment of the invention can automatically and accurately obtain the standard total value of the bill according to the bill value prediction model.

Exemplarily, according to the object detection method of the embodiment of the present disclosure, in the above embodiment, obtaining the bill value detection result according to the bill value association relationship of the plurality of bill objects may include: determining the sum of the bill values of at least one sub-bill object according to the incidence relation between the total bill object and the sub-bill objects to obtain the sum of the sub-bill values; comparing the total value of the document with the total sub-value of the document to obtain a document value comparison result; and determining a bill value detection result according to the bill value comparison result.

For example, when the total document object is a total invoice and the sub-document object is an inspection sheet, the association relationship between the total document object and the sub-document object includes: and the incidence relation between the sum of the bill values of all the sub-bill objects and the total bill value of the total bill object. For example, when the document value comparison result is that the total document value is less than or equal to the total sub-value of the documents, the document value detection result is determined to be passed.

In insurance claim settlement, the claim settlement conditions include: the total document value of the total document object is smaller than or equal to the total document sub-value, and the object detection method can detect the document value and determine the document value detection result by combining with the specific scene of insurance claim settlement.

It should be understood that when there is only one sub-sheet object, the above-mentioned "sum of sheet values of at least one sub-sheet object" should be understood as the sheet value of the only sub-sheet object.

According to the object detection method of the embodiment of the present disclosure, the total document value verification result may include pass and fail, and the object detection method may further include: and sending a manual auditing request in response to the result of the total document numerical value verification being failed.

Under some abnormal conditions, the total document value verification result may be wrong, and in order to ensure the accuracy of object detection, the total document value verification result may be reviewed manually.

The object detection method according to the embodiment of the present disclosure may further include: and generating a bill object detection report.

Illustratively, the document object detection report may include: the method comprises the steps of bill type, bill validity, bill value detection results and bill total value verification results.

Fig. 8 schematically shows a schematic diagram of an object detection method applied to a disease document object claim scene according to an embodiment of the present disclosure.

As shown in fig. 8, when a user uses a claim paper document object, the claim paper document object may be scanned into image information by a scanner, and the document object in an image format is input into a document classification model to obtain document categories, where the document categories may include: invoices, medical records, prescriptions, examination application forms, examination reports, and other document objects. And identifying the bill value of the bill object of each bill category, and detecting the bill value to obtain a bill value detection result. For example, if the document value detection result is failed, the document value problem of the claim document object of the user is indicated. After the receipt numerical value detection is carried out, the receipt effectiveness detection can be carried out, namely, the receipt object knowledge map elements relevant to the medicine are extracted firstly, and the receipt effectiveness is obtained according to the receipt object knowledge map element coverage rate of the receipt object knowledge map elements relative to the medical knowledge map. For example, if the document validity is failed, it indicates that the document validity problem occurs in the claim document object of the user. After the receipt validity detection is carried out, the total receipt value can be checked, namely, a corresponding initial word vector is determined according to a sub-receipt object entity of a sub-receipt object, a target word vector can be obtained after the initial word vector is pre-trained, and the sub-receipt object entity is a receipt object knowledge map element of the sub-receipt object. The target word vector is used as the input of the bill numerical prediction model, the bill standard total numerical value of the total bill object can be obtained through the prediction of the bill numerical prediction model, and the bill standard total numerical value and the bill total numerical value are compared to obtain the bill total numerical value verification result. For example, when the total document value verification result is that the total document value is not passed, the problem of the total document value of the claim document object of the surface user occurs, when at least one of the problem of the total document value, the problem of document validity and the problem of the total document value occurs, a mail feedback can be sent to related personnel, and when any one of the problems does not occur, the system can enter the claim settlement system for settlement.

Operations S810 to S840 and S860 shown in fig. 8 correspond to operations S210 to S240 and S460 of the foregoing embodiment, respectively, and are not described herein again, it should be understood that, in the embodiment shown in fig. 8, each operation is performed sequentially, and after the operation S840 is performed on the basis of the operation S830, a document detection result may be obtained according to a document value detection result obtained in the operation S830 and document validity obtained in the operation S840. Similarly, after operation S860, the total-document-target detection result may be obtained according to the document detection result of the total document object and the document total-value verification result.

According to an embodiment of the present disclosure, the present disclosure also provides an object detection apparatus.

As shown in fig. 9, an object detecting apparatus 900 according to an embodiment of the present disclosure includes: the system comprises a classification module 910, an identification module 920, a document value detection result determination module 930, a document validity determination module 940 and a document detection result determination module 950.

The classification module 910 is configured to classify a plurality of document objects to obtain document categories. In an embodiment, the classification module 910 may be configured to perform the operation S210, which is not described herein again.

The identifying module 920 is configured to identify the numerical value information of the plurality of document objects, respectively, to obtain document numerical values. In an embodiment, the identifying module 920 may be configured to perform the operation S220, which is not described herein again.

The document value detection result determining module 930 is configured to obtain a document value detection result according to the document value association relationships of the plurality of document objects. In an embodiment, the document value detection result determining module 930 may be configured to perform the operation S230, which is not described herein again.

The document validity determining module 940 is configured to perform validity detection on the plurality of document objects to obtain document validity. In an embodiment, the document validity determining module 940 may be configured to perform the operation S240, which is not described herein again.

And the document detection result determining module 950 is configured to determine a document detection result according to the document value detection result and the document validity. In an embodiment, the document detection result determining module 950 may be configured to perform the operation S250, which is not described herein again.

According to the object detection device of the embodiment of the present disclosure, the document object includes a total document object and a sub-document object, the value information of the total document object includes a total document value, and the object detection device may further include: the total document value verifying module and the total document target detection result determining module.

And the total document value checking module is used for checking the total document value of the total document object to obtain a total document value checking result.

And the total document target detection result determining module is used for determining a total document target detection result according to the document detection result of the total document object and the document total value verification result.

According to the object detection device of the embodiment of the disclosure, the document validity determination module may include: the document object knowledge map element determining sub-module, the document object knowledge map element coverage rate determining sub-module and the document effectiveness determining sub-module.

And the document object knowledge map element determining submodule is used for identifying the medical information of the plurality of document objects and obtaining document object knowledge map elements of the plurality of document objects.

And the document object knowledge map element coverage rate determining submodule is used for determining the document object knowledge map element coverage rate according to the document object knowledge map elements and the medical knowledge map.

And the document validity determining submodule is used for comparing the coverage rate of the document object knowledge graph elements with the validity threshold value to obtain the document validity.

According to the object detection device of the embodiment of the disclosure, the document object knowledge graph element may include a document object entity, and the document total value verification module may include: the system comprises a sub-document object entity determining sub-module, an initial word vector determining sub-module, a target word vector determining sub-module, a document standard total value determining sub-module and a document total value checking result determining sub-module.

And the sub-document object entity determining submodule is used for determining the sub-document object entity of the sub-document object.

And the initial word vector determining submodule is used for determining the initial word vector of the sub-document object entity.

And the target word vector determining submodule is used for independently pre-training each sub-document object to obtain the target word vector of each sub-document object.

And the document standard total value determining submodule is used for determining the document standard total value of the total document object according to the target word vector.

And the total document numerical value verification result determining submodule is used for determining a total document numerical value verification result according to the standard total document numerical value and the total document numerical value.

According to the object detection device of the embodiment of the present disclosure, the target word vector determination sub-module may include: the system comprises a disease characteristic determining unit, a target pre-training model determining unit and a target word vector determining unit.

And the disease characteristic determining unit is used for determining the disease characteristics of the document object.

And the target pre-training model determining unit is used for determining a target pre-training model according to the mapping relation between the disease characteristics and the pre-training model.

And the target word vector determining unit is used for pre-training the initial word vector of the sub-document object by using the target pre-training model to obtain the target word vector.

According to the object detection device of the embodiment of the disclosure, the document standard total value determining submodule may include: and the bill standard total value determining unit is used for inputting the target word vector into the bill numerical value prediction model to obtain the bill standard total value.

According to the object detection device of the embodiment of the disclosure, the document value detection result determining module may include: the system comprises a document sub-value sum determining submodule, a document value comparison result determining submodule and a document value detection result determining submodule.

And the document sub-value sum determining submodule is used for determining the sum of the document values of at least one sub-document object according to the incidence relation between the total document object and the sub-document objects to obtain the document sub-value sum.

And the document value comparison result determining submodule is used for comparing the total document value with the total document sub-value to obtain a document value comparison result.

And the document value detection result determining submodule is used for determining a document value detection result according to the document value comparison result.

According to the object detection device of the embodiment of the present disclosure, the bill value verification result may include pass and fail, and the object detection device may further include: and the manual auditing request sending module is used for responding to the failure of the total document numerical value checking result and sending a manual auditing request.

The object detection apparatus according to the embodiment of the present disclosure may further include: and the object detection report generating module is used for generating a document object detection report.

It should be understood that the embodiments of the apparatus part of the present disclosure are the same as or similar to the embodiments of the method part of the present disclosure, and the technical problems to be solved and the technical effects to be achieved are also the same as or similar to each other, and the detailed description of the present disclosure is omitted.

The present disclosure also provides an electronic device, a readable storage medium, and a computer program product according to embodiments of the present disclosure.

FIG. 10 illustrates a schematic block diagram of an example electronic device 1000 that can be used to implement embodiments of the present disclosure. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit implementations of the disclosure described and/or claimed herein.

As shown in fig. 10, the apparatus 1000 includes a computing unit 1001 that can perform various appropriate actions and processes according to a computer program stored in a Read Only Memory (ROM)1002 or a computer program loaded from a storage unit 1008 into a Random Access Memory (RAM) 1003. In the RAM 1003, various programs and data necessary for the operation of the device 1000 can also be stored. The calculation unit 1001, the ROM 1002, and the RAM 1003 are connected to each other by a bus 1004. An input/output (I/O) interface 1005 is also connected to bus 1004.

A number of components in device 1000 are connected to I/O interface 1005, including: an input unit 1006 such as a keyboard, a mouse, and the like; an output unit 1007 such as various types of displays, speakers, and the like; a storage unit 1008 such as a magnetic disk, an optical disk, or the like; and a communication unit 1009 such as a network card, a modem, a wireless communication transceiver, or the like. The communication unit 1009 allows the device 1000 to exchange information/data with other devices through a computer network such as the internet and/or various telecommunication networks.

Computing unit 1001 may be a variety of general and/or special purpose processing components with processing and computing capabilities. Some examples of the computing unit 1001 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various dedicated Artificial Intelligence (AI) computing chips, various computing units running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, and so forth. The calculation unit 1001 executes the respective methods and processes described above, such as the object detection method. For example, in some embodiments, the object detection method may be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 1008. In some embodiments, part or all of the computer program may be loaded and/or installed onto device 1000 via ROM 1002 and/or communications unit 1009. When the computer program is loaded into RAM 1003 and executed by the computing unit 1001, one or more steps of the object detection method described above may be performed. Alternatively, in other embodiments, the computing unit 1001 may be configured to perform the object detection method in any other suitable way (e.g., by means of firmware).

Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuitry, Field Programmable Gate Arrays (FPGAs), Application Specific Integrated Circuits (ASICs), Application Specific Standard Products (ASSPs), system on a chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, receiving data and instructions from, and transmitting data and instructions to, a storage system, at least one input device, and at least one output device.

Program code for implementing the methods of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the program codes, when executed by the processor or controller, cause the functions/operations specified in the flowchart and/or block diagram to be performed. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package partly on the machine and partly on a remote machine or entirely on the remote machine or server.

In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic, speech, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), Wide Area Networks (WANs), and the Internet.

The computer system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.

It should be understood that various forms of the flows shown above may be used, with steps reordered, added, or deleted. For example, the steps described in the present disclosure may be executed in parallel, sequentially, or in different orders, as long as the desired results of the technical solutions disclosed in the present disclosure can be achieved, and the present disclosure is not limited herein.

The above detailed description should not be construed as limiting the scope of the disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made in accordance with design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present disclosure should be included in the scope of protection of the present disclosure.

Claims

1. An object detecting apparatus comprising:

the classification module is used for classifying the plurality of document objects to obtain document categories;

the identification module is used for respectively identifying the numerical value information of the plurality of bill objects to obtain bill numerical values;

the bill numerical value detection result determining module is used for obtaining a bill numerical value detection result according to the bill numerical value incidence relations of the plurality of bill objects;

the document validity determining module is used for detecting the validity of the plurality of document objects to obtain the validity of the documents; and

and the bill detection result determining module is used for determining a bill detection result according to the bill value detection result and the bill validity.

2. The apparatus of claim 1, wherein the sheet object comprises an overall sheet object and a sub-sheet object, the value information of the overall sheet object comprises an overall sheet value, the object detection apparatus further comprising:

the total document value checking module is used for checking the total document value of the total document object to obtain a total document value checking result; and

and the total bill target detection result determining module is used for determining a total bill target detection result according to the bill detection result of the total bill object and the bill total value verification result.

3. The apparatus of claim 2, wherein the document validity determination module comprises:

the document object knowledge map element determining submodule is used for identifying the medical information of the plurality of document objects and obtaining document object knowledge map elements of the plurality of document objects;

the document object knowledge map element coverage rate determining sub-module is used for determining the document object knowledge map element coverage rate according to the document object knowledge map elements and the medical knowledge map; and

and the document validity determining submodule is used for comparing the coverage rate of the document object knowledge graph elements with a validity threshold value to obtain the document validity.

4. The apparatus of claim 3, wherein the document object knowledge graph element comprises a document object entity, the document total value verification module comprising:

the sub-document object entity determining sub-module is used for determining a sub-document object entity of the sub-document object;

the initial word vector determining submodule is used for determining the initial word vector of the sub-document object entity;

the target word vector determining sub-module is used for independently pre-training each sub-document object to obtain a target word vector of each sub-document object;

the bill standard total value determining submodule is used for determining a bill standard total value of the total bill object according to the target word vector; and

and the total document numerical value verification result determining submodule is used for determining the total document numerical value verification result according to the standard total document numerical value and the total document numerical value.

5. The apparatus of claim 4, wherein the target word vector determination sub-module comprises:

the disease characteristic determining unit is used for determining the disease characteristics of the document object;

the target pre-training model determining unit is used for determining a target pre-training model according to the mapping relation between the disease characteristics and the pre-training model; and

6. The apparatus of claim 4, wherein the document standard total value determination submodule comprises:

and the bill standard total value determining unit is used for inputting the target word vector into a bill numerical value prediction model to obtain the bill standard total value.

7. The apparatus of any of claims 2 to 6, wherein the document value detection result determination module comprises:

the bill sub-value sum determining submodule is used for determining the sum of the bill values of at least one sub-bill object according to the incidence relation between the total bill object and the sub-bill objects to obtain the sum of the bill sub-values;

the bill numerical value comparison result determining submodule is used for comparing the total bill numerical value with the sum of the sub-bill numerical values to obtain a bill numerical value comparison result; and

and the receipt value detection result determining submodule is used for determining the receipt value detection result according to the receipt value comparison result.

8. The apparatus of any of claims 2 to 6, the document value verification result comprising a pass and a fail, the object detection apparatus further comprising:

and the manual auditing request sending module is used for responding to the result that the total document numerical value verification result is not passed and sending a manual auditing request.

9. The apparatus of any of claims 1 to 6, further comprising:

and the object detection report generating module is used for generating a document object detection report.

10. An object detection method, comprising:

classifying the plurality of document objects to obtain document categories;

respectively identifying the numerical value information of the plurality of bill objects to obtain bill numerical values;

obtaining a receipt value detection result according to the receipt value incidence relations of the plurality of receipt objects;

carrying out validity detection on the plurality of document objects to obtain document validity; and

and determining a bill detection result according to the bill value detection result and the bill validity.

11. The method of claim 10, wherein the sheet object includes an overall sheet object and sub-sheet objects, the value information for the overall sheet object includes an overall sheet value, the object detection method further comprising:

verifying the total document value of the total document object to obtain a total document value verification result; and

and determining a total bill target detection result according to the bill detection result of the total bill object and the bill total value verification result.

12. The method of claim 11, wherein the detecting validity of the plurality of document objects, obtaining document validity comprises:

identifying medical information of the plurality of document objects to obtain document object knowledge map elements of the plurality of document objects;

determining the coverage rate of the document object knowledge map elements according to the document object knowledge map elements and the medical knowledge map; and

and comparing the coverage rate of the document object knowledge graph element with an effectiveness threshold value to obtain the effectiveness of the document.

13. The method of claim 12, wherein the document object knowledge graph elements include document object entities, and the verifying the total document value of the total document object to obtain a total document value verification result comprises:

determining a sub-document object entity of the sub-document object;

determining an initial word vector of the sub-document object entity;

pre-training each sub-document object independently to obtain a target word vector of each sub-document object;

determining a bill standard total value of the total bill object according to the target word vector; and

and determining the total bill value verification result according to the total standard bill value and the total bill value.

14. The method of claim 13, wherein the pre-training each of the sub-document objects independently to obtain a target word vector for each of the sub-document objects comprises:

determining disease characteristics of the document object;

determining a target pre-training model according to the mapping relation between the disease characteristics and the pre-training model; and

and pre-training the initial word vector of the sub-document object by using the target pre-training model to obtain the target word vector.

15. The method of claim 13, wherein the determining a document standard total value for the total document object from the target word vector comprises:

and inputting the target word vector into a bill numerical value prediction model to obtain the standard total numerical value of the bill.

16. The method according to any one of claims 11 to 15, wherein the obtaining a document value detection result according to the document value incidence relations of the plurality of document objects comprises:

determining the sum of the bill numerical values of at least one sub-bill object according to the incidence relation between the total bill object and the sub-bill objects to obtain the sum of the sub-numerical values of the bills;

comparing the total value of the bill with the total sum of the sub-values of the bill to obtain a bill value comparison result; and

and determining the bill value detection result according to the bill value comparison result.

17. The method of any of claims 11 to 15, the document value verification result comprising a pass and a fail, the object detection method further comprising:

and sending a manual auditing request in response to the result of the total document numerical value verification being failed.

18. The method of any of claims 10 to 15, further comprising:

and generating a bill object detection report.

19. An electronic device, comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 10-18.

20. A non-transitory computer readable storage medium having stored thereon computer instructions for causing the computer to perform the method of any one of claims 10-18.

21. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any one of claims 10-18.