A method of processing medical tickets and a processing system

Abstract

The application provides a medical bill processing method and a processing system. The processing method comprises the following steps: outputting the position information corresponding to the text and the text box on the bill through image data by OCR character recognition, and then extracting the required entity elements from each text box based on the labels defined based on prior knowledge through entity extraction; and the entity elements are subjected to structured processing by using a structured processing method, so as to achieve a result consistent with the details of the bill. The processing method can efficiently process and audit the medical bill, has high accuracy, avoids the need for manual recording in the prior art, and avoids the problem of gradually increasing error rate and increasing the difficulty of auditing.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of medical bill processing, in particular, to a medical bill processing method and system. BACKGROUND

[0002] At present, the process of claim settlement records in the insurance industry generally relies on manual input and manual checking and correcting. This traditional mode requires a large amount of manpower and material resources. For a long time, a large number of claim settlement records need to be processed, and the continuous low efficiency of manual work and the increase of business volume form a great contradiction. In addition, different personnel have different understandings of the business, and it is difficult to form a standardized claim settlement record standard. Therefore, it is necessary to take innovative intelligent means to improve and solve the known pain points in the insurance industry, and it is imperative to build an efficient and accurate intelligent claim settlement input system.

[0003] In the flood of structured data and artificial intelligence, it is an urgent task to use innovative technology to improve the work efficiency of the traditional insurance industry, improve customer experience, reduce false report rate, and standardize claim settlement.

[0004] Therefore, the present application is proposed. SUMMARY

[0005] Therefore, the present application discloses a medical bill processing method and system. The present application constructs an intelligent claim settlement record system by using multi-modal deep learning and structured post-processing, thereby efficiently processing and auditing medical bills with high accuracy, avoiding the need for manual input in the prior art, and avoiding the problem of increasing error rate and increasing the difficulty of auditing.

[0006] Specifically, the present application is realized by the following technical solutions:

[0007] In a first aspect, the present application discloses a medical bill processing method, comprising the following steps:

[0008] The image data is output through OCR text recognition to obtain the position information of the text and the text box on the bill, and then the required entity elements are extracted from each text box based on the labels defined by prior knowledge through entity extraction.

[0009] The entity elements are structured by using a structured processing method to achieve a result consistent with the details of the bill.

[0010] In a second aspect, the present application discloses a medical bill processing system, comprising:

[0011] The extraction module is used to output the text on the bill and the position information corresponding to the text box through OCR text recognition, and then extract the required entity elements from each text box based on the labels defined based on prior knowledge through entity extraction.

[0012] The structured processing module is used to perform structured processing on the entity elements by using a structured processing method to achieve a result consistent with the details of the bill.

[0013] In a third aspect, the present application discloses a computer readable storage medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the steps of the medical bill processing method of the first aspect.

[0014] In a fourth aspect, the present application discloses a computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor implements the steps of the medical bill processing method of the first aspect when executing the program.

[0015] In summary, the medical bill processing method and processing system proposed by the present application construct an intelligent claim recording system by using multi-modal deep learning and structured post-processing, thereby efficiently processing and auditing medical bills with high accuracy, avoiding the need for manual recording in the prior art, and avoiding the problem of gradually increasing error rate and increasing the difficulty of auditing. BRIEF DESCRIPTION OF DRAWINGS

[0016] Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The accompanying drawings are included only to illustrate preferred embodiments and are not intended to limit the present application. Moreover, the same reference numerals are used throughout the drawings to designate the same components. In the drawings:

[0017] Figure 1 A flowchart of the medical bill processing method provided by the embodiment of the present application is shown in the figure;

[0018] Figure 2 A flowchart of the Y threshold method provided by the embodiment of the present application is shown in the figure;

[0019] Figure 3 A flowchart of the common part ratio judgment method provided by the embodiment of the present application is shown in the figure;

[0020] Figure 4 A flowchart of the vertical group slope splicing method provided by the embodiment of the present application is shown in the figure;

[0021] Figure 5 A flowchart of a computer device provided by the embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0022] The exemplary embodiments will be described in detail herein with reference to the attached drawings. The description herein relates to the drawings, which show by way of illustration, rather than limitation, the exemplary embodiments. Unless otherwise expressly provided, the same numbers in different drawings can represent the same or similar elements. The following exemplary embodiments described are not meant to be an all-inclusive description of all aspects of the disclosure. Rather, they are merely examples that can be practiced with apparatus and methods consistent with the disclosure as detailed in the appended claims.

[0023] The terminology used in the disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. As used in the description of the disclosure and the appended claims, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will also be understood that the term "and / or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.

[0024] It will be understood that, although the terms first, second, third, etc. can be used herein to describe various information, these terms are not intended to denote a temporal or chronological order. Rather, these terms are used only as a shorthand notation to distinguish one item from another. For example, in the absence of a disclosure to the contrary, a first item can be termed a second item, and similarly, a second item can be termed a first item. Depending on the context, the word "if' as used herein can be interpreted to mean "when" or "in response to determining".

[0025] A medical bill processing method is disclosed, referring to Figure 1 as shown, comprising the following steps:

[0026] The image data is output through OCR text recognition to output the text on the bill and the position information corresponding to the text box, and then the entity extraction is performed to extract the required entity elements from each text box based on the labels defined by prior knowledge.

[0027] The entity elements are structured by using a structured processing method to achieve a result consistent with the details of the bill.

[0028] The image data mentioned above is divided into internal data of insurance companies (historical data related to all customers, such as invoices, case books, claim application forms, medical detail lists, etc.) and external data (such as third-party output provided by physical examination companies, customer medical data provided by cooperative hospitals, etc.).

[0029] The data structure is unstructured data, and all are images uploaded by insurance claim users.

[0030] The structured processing process is a key process for structuring medical invoice details. The result text, position information and label three elements input by the character recognition and entity extraction layer cannot form structured business requirement detail information. Therefore, the structured processing method needs to provide three methods to restore the unstructured result to the same as the invoice details. The three methods here are Y threshold method, public part ratio judgment method, and vertical group slope splicing method. The most suitable way can be selected for processing according to the actual situation.

[0031] The Y threshold method includes the following steps, which can be seen from the specific Figure 2

[0032] Separate the part containing the label from the entity element, and take out the item containing the medical detail cost item;

[0033] Set a threshold range according to the Y value of the fixed position information to obtain all other items in the same row;

[0034] Sort all items according to X coordinates, and traverse all non-cost items until the next cost item is encountered.

[0035] The Y threshold method starts from the position information and label of the three elements output by the text recognition + entity extraction layer, finds the head of a certain detail information (that is, the name of the cost item), and then uses a fixed Y threshold to find the remaining other information belonging to this detail, to form a complete detail information for output.

[0036] The public part ratio judgment method includes the following steps, which can be seen from the specific Figure 3

[0037] Separate the part containing the label from the entity element, and take out the item containing the medical detail cost item, and take the item as A item and another item as B item;

[0038] Calculate the height and specific coordinates of the left side of A item and B item text, and then calculate the length ratio of the public part;

[0039] When the length ratio of the public part is greater than 50%, it is determined that B item and A item are in the same row, and B item is changed to A item, and then a new B item is searched;

[0040] Sort all items in the same row according to the ascending order of X coordinates, and then cut according to the cost item label. The several items in each interval after cutting form the detailed medical details for output.

[0041] ​​The public part ratio judgment method uses position information to find the public part ratio of the vertical edges of the two text boxes, and uses chain relationship to judge the entries in the same row one by one. In the special case of a bill image sample that is inclined and folded, the method can achieve better results than the Y threshold method.

[0042] The vertical group slope splicing method includes the following steps, which are specifically shown in Figure 4

[0043] Separating the part containing the label from the entity element, taking out the entry containing the medical detailed expense item, and dividing the entry into the same number of expense item groups and amount groups;

[0044] Assembling the group serial numbers of the expense item groups and the amount groups to form a detailed entry containing a head and a tail;

[0045] Calculating the slope K value of the expense item and the amount item composed of the head and the tail in the detailed entry;

[0046] Calculating the K value of other results not included in the expense item groups and the amount groups, and including the results meeting the conditions in the entry.

[0047] Preferably, the judgment method of whether the number of the expense item groups and the amount groups is the same includes:

[0048] Dividing the expense item groups into several small groups according to the X threshold value to obtain a small group number A;

[0049] Dividing the amount groups into several small groups according to the X threshold value to obtain a small group number B;

[0050] Judging whether A is equal to B, otherwise discarding the method, and if yes, using the method for subsequent steps.

[0051] The vertical group slope splicing method is slightly more complex than the previous two methods, and can only be used under the condition that the number of expense item groups and the number of amount groups are the same, otherwise it will lead to a messy splicing result. Therefore, when using this method, it is necessary to judge the number of expense item groups and the number of amount groups. This method first finds and forms the expense item groups and the amount groups from the entity extraction results according to the label, then assembles the same group serial numbers in the expense item groups and the amount groups to form several predefined detailed entries containing only a head and a tail, and calculates the slope K value of the expense item and the amount item composed of the head and the tail in these detailed entries. Next, in other results not included in the expense item groups and the amount item groups, it is calculated whether the K value of a certain predefined detailed entry is the same, and if yes, it is included in this detailed entry, at which time this detailed entry is spliced. Repeat until all predefined detailed entries are inserted, the expense item groups and the amount groups have no available items, and finally output the result.

[0052] ​In summary, in the above three methods, different complex bill images are dealt with, and each method has its own advantages. For example, the Y threshold method has low time complexity and fast implementation for regular invoices, and the result is accurate; for invoices with slight tilt, the public part ratio judgment method can accurately find and splice the items in the same row under the adverse conditions of tilt to form a complete result output; for invoices with severe tilt and folding, the vertical group slope splicing method can solve the splicing problem to a certain extent, but the last method has relatively harsh conditions for use, and when the entity extraction result is abnormal, this method cannot be used.

[0053] The final output result needs to judge whether there is an abnormal situation in the results of the three methods, and the judgment basis is as follows: whether there is only one item in the detailed item, and how many there are in the result returned by the three methods. After two judgments, the result of the method that does not contain only one item in the detailed item is returned, or the result of the method with the least number of only one item in the detailed item in the three methods.

[0054] In addition, the present application also provides a medical bill processing system, specifically comprising:

[0055] The extraction module is used to output the position information of the text and the text box on the bill through OCR text recognition of the image data, and then extract the required entity elements from each text box based on the label defined based on prior knowledge through entity extraction;

[0056] The structured processing module is used to structure the entity elements by using a structured processing method to achieve a result consistent with the details of the bill.

[0057] The processing system is mainly composed of the above modules, and the construction of the system well solves the problem of improving the work efficiency of recording.

[0058] In specific implementation, each of the above modules can be implemented as an independent entity, or can be combined as the same or several entities, and the specific implementation of each unit can be referred to the method embodiment above, which will not be repeated here.

[0059] Figure 5 A structural diagram of a computer device is disclosed in the present application. Referring to Figure 5 The computer device 400 includes at least a memory 402 and a processor 401; the memory 402 is connected with the processor through a communication bus 403, and is used to store computer instructions executable by the processor 401; the processor 401 is used to read computer instructions from the memory 402 to implement the steps of the medical bill processing method described in any of the above embodiments.

[0060] For the above apparatus embodiments, since they basically correspond to the method embodiments, the relevant parts can be seen from the parts of the method embodiments. The apparatus embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the present disclosure according to actual needs. Those skilled in the art can understand and implement it without creative labor.

[0061] Computer readable media suitable for storing computer program instructions and data include all forms of non-volatile memory, media and memory devices, including by way of example semiconductor memory devices, e.g., EPROM, EEPROM, and flash memory devices; magnetic disks, e.g., internal hard disks or removable disks; magneto-optical disks; and CD-ROM and DVD-ROM disks. The processor and the memory can be supplemented by, or incorporated in, special purpose logic circuitry.

[0062] It should also be understood that, although the terms "first" and "second" can be used herein to describe various elements, these elements should not be limited by these terms. These terms are generally only used to distinguish one element from another. Thus, a first element discussed below could be termed a second element without departing from the scope of the present application. Similarly, it should be understood that, although the terms "including", "includes" and "included" are used herein, these terms are used in the sense of "including but not limited to", and are not used in the sense of "consist(ing) only of".

[0063] Similarly, while operations are depicted in the drawings in a particular order, this should not be understood as requiring such order, nor that all illustrated operations be performed, to achieve desirable results. In certain circumstances, multitasking and parallel processing can be advantageous. Moreover, the separation of various system modules and components in the embodiments described above should not be understood as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software product or packaged into multiple software products.

[0064] Thus, particular embodiments of the subject matter have been described. Other embodiments are within the scope of the following claims. In some cases, actions recited in the claims can be performed in a different order and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve the desired results. In certain implementations, multitasking and parallel processing can be advantageous.

[0065] The above-described embodiments are merely possible implementations of the present disclosure, and do not limit the present disclosure. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present disclosure shall be included in the scope of the present disclosure.

Claims

1. A method for processing medical invoices, characterized in that, Includes the following steps: The image data is processed by OCR text recognition to output the text on the ticket and the location information corresponding to the text boxes. Then, entity extraction is performed to extract the required entity elements from each text box based on the labels defined by prior knowledge. The entity elements are structured using a structured processing method to achieve a result consistent with the details on the ticket. The structured processing method includes the Y-threshold method, the common part ratio judgment method, and the vertical group slope splicing method, and one or a combination of several methods may be selected. The method for determining the proportion of common parts includes the following steps: Separate the part containing the label from the entity element, extract the item containing the medical expense details, take this item as item A, and take another item as item B; Calculate the height and specific coordinates of the left side of the text in items A and B, and then calculate the ratio of the length of the common part; When the proportion of the length of the common part is greater than 50%, it is determined that item B and item A are in the same row, and item B is changed to item A, and then the search for new item B continues; Sort all entries in the same row in ascending order of X coordinate, then divide them according to the cost item label, and output the detailed medical information composed of several entries in each divided interval. The Y-threshold method includes the following steps: Separate the portion containing the label from the entity element and extract the entry containing the medical expense details; To retrieve all other entries in the same row, a threshold range is set based on the Y value of the fixed location information. Sort all entries according to their X coordinates, and iterate through all non-cost entries until the next cost item is encountered. The vertical group slope splicing method includes the following steps: Separate the part containing the label from the entity element, and take out the items containing medical expense details and divide them into expense item groups and amount groups with the same number of items; The group numbers of the expense items and the amount groups are combined to form a detailed entry including a beginning and an end; Calculate the slope K value formed by the expense item and the amount item composed of the beginning and end of the detailed entry; Calculate the K value of other expense items not included in the expense group and the amount group, and include those that meet the criteria in the entry; The processing of expense item groups and amount groups includes: The cost item group is divided into several subgroups according to the X threshold, resulting in a group size of A. The amount group is divided into several subgroups according to the X threshold, resulting in a subgroup number of B; Determine if A is equal to B; otherwise, discard this method; otherwise, use this method for subsequent steps.

2. A processing system employing the medical invoice processing method of claim 1, characterized in that, include: Extraction module: It is used to output the text on the ticket and the corresponding position information of the text box through OCR text recognition of image data, and then extract the required entity elements from each text box based on the labels defined by prior knowledge. Structured processing module: Used to perform structured processing on the entity elements using structured processing methods to achieve a result consistent with the details on the ticket.

3. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed, it implements the steps of the medical invoice processing method of claim 1.

4. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the medical invoice processing method as described in claim 1.

Images