Loss assessment and claim settlement method, device and equipment and storage medium

Abstract

The application relates to artificial intelligence and provides a loss determination and claim settlement method, device, equipment and storage medium. The method obtains an initial label, initial label information and a claim form image of a loss determination object, performs image recognition on the claim form image, determines keyword label information according to a keyword label generated after the image recognition and a text keyword, calculates a general coefficient and a characteristic coefficient of a preset loss item based on the initial label, the initial label information, the keyword label, the keyword label information, a preset general coefficient model and a preset characteristic coefficient model, respectively adjusts and corrects the general coefficient based on the characteristic coefficient and a preset loss determination standard library to obtain a loss determination coefficient of the preset loss item, and calculates claim settlement data of the loss determination object according to a plurality of preset loss items and a plurality of loss determination coefficients, so that the accuracy of human injury loss determination can be improved. In addition, the application also relates to blockchain technology, and the claim settlement data can be stored in the blockchain.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, in particular to a loss assessment and claim settlement method, device, equipment and storage medium. BACKGROUND

[0002] In the current loss assessment scheme of human injury, since loss assessment of human injury needs to analyze a large amount of data, the process of loss assessment of human injury is complex, which makes the accuracy of loss assessment of human injury poor. Therefore, how to improve the accuracy of loss assessment of human injury has become a problem to be solved. SUMMARY

[0003] In view of the above, it is necessary to provide a loss assessment and claim settlement method, device, equipment and storage medium, which can solve the technical problem of how to improve the accuracy of loss assessment of human injury.

[0004] In one aspect, the present application provides a loss assessment and claim settlement method, which comprises:

[0005] An initial label of a loss assessment object, initial label information corresponding to the initial label, and a claim form image are obtained, and a pre-trained character recognition model is used to recognize the claim form image to obtain a keyword label and a text keyword of the claim form image. The keyword label information corresponding to the keyword label is generated based on the label position and the text position corresponding to the keyword label and the text keyword in the claim form image. The initial label, the initial label information, the keyword label and the keyword label information are input into a preset general coefficient model and a preset feature coefficient model respectively to calculate a general coefficient of a preset loss item and a feature coefficient of the preset loss item. The general coefficient is adjusted based on the feature coefficient to obtain a target coefficient of the preset loss item, and the target coefficient is corrected based on a preset loss assessment standard library to obtain a loss assessment coefficient of the preset loss item. The claim settlement data of the loss assessment object is calculated according to a plurality of preset loss items and the loss assessment coefficient corresponding to each preset loss item.

[0006] According to an optional embodiment of the present application, the pre-trained character recognition model comprises a recurrent sequence generation network and a decoding network, and the use of the pre-trained character recognition model to recognize the claim form image to obtain the keyword label and the text keyword of the claim form image comprises:

[0007] Position the position of the text information in the claim form image based on the pixel value of the pixel point in the claim form image, obtain the text position, perform feature extraction on the text information based on the text position, obtain a feature sequence, input the feature sequence into the recurrent sequence generation network, obtain a recurrent sequence, decode the recurrent sequence based on the decoding network, obtain the text information of each claim image, determine a first preset keyword in the text information as the keyword label, and determine a second preset keyword in the text information as the text keyword.

[0008] According to an optional embodiment of the present application, the generating the keyword label information corresponding to the keyword label based on the keyword label and the label position and the text position corresponding to the text keyword in the claim form image comprises:

[0009] According to the label position and the text position, it is detected whether the label rectangular region corresponding to the keyword label intersects with the text rectangular region corresponding to the text keyword. If the label rectangular region intersects with the text rectangular region, the intersection region of the label rectangular region and each intersecting text rectangular region is calculated, the first area ratio of the intersection region on the label rectangular region is calculated, and the second area ratio of the intersection region on the text rectangular region is calculated. The text rectangular region with both the first area ratio and the second area ratio greater than a preset threshold value is selected as a target rectangular region, and the keyword label information corresponding to the keyword label is generated based on the number of target rectangular regions and the text keyword corresponding to the target rectangular region.

[0010] According to an optional embodiment of the present application, the generating the keyword label information corresponding to the keyword label based on the keyword label and the label position and the text position corresponding to the text keyword in the claim form image comprises:

[0011] If the number of target rectangular regions is single, the text keyword corresponding to the target rectangular region is determined as the keyword label information corresponding to the keyword label, or if the number of target rectangular regions is multiple, a weighted sum operation is performed on each first area ratio and the corresponding second area ratio to obtain a final score value of each target rectangular region, and the text keyword corresponding to the target rectangular region with the maximum final score value is selected as the keyword label information corresponding to the keyword label.

[0012] According to an optional embodiment of the present application, before the initial label, the initial label information, the keyword label and the keyword label information are input into a preset feature coefficient model, the method further comprises:

[0013] Obtaining a preset adversarial neural network, obtaining position information to which the loss-determining object belongs and training data corresponding to the position information, selecting preset feature data from the training data based on a preset feature keyword, training the preset adversarial neural network based on the preset feature data, and obtaining a preset feature coefficient model.

[0014] According to an optional embodiment of the present application, the universal coefficient model comprises a plurality of preset universal impairment grades and a universal corresponding relationship corresponding to each preset universal impairment grade, and the universal coefficient comprises a universal impairment coefficient, a universal base number and a universal month number. The universal coefficient model is used to calculate the universal coefficient of the preset loss item, which comprises:

[0015] The impairment label is selected from the initial label and the keyword label based on a preset impairment keyword, and a value corresponding to the impairment label in the claim form image is taken as an initial impairment grade of the loss-determining object. The universal proportion coefficient corresponding to the preset universal impairment grade identical to the initial impairment grade in the universal coefficient model is determined as the universal impairment coefficient. The universal corresponding relationship corresponding to the preset universal impairment grade in the universal coefficient model is determined as a target corresponding relationship. The initial base number label corresponding to the universal base number label in the target corresponding relationship and the initial month number label corresponding to the universal month number label in the target corresponding relationship are identified from the initial label and the keyword label. The universal base number is calculated based on an initial base value of the initial base number label and a universal base number calculation mode in the target corresponding relationship. The universal month number is calculated based on an initial month value of the initial month number label and a universal month number calculation mode in the target corresponding relationship.

[0016] According to an optional embodiment of the present application, the universal coefficient is adjusted based on the feature coefficient to obtain a target coefficient of the preset loss item, which comprises:

[0017] If the feature coefficient does not correspond to each universal coefficient, the universal coefficient is determined as the target coefficient. Alternatively, if at least one feature coefficient corresponds to the universal coefficient, the universal coefficient is replaced by the corresponding feature coefficient, and the replaced universal coefficient is determined as the target coefficient.

[0018] On the other hand, the present application also proposes a loss-determining claim settlement device, which comprises:

[0019] The acquisition unit is configured to acquire an initial label of a loss settlement object, initial label information corresponding to the initial label, and a claim form image, and use a pre-trained character recognition model to recognize the claim form image to obtain a keyword label of the claim form image and a text keyword; the generation unit is configured to generate keyword label information corresponding to the keyword label based on a label position and a text position corresponding to the keyword label and the text keyword in the claim form image; the input unit is configured to input the initial label, the initial label information, the keyword label, and the keyword label information into a preset general coefficient model and a preset feature coefficient model respectively to calculate a general coefficient of a preset loss item and a feature coefficient of the preset loss item; the adjustment unit is configured to adjust the general coefficient based on the feature coefficient to obtain a target coefficient of the preset loss item, and correct the target coefficient based on a preset loss settlement standard library to obtain a loss settlement coefficient of the preset loss item; and the calculation unit is configured to calculate claim data of the loss settlement object according to a plurality of preset loss items and the loss settlement coefficient corresponding to each preset loss item.

[0020] In another aspect, the present application also provides an electronic device, which comprises:

[0021] a memory configured to store computer-readable instructions; and

[0022] a processor configured to execute the computer-readable instructions stored in the memory to implement the loss settlement and claim method.

[0023] In another aspect, the present application also provides a computer-readable storage medium, which stores computer-readable instructions, and the computer-readable instructions are executed by a processor in an electronic device to implement the loss settlement and claim method.

[0024] It can be seen from the above technical solutions that, by identifying the claim form image, the information of the loss-determining object can be comprehensively obtained, the text keywords and numerical keywords in the claim form image information can be extracted, redundant information can be removed, the keyword label information corresponding to the keyword label is determined based on the label position and the text position, the keyword label and the keyword label information are corresponded, which is beneficial to quickly calculating the general coefficient and the characteristic coefficient based on the corresponding relationship between the keyword label and the keyword label information, adjusting the general coefficient based on the characteristic coefficient, obtaining the target coefficient corresponding to the preset loss-determining item, and improving the accuracy of the characteristic coefficient model trained by using the characteristic data, so that the characteristic coefficient output by the characteristic coefficient model is more accurate, the general coefficient is adjusted based on the characteristic coefficient, the accuracy of the target coefficient can be improved, and the target coefficient is corrected based on the preset loss-determining standard library, the target coefficient can be corrected based on the accurate loss-determining result output by the loss-determining standard library, and the accuracy of the claim data can be ensured. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 is a flowchart of a preferred embodiment of the loss-determining claim method of the present application.

[0026] Figure 2 is a functional module diagram of a preferred embodiment of the loss-determining claim method of the present application.

[0027] Figure 3 is a structural schematic diagram of an electronic device for implementing a preferred embodiment of the loss-determining claim method of the present application. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be described in detail below in combination with the drawings and specific embodiments.

[0029] As shown in Figure 1 is a flowchart of a preferred embodiment of the loss-determining claim method of the present application. The order of steps in the flowchart can be changed according to different needs, and some steps can be omitted.

[0030] The loss-determining claim method can acquire and process related data based on artificial intelligence technology. Artificial intelligence (AI) is the theory, method, technology and application system for simulating, extending and expanding human intelligence by using digital computers or machines controlled by digital computers to perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0031] The basic technologies of artificial intelligence generally include technologies such as sensors, special artificial intelligence chips, cloud computing, distributed storage, big data processing, operation / interaction systems, mechatronics, etc. The software technologies of artificial intelligence mainly include computer vision technology, robot technology, biometric identification technology, speech processing technology, natural language processing technology, and machine learning / deep learning, etc.

[0032] The loss settlement and claim method is applied to one or more electronic devices, which is a device capable of automatically performing numerical calculation and / or information processing according to pre-set or stored computer readable instructions. The hardware thereof includes but is not limited to microprocessors, application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0033] The electronic device can be any electronic product capable of human-computer interaction with the user, such as a personal computer, a tablet computer, a smart phone, a personal digital assistant (PDA), a game console, an interactive Internet Protocol Television (IPTV), a smart wearable device, etc.

[0034] The electronic device can include network devices and / or user devices. The network devices include, but are not limited to, single network electronic devices, groups of electronic devices composed of multiple network electronic devices, or clouds composed of a large number of hosts or network electronic devices based on cloud computing.

[0035] The network in which the electronic device is located includes, but is not limited to, the Internet, a wide area network, a metropolitan area network, a local area network, a virtual private network (VPN), etc.

[0036] 101, obtaining an initial label of a loss settlement object, initial label information corresponding to the initial label, and a claim form image, and using a pre-trained character recognition model to recognize the claim form image to obtain a keyword label and a text keyword of the claim form image.

[0037] In at least one embodiment of the present application, the initial label refers to partial text data of the loss object, the initial label information refers to keyword information corresponding to the initial label, and the claim form image refers to an image containing information required when the loss object claims. For example, the initial label includes, but is not limited to, the unified plan area where the loss object is located, and the social security card number of the loss object. When the initial label is the unified plan area where the loss object is located, the initial label information corresponding to the initial label may be, for example, Shanghai. The claim form image contains previous disability identification information of the loss object, labor contract performance, household nature, gender, and date of birth, and the like.

[0038] In at least one embodiment of the present application, the electronic device provides an information input interface for the user to input the initial label information and import the claim form image through the information input interface. The information input interface includes an input label, an input button corresponding to the input label, an information input field corresponding to the input label, an image upload button, and an image import field corresponding to the image upload button. The electronic device receives a first trigger instruction generated by the user pressing the input button, and receives the initial label information input by the user in the information input field corresponding to the input label according to the first trigger instruction. The electronic device receives a second trigger instruction generated by the user pressing the image upload button, and receives the claim form image uploaded by the user in the image import field according to the second trigger instruction.

[0039] In at least one embodiment of the present application, the electronic device uses a pre-trained text recognition model to recognize the claim form image to obtain the keyword label and the text keyword of the claim form image.

[0040] In at least one embodiment of the present application, the text recognition model refers to a model for recognizing text information and digital information in the claim form image.

[0041] In at least one embodiment of the present application, the keyword label includes a first keyword label and a second keyword label, the text keyword includes a literal keyword and a numerical keyword, the first keyword label refers to a label corresponding to the literal keyword, the second keyword label refers to a label corresponding to the numerical keyword, the literal keyword refers to literal specific information of a loss object corresponding to the first keyword label, and the numerical keyword refers to numerical specific information of the loss object corresponding to the second keyword label. For example, when the first keyword label is gender, the literal keyword is male, when the second keyword label is birth date, the numerical keyword is 1987-12-28, when the second keyword label is age, the numerical keyword is 35, and so on.

[0042] In at least one embodiment of the present application, the text recognition model includes a recurrent sequence generation network and a decoding network, and the electronic device uses a pre-trained text recognition model to recognize the claim form image to obtain keyword labels and text keywords of the claim form image, including:

[0043] The electronic device locates the position of the text information in the claim form image based on the pixel value of the pixel point in the claim form image to obtain a literal position, further, the electronic device extracts features of the text information based on the literal position to obtain a feature sequence, and still further, the electronic device inputs the feature sequence into the recurrent sequence generation network to obtain a recurrent sequence, and still further, the electronic device decodes the recurrent sequence based on the decoding network to obtain literal information of each loss image, and still further, the electronic device determines a first preset keyword in the literal information as the keyword label and a second preset keyword in the literal information as the text keyword.

[0044] In the present embodiment, the first preset keyword includes, but is not limited to, gender, occupation category, age, birth date, and monthly average salary, and so on, and the second preset keyword includes, but is not limited to, male, female, technical personnel, 35, 1987-12-28, and 8000, and so on.

[0045] Specifically, the electronic device extracts features of the text information based on a feature extraction network to obtain the feature sequence, the feature extraction network can be a resnet50 network, the recurrent sequence generation network can be a bidirectional long short-term memory neural network, and the decoding network can be a neural network-based time series classification algorithm (Connectionist temporal classification, CTC).

[0046] Specifically, the electronic device locates a position of text information in the claim form image based on pixel values of pixel points in the claim form image, to obtain a character position, including:

[0047] The electronic device obtains pixel positions of all pixel points in the claim form image, and traverses pixel values of all pixel points in the claim form image. Further, the electronic device determines a plurality of pixel points corresponding to a pixel value greater than a preset value as text information, and determines pixel positions of the corresponding plurality of pixel points as the character position.

[0048] Specifically, the recurrent sequence generation network includes a convolution layer and a pooling layer. The electronic device inputs the feature sequence into the recurrent sequence generation network to obtain a recurrent sequence, including:

[0049] The electronic device performs segmentation processing on the feature sequence to obtain a plurality of segmented features. Further, the electronic device performs convolution operation on the plurality of segmented features based on the convolution layer to obtain a convolution result. Further, the electronic device performs pooling processing on the convolution result based on the pooling layer to obtain a feature vector corresponding to each segmented feature. Further, the electronic device splices a plurality of the feature vectors to obtain the recurrent sequence.

[0050] Wherein, the segmentation processing process is prior art, which is not described herein.

[0051] Specifically, the electronic device decodes the recurrent sequence based on the decoding network to obtain text information of each loss-determining image, including:

[0052] The electronic device obtains a feature order corresponding to each feature vector in the recurrent sequence, and performs mapping processing on each feature vector in the recurrent sequence based on a pre-constructed dictionary to obtain a feature character corresponding to each feature vector. The electronic device splices a plurality of the feature characters according to the feature order to obtain text information of each loss-determining image.

[0053] Wherein, the pre-constructed dictionary stores a corresponding relationship between each feature vector and each feature character. A plurality of the feature characters are spliced according to the corresponding feature order to obtain text information of each loss-determining image.

[0054] In this embodiment, combining a plurality of the feature characters according to the feature order can ensure the accuracy of the text information.

[0055] 102, generate the keyword label information corresponding to the keyword label based on the keyword label and the corresponding label position and text position of the text keyword in the claim form image.

[0056] In at least one embodiment of the present application, the electronic device generates the keyword label information corresponding to the keyword label based on the keyword label and the text keyword and the label position and the text position corresponding to the claim form image, comprising:

[0057] The electronic device detects whether the label rectangular area corresponding to the keyword label and the text rectangular area corresponding to the text keyword intersect according to the label position and the text position. If the label rectangular area and the text rectangular area intersect, the electronic device calculates the intersection area of the label rectangular area and each intersecting text rectangular area. Further, the electronic device calculates the first area ratio of the intersection area on the label rectangular area and the second area ratio of the intersection area on the text rectangular area. Still further, the electronic device selects the text rectangular area with both the first area ratio and the second area ratio greater than a preset threshold as a target rectangular area, and generates the keyword label information corresponding to the keyword label based on the number of target rectangular areas and the text keyword corresponding to the target rectangular area.

[0058] Specifically, the positioning method of the label position comprises:

[0059] The electronic device obtains the minimum horizontal coordinate value corresponding to the pixel points in the label rectangular area, and obtains the minimum vertical coordinate value corresponding to the pixel points in the label rectangular area. The electronic device obtains the maximum horizontal coordinate value corresponding to the pixel points in the label rectangular area, and obtains the maximum vertical coordinate value corresponding to the pixel points in the label rectangular area. The electronic device takes the first horizontal coordinate interval composed of the minimum horizontal coordinate value and the maximum horizontal coordinate value and the first vertical coordinate interval composed of the minimum vertical coordinate value and the maximum vertical coordinate value as the label position.

[0060] Specifically, the text position comprises a second horizontal coordinate interval and a second vertical coordinate interval. The electronic device detects whether the label rectangular area corresponding to the keyword label and the text rectangular area corresponding to the text keyword intersect according to the label position and the text position, comprising:

[0061] The electronic device identifies whether the first horizontal coordinate interval and the second horizontal coordinate interval overlap, and identifies whether the first vertical coordinate interval and the second vertical coordinate interval overlap. If the first horizontal coordinate interval and the second horizontal coordinate interval overlap, and the first vertical coordinate interval and the second vertical coordinate interval overlap, it is determined that the label rectangular area and the text rectangular area intersect.

[0062] Specifically, the electronic device generates the keyword label information corresponding to the keyword label based on the number of the target rectangular regions and the text keyword corresponding to the target rectangular regions, including:

[0063] If the number of the target rectangular regions is one, the electronic device determines the text keyword corresponding to the target rectangular region as the keyword label information corresponding to the keyword label, or if the number of the target rectangular regions is multiple, the electronic device performs weighted sum operation on each first area ratio and the corresponding second area ratio to obtain a final score value of each target rectangular region, and selects the text keyword corresponding to the target rectangular region with the maximum final score value as the keyword label information corresponding to the keyword label.

[0064] In the embodiment, when the number of the target rectangular regions is multiple, the text keyword corresponding to the target rectangular region with the maximum final score value is selected as the keyword label information corresponding to the keyword label. Since the final score value is generated by weighted operation, the keyword label information most matched with the keyword label can be reasonably selected.

[0065] 103, input the initial label, the initial label information, the keyword label and the keyword label information into the preset general coefficient model and the preset feature coefficient model respectively to calculate the general coefficient of the preset loss term and the feature coefficient of the preset loss term.

[0066] In at least one embodiment of the present application, the general coefficient model is used to calculate the general coefficient, the feature coefficient model is used to calculate the feature coefficient, and there is an intersection between the preset general data and the preset feature data. For example, the general coefficient includes a general base 8600, a general month 18 and a general disability coefficient 0.8, and the feature coefficient includes a feature base 8700 and a feature disability coefficient 0.8.

[0067] In at least one embodiment of the present application, before inputting the initial label, the initial label information, the keyword label and the keyword label information into the preset feature coefficient model, the method further includes:

[0068] The electronic device obtains a preset adversarial neural network, obtains position information to which the loss-determining object belongs and training data corresponding to the position information, selects preset feature data from the training data based on a preset feature keyword, and further trains the preset adversarial neural network based on the preset feature data to obtain the preset feature coefficient model.

[0069] The training data corresponding to the position information refers to claim information of a plurality of loss-determining objects in the position information, and the training data includes the preset general data and the preset characteristic data. The preset characteristic keyword includes a professional category of the loss-determining object, a type of a labor contract, and the like. For example, the preset characteristic keyword can be 2022106 road and bridge engineering technician, fixed-term labor contract, and the like.

[0070] In at least one embodiment of the present application, the general coefficient model includes a plurality of preset general disability grades, a general proportion coefficient corresponding to each preset general disability grade, and a general corresponding relationship corresponding to each preset general disability grade. The general corresponding relationship corresponding to each preset general disability grade includes a general base number label and a calculation method of a general base number corresponding to the general base number label, a general month number label and a calculation method of a general month number corresponding to the general month number label. The general coefficient includes a general disability coefficient, a general base number, and a general month number.

[0071] In at least one embodiment of the present application, the electronic device inputs the initial label, the initial label information, the keyword label, and the keyword label information into a preset general coefficient model to calculate a general coefficient of a preset loss item, which includes:

[0072] The electronic device selects a disability label from the initial label and the keyword label based on a preset disability keyword, and takes a value corresponding to the disability label in the claim form image as an initial disability grade of the loss-determining object. Further, the electronic device determines a general proportion coefficient corresponding to a preset general disability grade identical to the initial disability grade in the general coefficient model as the general disability coefficient, and determines a general corresponding relationship corresponding to the preset general disability grade in the general coefficient model as a target corresponding relationship. Still further, the electronic device identifies an initial base number label corresponding to the general base number label in the target corresponding relationship and an initial month number label corresponding to the general month number label in the target corresponding relationship from the initial label and the keyword label. Still further, the electronic device calculates the general base number based on an initial base number value of the initial base number label and a general base number calculation method in the target corresponding relationship, and calculates the general month number based on an initial month number value of the initial month number label and a general month number calculation method in the target corresponding relationship.

[0073] The preset loss item includes, but is not limited to, a one-time work injury compensation and a one-time disability employment gold.

[0074] For example, the initial label and the initial label information corresponding to the initial label are: working address: Shanghai and social security card number: 000888XXXXX, the keyword label and the keyword label information corresponding to the keyword label are: monthly average salary: 8000, gender: male, date of birth: 1987-12-28, age: 35, contract type: no fixed-term labor contract, and the like.

[0075] The initial base label includes the monthly average salary, and the initial month label includes the disability grade, gender, date of birth, and age. The preset disability keyword can be the disability grade, and the preset general disability grade includes grades from first-grade disability to tenth-grade disability. The range of the proportion coefficient corresponding to each preset general disability grade is [0, 1]. For example, the initial disability grade is third-grade disability, and the proportion coefficient corresponding to the third-grade disability is 0.7.

[0076] In this embodiment, when the initial label includes the initial base label, the initial base value is the initial label information corresponding to the initial label, and when the keyword label includes the initial base label, the initial base value is the keyword label information corresponding to the keyword label.

[0077] The calculation method of the general base includes:

[0078] The reference base value is determined according to the region where the loss-determining object is located, and the first salary threshold and the second salary threshold are determined according to the reference base value. The first salary threshold is greater than the second salary threshold. The first salary threshold is compared with the initial base value. If the initial base value is greater than or equal to the first salary threshold, the first salary threshold is determined as the general base. If the initial base value is less than or equal to the second salary threshold, the second salary threshold is determined as the general base. If the initial base value corresponding to the initial base label is between the first salary threshold and the second salary threshold, the initial base value is determined as the general base.

[0079] For example, the reference base value can be the average salary of workers in the region where the loss-determining object is located, the first salary threshold can be 300% of the average salary of workers, and the second salary threshold can be 60% of the average salary of workers.

[0080] The calculation method of the general month number includes:

[0081] A month number mapping table is constructed based on a preset loss item, a preset date of birth, a preset age, a preset disability grade, and a preset month value. The initial month label is mapped based on the month number mapping table, and the preset month value corresponding to the initial month label is determined as the general month number.

[0082] In this embodiment, the universal coefficient can be accurately calculated based on the specific information of the damage assessment object based on the calculation method of the universal base and the calculation method of the universal number of months.

[0083] In at least one embodiment of the present application, the characteristic coefficient model includes a label base value corresponding to the characteristic base label, a characteristic month label and a label month value corresponding to the characteristic month label, a characteristic coefficient label and a characteristic proportional coefficient corresponding to the characteristic coefficient label, and the characteristic coefficient includes a characteristic base, a characteristic month and a characteristic disability coefficient.

[0084] In at least one embodiment of the present application, the electronic device inputs the initial tag, the initial tag information, the keyword tag, and the keyword tag information into a preset characteristic coefficient model to calculate the characteristic coefficient of the preset loss item, including:

[0085] The electronic device obtains a cardinality tag type corresponding to the characteristic cardinality tag, and detects whether there is a tag in the initial tag and the keyword tag corresponding to the characteristic cardinality tag based on the cardinality tag type.

[0086] In this embodiment, if there is a label in the initial label or the keyword label that corresponds to the characteristic cardinality label, the electronic device determines the label cardinality value corresponding to the characteristic cardinality label as the characteristic cardinality, the electronic device obtains the month label type corresponding to the characteristic month label, and detects whether there is a label in the initial label and the keyword label that corresponds to the characteristic month label based on the month label type.

[0087] In this embodiment, if there is a label in the initial label or the keyword label that corresponds to the characteristic month label, the electronic device determines the label month value corresponding to the characteristic month label as the characteristic month, the electronic device obtains the coefficient label type corresponding to the characteristic coefficient label, and detects whether there is a label in the initial label and the keyword label that corresponds to the characteristic coefficient label based on the coefficient label type.

[0088] In this embodiment, if there is a tag in the initial tag or the keyword tag that corresponds to the characteristic coefficient tag, the electronic device determines the tag coefficient value corresponding to the characteristic coefficient tag as the characteristic disability coefficient.

[0089] The characteristic base number label includes, but is not limited to, disability grade, average monthly salary, household registration nature, whether a labor contract is signed, work address, nature of the employment unit, type of the labor contract if signed, and occupation category. The characteristic monthly number label includes, but is not limited to, disability grade, age, gender, occupation category, household registration nature, work address, whether a labor contract is signed, and whether a labor relationship with the employment unit is maintained so far. The characteristic coefficient label includes, but is not limited to, disability grade and household registration nature.

[0090] According to the above embodiments, since the characteristic base number label, the characteristic monthly number label, and the characteristic coefficient label involve more specific information related to the loss object, the characteristic coefficient is more accurate than the general coefficient.

[0091] 104, adjusting the general coefficient based on the characteristic coefficient to obtain a target coefficient of the preset loss item, and correcting the target coefficient based on a preset loss standard library to obtain a loss coefficient of the preset loss item.

[0092] In at least one embodiment of the present application, the electronic device detects whether the characteristic coefficient corresponds to each general coefficient. If the characteristic coefficient does not correspond to each general coefficient, the electronic device determines the general coefficient as the target coefficient. Alternatively, if at least one characteristic coefficient corresponds to the general coefficient, the electronic device replaces the general coefficient with the corresponding characteristic coefficient and determines the replaced general coefficient as the target coefficient.

[0093] For example, the characteristic base number corresponds to the general base number, the characteristic proportion coefficient corresponds to the general proportion coefficient, and the general disability coefficient corresponds to the characteristic disability coefficient. For example, when the general base number is 8600, the general monthly number is 18, and the general disability coefficient is 0.8, the characteristic base number is 8700, and the characteristic disability coefficient is 0.8, the general base number 8600 corresponds to the characteristic base number 8700, and the general disability coefficient 0.8 corresponds to the characteristic disability coefficient 0.8.

[0094] According to the above embodiments, when at least one characteristic coefficient corresponds to the general coefficient, the general coefficient is replaced with the corresponding characteristic coefficient to obtain the target coefficient. Since the characteristic coefficient is more accurate than the general coefficient, the accuracy of the target coefficient can be improved.

[0095] In at least one embodiment of the present application, the preset loss-determination standard library includes, but is not limited to, a disability national standard library, a compensation standard library, and a disability-determination-by-injury library. The disability national standard library is used to store respective disability judgment standards corresponding to respective geographical regions, the compensation standard library is used to store compensation standards issued by the National Bureau of Statistics according to regions, and the disability-determination-by-injury library is used to store corresponding relationships between geographical regions, lower age limits, upper age limits, and compensation-month evaluation algorithms.

[0096] In at least one embodiment of the present application, the electronic device corrects the target coefficient based on the preset loss-determination standard library to obtain a loss-determination coefficient of the preset loss-determination item, including:

[0097] The electronic device inputs the initial label, the initial label information, the keyword label, and the keyword label information into the preset loss-determination standard library, and determines a loss-determination label associated with each preset loss-determination item according to a plurality of preset loss-determination items matched by the loss-determination object. Further, the electronic device determines matching parameters of the loss-determination object on the loss-determination label associated with each preset loss-determination item according to information data stored in the preset loss-determination standard library. Further, the electronic device compares the target coefficient with the matching parameters. If a difference between the target coefficient and the matching parameters is less than or equal to a configuration value, the electronic device determines the target coefficient as the loss-determination coefficient. Alternatively, if the difference between the target coefficient and the matching parameters is greater than the configuration value, the electronic device determines the matching parameters as the loss-determination coefficient.

[0098] For example, when a preset loss-determination item is one-time work injury compensation, the loss-determination label of the one-time work injury compensation includes a compensation base value, a disability coefficient, and a compensation month number. The compensation base value stored in the loss-determination standard library is 8000, the disability coefficient is 0.8, and the compensation month number is 16. When the target coefficient includes a target base value 8050, a target disability coefficient 0.8, and a target compensation month number 18, a first configuration value corresponding to the compensation base value 8000 and the target base value 8050 is 100, a second configuration value corresponding to the disability coefficient 0.8 and the target disability coefficient 0.8 is 0.1, and a third configuration value corresponding to the compensation month number 16 and the target compensation month number 18 is 1. The loss-determination base of the loss-determination coefficient is 8000, the loss-determination disability coefficient of the loss-determination coefficient is 0.8, and the loss-determination month number of the loss-determination coefficient is 18.

[0099] According to the above embodiment, the matching parameters are generated by analyzing all information of the loss-determination object through the preset loss-determination standard library, and therefore, logical operation processing is performed according to the obtained matching parameters to complete output of the loss-determination coefficient, which not only realizes automatic calculation of personal injury loss determination, but also improves accuracy of the loss-determination coefficient.

[0100] 105, the claim data of the loss-determination object is calculated according to the plurality of preset loss items and the loss-determination coefficient corresponding to each preset loss item.

[0101] It should be emphasized that, in order to further ensure the privacy and security of the above-mentioned claim data, the claim data can also be stored in a node of a block chain.

[0102] In at least one embodiment of the present application, the loss-determination coefficient includes a loss-determination base, a loss-determination month number, and a loss-determination damage residual coefficient, and the electronic device calculates the claim data of the loss-determination object according to the plurality of preset loss items and the loss-determination coefficient corresponding to each preset loss item, which includes:

[0103] The electronic device performs multiplication operation on the loss-determination base, the loss-determination month number, and the loss-determination damage residual coefficient to obtain the loss-determination amount corresponding to each preset loss item, and further, the electronic device performs addition operation on the loss-determination amounts corresponding to the plurality of preset loss items to obtain the claim data.

[0104] According to the above embodiment, multiplication and addition operation are performed on the loss-determination base, the loss-determination month number, and the loss-determination damage residual coefficient corresponding to the plurality of preset loss items, which can quickly calculate the claim data.

[0105] As can be seen from the above technical solutions, the present application can comprehensively obtain information of the loss-determination object by recognizing the claim form image, extract text keywords and numerical keywords in the claim form image information, remove redundant information, determine keyword label information corresponding to the keyword label based on the label position and the text position, correspond the keyword label and the keyword label information, which is conducive to quickly calculating the general coefficient and the characteristic coefficient based on the corresponding relationship between the keyword label and the keyword label information, adjusting the general coefficient based on the characteristic coefficient to obtain the target coefficient corresponding to the preset loss item, since the preset characteristic data is more specific information than the preset general data, the accuracy of the characteristic coefficient model trained using the characteristic data can be improved, so that the characteristic coefficient output by the characteristic coefficient model is more accurate, adjusting the general coefficient based on the characteristic coefficient can improve the accuracy of the target coefficient, and modifying the target coefficient based on the preset loss-determination standard library can modify the target coefficient based on the accurate loss-determination result output by the loss-determination standard library, which can ensure the accuracy of the claim data.

[0106] As Figure 2 shown in FIG. 1, is a functional module diagram of a preferred embodiment of the loss settlement and claim device. The loss settlement and claim device 11 includes an acquisition unit 110, a generation unit 111, an input unit 112, an adjustment unit 113, and a calculation unit 114. The modules / units referred to in the present application refer to a series of computer-readable instruction segments capable of being acquired by the processor 13 and capable of completing a fixed function, which are stored in the memory 12. In the present embodiment, the functions of each module / unit will be described in detail in subsequent embodiments.

[0107] The acquisition unit 110 acquires an initial label of a loss settlement object, initial label information corresponding to the initial label, and a claim form image, and uses a pre-trained text recognition model to recognize the claim form image to obtain a keyword label and a text keyword of the claim form image.

[0108] In at least one embodiment of the present application, the initial label refers to partial text data of the loss settlement object, the initial label information refers to keyword information corresponding to the initial label, and the claim form image refers to an image containing information required when the loss settlement object claims. For example, the initial label includes, but is not limited to, the unified planning area where the loss settlement object is located, the social security card number of the loss settlement object. When the initial label is the unified planning area where the loss settlement object is located, the initial label information corresponding to the initial label may be, for example, Shanghai, and the claim form image contains the loss settlement object's previous disability identification information, labor contract performance, household nature, gender, and date of birth, etc.

[0109] In at least one embodiment of the present application, the acquisition unit 110 provides an information input interface for a user to input the initial label information and import the claim form image through the information input interface. The information input interface includes an input label, an input button corresponding to the input label, an information input field corresponding to the input label, an image upload button, and an image import field corresponding to the image upload button. The acquisition unit 110 receives a first trigger instruction generated by the user pressing the input button, and receives the initial label information input by the user in the information input field corresponding to the input label according to the first trigger instruction. The acquisition unit 110 receives a second trigger instruction generated by the user pressing the image upload button, and receives the claim form image uploaded by the user in the image import field according to the second trigger instruction.

[0110] In at least one embodiment of the present application, the obtaining unit 110 uses a pre-trained character recognition model to recognize the claim form image, to obtain the keyword label and the text keyword of the claim form image.

[0111] In at least one embodiment of the present application, the character recognition model refers to a model for recognizing character information and numerical information in the claim form image.

[0112] In at least one embodiment of the present application, the keyword label includes a first keyword label and a second keyword label, the text keyword includes a character keyword and a numerical keyword, the first keyword label refers to a label corresponding to the character keyword, the second keyword label refers to a label corresponding to the numerical keyword, the character keyword refers to the character specific information of the loss object corresponding to the first keyword label, and the numerical keyword refers to the numerical specific information of the loss object corresponding to the second keyword label. For example, when the first keyword label is gender, the character keyword is male, when the second keyword label is birth date, the numerical keyword is 1987-12-28, when the second keyword label is age, the numerical keyword is 35, and so on.

[0113] In at least one embodiment of the present application, the character recognition model includes a recurrent sequence generation network and a decoding network, and the obtaining unit 110 uses a pre-trained character recognition model to recognize the claim form image, to obtain the keyword label and the text keyword of the claim form image, including:

[0114] The obtaining unit 110 locates the position of the text information in the claim form image based on the pixel value of the pixel point in the claim form image, to obtain a character position, further, the obtaining unit 110 performs feature extraction on the text information based on the character position, to obtain a feature sequence, still further, the obtaining unit 110 inputs the feature sequence into the recurrent sequence generation network, to obtain a recurrent sequence, still further, the obtaining unit 110 decodes the recurrent sequence based on the decoding network, to obtain the character information of each loss image, still further, the obtaining unit 110 determines a first preset keyword in the character information as the keyword label, and determines a second preset keyword in the character information as the text keyword.

[0115] In this embodiment, the first preset keyword includes, but is not limited to, gender, occupation category, age, birth date, and monthly average salary, and so on, and the second preset keyword includes, but is not limited to, male, female, technical personnel, 35, 1987-12-28, and 8000, and so on.

[0116] Specifically, the obtaining unit 110 performs feature extraction on the text information based on a feature extraction network to obtain the feature sequence, the feature extraction network can be a resnet50 network, the recurrent sequence generation network can be a bidirectional long short-term memory neural network, and the decoding network can be a connectionist temporal classification (CTC) algorithm based on a neural network.

[0117] Specifically, the obtaining unit 110 locates the position of the text information in the claim form image based on the pixel value of the pixel point in the claim form image to obtain the character position, including:

[0118] The obtaining unit 110 obtains the pixel position of all pixel points in the claim form image, and the obtaining unit 110 traverses the pixel value of all pixel points in the claim form image. Further, the obtaining unit 110 determines a plurality of pixel points corresponding to a pixel value greater than a preset value as text information, and determines the pixel position of the corresponding plurality of pixel points as the character position.

[0119] Specifically, the recurrent sequence generation network includes a convolution layer and a pooling layer, and the obtaining unit 110 inputs the feature sequence into the recurrent sequence generation network to obtain a recurrent sequence, including:

[0120] The obtaining unit 110 performs segmentation processing on the feature sequence to obtain a plurality of segmented features. Further, the obtaining unit 110 performs convolution operation on the plurality of segmented features based on the convolution layer to obtain a convolution result. Further, the obtaining unit 110 performs pooling processing on the convolution result based on the pooling layer to obtain a feature vector corresponding to each segmented feature. Further, the obtaining unit 110 splices a plurality of the feature vectors to obtain the recurrent sequence.

[0121] The segmentation processing process is a prior art, which is not described herein.

[0122] Specifically, the obtaining unit 110 decodes the recurrent sequence based on the decoding network to obtain the text information of each loss-determining image, including:

[0123] The obtaining unit 110 obtains the feature order corresponding to each feature vector in the recurrent sequence, and performs mapping processing on each feature vector in the recurrent sequence based on a pre-constructed dictionary to obtain a feature character corresponding to each feature vector. The obtaining unit 110 splices a plurality of the feature characters according to the feature order to obtain the text information of each loss-determining image.

[0124] The pre-constructed dictionary stores a corresponding relationship between each feature vector and each feature character, and the plurality of feature characters are spliced according to the corresponding feature sequence to obtain the text information of each loss determination image.

[0125] In this embodiment, the plurality of feature characters are combined according to the feature sequence, so that the accuracy of the text information can be ensured.

[0126] The generation unit 111 generates keyword label information corresponding to the keyword label based on the keyword label and the label position and the text position corresponding to the text keyword in the claim form image.

[0127] In at least one embodiment of the present application, the generation unit 111 generates keyword label information corresponding to the keyword label based on the keyword label and the label position and the text position corresponding to the text keyword in the claim form image, including:

[0128] The generation unit 111 detects whether the label rectangular region corresponding to the keyword label intersects with the text rectangular region corresponding to the text keyword according to the label position and the text position. If the label rectangular region intersects with the text rectangular region, the generation unit 111 calculates the intersection region of the label rectangular region and each intersecting text rectangular region. Further, the generation unit 111 calculates a first area ratio of the intersection region on the label rectangular region and a second area ratio of the intersection region on the text rectangular region. Still further, the generation unit 111 selects a text rectangular region with both the first area ratio and the second area ratio greater than a preset threshold as a target rectangular region, and generates keyword label information corresponding to the keyword label based on the number of target rectangular regions and the text keyword corresponding to the target rectangular region.

[0129] Specifically, the positioning mode of the label position includes:

[0130] The generation unit 111 obtains the smallest horizontal coordinate value corresponding to the pixel points in the label rectangular region, and obtains the smallest vertical coordinate value corresponding to the pixel points in the label rectangular region. The generation unit 111 obtains the largest horizontal coordinate value corresponding to the pixel points in the label rectangular region, and obtains the largest vertical coordinate value corresponding to the pixel points in the label rectangular region. The generation unit 111 takes the first horizontal coordinate interval formed by the smallest horizontal coordinate value and the largest horizontal coordinate value and the first vertical coordinate interval formed by the smallest vertical coordinate value and the largest vertical coordinate value as the label position.

[0131] Specifically, the text position includes a second horizontal coordinate interval and the second vertical coordinate interval, and the generation unit 111 detects whether the label rectangle region corresponding to the keyword label and the text rectangle region corresponding to the text keyword intersect according to the label position and the text position, including:

[0132] The generation unit 111 identifies whether the first horizontal coordinate interval and the second horizontal coordinate interval overlap, and identifies whether the first vertical coordinate interval and the second vertical coordinate interval overlap, and determines that the label rectangle region and the text rectangle region intersect if the first horizontal coordinate interval and the second horizontal coordinate interval overlap and the first vertical coordinate interval and the second vertical coordinate interval overlap.

[0133] Specifically, the generation unit 111 generates the keyword label information corresponding to the keyword label based on the number of target rectangular regions and the text keyword corresponding to the target rectangular region, including:

[0134] If the number of target rectangular regions is single, the generation unit 111 determines the text keyword corresponding to the target rectangular region as the keyword label information corresponding to the keyword label, or if the number of target rectangular regions is multiple, the generation unit 111 performs weighted sum operation on each first area ratio and the corresponding second area ratio to obtain the final score value of each target rectangular region, and selects the text keyword corresponding to the target rectangular region with the maximum final score value as the keyword label information corresponding to the keyword label.

[0135] In the embodiment, when the number of target rectangular regions is multiple, the text keyword corresponding to the target rectangular region with the maximum final score value is selected as the keyword label information corresponding to the keyword label. Since the final score value is generated by weighted operation, the keyword label information that best matches the keyword label can be reasonably selected.

[0136] The input unit 112 inputs the initial label, the initial label information, the keyword label and the keyword label information into a preset general coefficient model and a preset feature coefficient model respectively to calculate the general coefficient of the preset loss term and the feature coefficient of the preset loss term.

[0137] In at least one embodiment of the present application, the general coefficient model is used to calculate the general coefficient, the feature coefficient model is used to calculate the feature coefficient, and there is an intersection between the preset general data and the preset feature data. For example, the general coefficient includes a general base 8600, a general monthly number 18 and a general disability coefficient 0.8, and the feature coefficient includes a feature base 8700 and a feature disability coefficient 0.8.

[0138] In at least one embodiment of the present application, before the initial label, the initial label information, the keyword label and the keyword label information are input into the preset characteristic coefficient model, the method further comprises:

[0139] The input unit 112 is also used to obtain a preset adversarial neural network, and obtain position information to which the loss object belongs and training data corresponding to the position information, and select preset feature data from the training data based on a preset feature keyword. Further, the input unit 112 trains the preset adversarial neural network based on the preset feature data to obtain the preset characteristic coefficient model.

[0140] The training data corresponding to the position information refers to the claim information of a plurality of loss objects in the position information, and the training data includes the preset general data and the preset feature data. The preset feature keyword includes the occupation category of the loss object, the type of labor contract, etc. For example, the preset feature keyword can be 2022106 road and bridge engineering technician, fixed-term labor contract, etc.

[0141] In at least one embodiment of the present application, the general coefficient model includes a plurality of preset general injury grades, a general proportion coefficient corresponding to each preset general injury grade, and a general corresponding relationship corresponding to each preset general injury grade. Each preset general injury grade corresponding to the general corresponding relationship includes a general base number label and a calculation method of a general base number corresponding to the general base number label, a general month label and a calculation method of a general month number corresponding to the general month label. The general coefficient includes a general injury coefficient, a general base number and a general month number.

[0142] In at least one embodiment of the present application, the input unit 112 inputs the initial label, the initial label information, the keyword label and the keyword label information into the preset general coefficient model to calculate the general coefficient of the preset loss item, which includes:

[0143] The input unit 112 selects a disability label from the initial label and the keyword label based on a preset disability keyword, and takes a value corresponding to the disability label in the claim form image as an initial disability grade of the loss assessment object. Further, the input unit 112 determines a universal proportion coefficient corresponding to a preset universal disability grade same as the initial disability grade in the universal coefficient model as the universal disability coefficient, and determines a universal corresponding relationship corresponding to the preset universal disability grade in the universal coefficient model as a target corresponding relationship. Further, the input unit 112 identifies an initial base label corresponding to the universal base label in the target corresponding relationship and an initial month label corresponding to the universal month label in the target corresponding relationship from the initial label and the keyword label. Further, the input unit 112 calculates the universal base based on an initial base value of the initial base label and a universal base calculation mode in the target corresponding relationship, and calculates the universal month based on an initial month value of the initial month label and a universal month calculation mode in the target corresponding relationship.

[0144] The preset loss items include, but are not limited to, one-time work injury compensation and one-time disability employment gold.

[0145] For example, the initial label and the initial label information corresponding to the initial label are: work address: Shanghai and social security card number: 000888XXXXX, the keyword label and the keyword label information corresponding to the keyword label are: monthly average salary: 8000, gender: male, date of birth: 1987-12-28, age: 35, contract type: no fixed-term labor contract, etc.

[0146] The initial base label includes monthly average salary, and the initial month label includes disability grade, gender, date of birth, and age. The preset disability keyword can be disability grade. The preset universal disability grade includes first-class disability to tenth-class disability, and the proportion coefficient corresponding to each preset universal disability grade ranges from 0 to 1. For example, the initial disability grade is third-class disability, and the proportion coefficient corresponding to third-class disability is 0.7.

[0147] In the embodiment, when the initial label includes the initial base label, the initial base value is the initial label information corresponding to the initial label, and when the keyword label includes the initial base label, the initial base value is the keyword label information corresponding to the keyword label.

[0148] The calculation mode of the universal base includes:

[0149] The reference base value is determined according to a region where the loss-determining object is located, and the first wage threshold and the second wage threshold are determined according to the reference base value, the first wage threshold is greater than the second wage threshold, the first wage threshold is compared with the initial base value, if the initial base value is greater than or equal to the first wage threshold, the first wage threshold is determined as the universal base, if the initial base value is less than or equal to the second wage threshold, the second wage threshold is determined as the universal base, and if the initial base value corresponding to the initial base label is between the first wage threshold and the second wage threshold, the initial base value is determined as the universal base.

[0150] For example, the reference base value can be the average wage of workers in the region where the loss-determining object is located, the first wage threshold can be 300% of the average wage of workers, and the second wage threshold can be 60% of the average wage of workers.

[0151] The calculation method of the universal month number includes:

[0152] The month number mapping table is constructed based on the preset loss item, the preset birth date, the preset age, the preset disability grade and the preset month value, the initial month number label is mapped based on the month number mapping table, and the preset month value corresponding to the initial month number label is determined as the universal month number.

[0153] In the embodiment, the calculation method of the universal base and the universal month number calculation method can accurately calculate the universal coefficient according to the specific information of the loss-determining object.

[0154] In at least one embodiment of the present application, the feature coefficient model includes the label base value corresponding to the feature base label, the feature month number label and the label month value corresponding to the feature month number label, the feature coefficient label and the feature proportion coefficient corresponding to the feature coefficient label, and the feature coefficient includes the feature base, the feature month number and the feature disability coefficient.

[0155] In at least one embodiment of the present application, the input unit 112 inputs the initial label, the initial label information, the keyword label and the keyword label information into the preset feature coefficient model to calculate the feature coefficient of the preset loss item, which includes:

[0156] The input unit 112 acquires the base label type corresponding to the feature base label, and detects whether there is a label corresponding to the feature base label in the initial label and the keyword label based on the base label type.

[0157] In the embodiment, if there is a label corresponding to the characteristic base number label in the initial label or the keyword label, the input unit 112 determines the label base number value corresponding to the characteristic base number label as the characteristic base number, and the input unit 112 acquires the month label type corresponding to the characteristic month label, and detects whether there is a label corresponding to the characteristic month label in the initial label and the keyword label based on the month label type.

[0158] In the embodiment, if there is a label corresponding to the characteristic month label in the initial label or the keyword label, the input unit 112 determines the label month value corresponding to the characteristic month label as the characteristic month, and the input unit 112 acquires the coefficient label type corresponding to the characteristic coefficient label, and detects whether there is a label corresponding to the characteristic coefficient label in the initial label and the keyword label based on the coefficient label type.

[0159] In the embodiment, if there is a label corresponding to the characteristic coefficient label in the initial label or the keyword label, the input unit 112 determines the label coefficient value corresponding to the characteristic coefficient label as the characteristic disability coefficient.

[0160] The characteristic base number label includes, but is not limited to, disability grade, average monthly salary, household nature, whether a labor contract is signed, work address, nature of the employment unit, type of the labor contract if signed, and occupation category, the characteristic month label includes, but is not limited to, disability grade, age, gender, occupation category, household nature, work address, whether a labor contract is signed, and whether the labor relationship with the employment unit is maintained so far, and the characteristic coefficient label includes, but is not limited to, disability grade and household nature.

[0161] According to the above-mentioned embodiments, since the characteristic base number label, the characteristic month label and the characteristic coefficient label involve more specific information related to the damage object, the characteristic coefficient can be more accurate than the general coefficient.

[0162] The adjustment unit 113 adjusts the general coefficient based on the characteristic coefficient to obtain a target coefficient of the preset damage item, and corrects the target coefficient based on the preset damage standard library to obtain a damage coefficient of the preset damage item.

[0163] In at least one embodiment of the present application, the adjustment unit 113 detects whether the characteristic coefficient corresponds to each general coefficient. If the characteristic coefficient does not correspond to each general coefficient, the adjustment unit 113 determines the general coefficient as the target coefficient. Alternatively, if there is at least one characteristic coefficient corresponding to the general coefficient, the adjustment unit 113 replaces the general coefficient with the corresponding characteristic coefficient and determines the replaced general coefficient as the target coefficient.

[0164] For example, the characteristic base corresponds to the general base, the characteristic proportion coefficient corresponds to the general proportion coefficient, and the general disability coefficient corresponds to the characteristic disability coefficient. For example, when the general base is 8600, the general month is 18, and the general disability coefficient is 0.8, the characteristic base is 8700, and the characteristic disability coefficient is 0.8, the general base 8600 corresponds to the characteristic base 8700, and the general disability coefficient 0.8 corresponds to the characteristic disability coefficient 0.8.

[0165] According to the above embodiment, when there is at least one characteristic coefficient corresponding to the general coefficient, the general coefficient is replaced with the corresponding characteristic coefficient to obtain the target coefficient. Since the characteristic coefficient is more accurate than the general coefficient, the accuracy of the target coefficient can be improved.

[0166] In at least one embodiment of the present application, the preset loss determination standard library includes, but is not limited to, a disability national standard library, a compensation standard library, and a disability determination library. The disability national standard library is used to store the respective disability judgment standards corresponding to each geographical area. The compensation standard library is used to store the compensation standards issued by the National Bureau of Statistics according to the region. The disability determination library is used to store the corresponding relationship between the geographical area, the lower limit of age, the upper limit of age, and the compensation month evaluation algorithm.

[0167] In at least one embodiment of the present application, the adjustment unit 113 corrects the target coefficient based on the preset loss determination standard library to obtain the loss determination coefficient of the preset loss item, including:

[0168] The adjustment unit 113 inputs the initial label, the initial label information, the keyword label and the keyword label information into the preset loss standard library, and determines the loss label associated with each preset loss item according to the plurality of preset loss items matched by the loss object. Further, the adjustment unit 113 determines the matching parameter of the loss object on the loss label associated with each preset loss item according to the information data stored in the preset loss standard library. Further, the adjustment unit 113 compares the target coefficient with the matching parameter. If the difference between the target coefficient and the matching parameter is less than or equal to the configuration value, the adjustment unit 113 determines the target coefficient as the loss coefficient. Alternatively, if the difference between the target coefficient and the matching parameter is greater than the configuration value, the adjustment unit 113 determines the matching parameter as the loss coefficient.

[0169] For example, when a preset loss item is one-time compensation for work injury, the loss label of the one-time compensation for work injury includes compensation base value, disability coefficient and compensation month. The compensation base value stored in the loss standard library is 8000, the disability coefficient is 0.8, and the compensation month is 16. When the target coefficient includes a target base value 8050, a target disability coefficient 0.8 and a target compensation month 18, the first configuration value corresponding to the compensation base value 8000 and the target base value 8050 is 100, the second configuration value corresponding to the disability coefficient 0.8 and the target disability coefficient 0.8 is 0.1, and the third configuration value corresponding to the compensation month 16 and the target compensation month 18 is 1. The loss base of the loss coefficient is 8000, the loss disability coefficient of the loss coefficient is 0.8, and the loss month of the loss coefficient is 18.

[0170] Through the above embodiment, the matching parameter is generated by analyzing all information of the loss object through the preset loss standard library. Therefore, logical operation processing is performed according to the obtained matching parameter to complete the output of the loss coefficient. Not only can the automatic calculation of personal injury loss be realized, but also the accuracy of the loss coefficient can be improved.

[0171] The calculation unit 114 calculates the claim data of the loss object according to the plurality of preset loss items and the loss coefficient corresponding to each preset loss item.

[0172] It should be emphasized that, in order to further ensure the privacy and security of the above claim data, the above claim data can also be stored in a node of a block chain.

[0173] In at least one embodiment of the present application, the loss-determining coefficient includes a loss-determining base, a loss-determining month number, and a loss-determining damage coefficient, and the calculation unit 114 calculates the claim data of the loss-determining object according to the plurality of preset loss-determining items and the loss-determining coefficient corresponding to each preset loss-determining item, which includes:

[0174] The calculation unit 114 multiplies the loss-determining base, the loss-determining month number, and the loss-determining damage coefficient to obtain the loss-determining amount corresponding to each preset loss-determining item, and further adds the loss-determining amounts corresponding to the plurality of preset loss-determining items to obtain the claim data.

[0175] Through the above implementation, the loss-determining base, the loss-determining month number, and the loss-determining damage coefficient corresponding to the plurality of preset loss-determining items are multiplied and added to quickly calculate the claim data.

[0176] As can be seen from the above technical solutions, the present application can comprehensively obtain the information of the loss-determining object by recognizing the claim form image, extract the text keywords and numerical keywords in the claim form image information, remove redundant information, determine the keyword label information corresponding to the keyword label based on the label position and the text position, correspond the keyword label and the keyword label information, which is conducive to quickly calculating the general coefficient and the characteristic coefficient based on the corresponding relationship between the keyword label and the keyword label information, adjusting the general coefficient based on the characteristic coefficient to obtain the target coefficient corresponding to the preset loss-determining item, since the preset characteristic data is more specific information than the preset general data, the accuracy of the characteristic coefficient model trained using the characteristic data can be improved, so that the characteristic coefficient output by the characteristic coefficient model is more accurate, adjusting the general coefficient based on the characteristic coefficient can improve the accuracy of the target coefficient, and modifying the target coefficient based on the preset loss-determining standard library can modify the target coefficient based on the accurate loss-determining result output by the loss-determining standard library, which can ensure the accuracy of the claim data.

[0177] As shown in Figure 3 FIG. 1 is a structural schematic diagram of an electronic device for implementing the loss-determining claim method according to a preferred embodiment of the present application.

[0178] In one embodiment of the present application, the electronic device 1 includes, but is not limited to, a memory 12, a processor 13, and computer readable instructions stored in the memory 12 and executable on the processor 13, such as a loss-determining claim program.

[0179] Those skilled in the art can understand that the schematic diagram is only an example of the electronic device 1, and does not constitute a limitation on the electronic device 1, and can include more or less components than the diagram, or combine certain components, or different components, for example, the electronic device 1 can also include an input / output device, a network access device, a bus, etc.

[0180] The processor 13 can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor or the like. The processor 13 is the operation core and control center of the electronic device 1, and connects various parts of the entire electronic device 1 through various interfaces and lines, and executes the operating system and various application programs, program codes, etc. installed in the electronic device 1.

[0181] For example, the computer readable instructions can be divided into one or more modules / units, for example Figure 2 The acquisition unit, the generation unit, the input unit, the adjustment unit, the calculation unit, etc. shown are stored in the memory 12 and executed by the processor 13 to complete the present application. The one or more modules / units can be a series of computer readable instruction segments capable of completing a specific function, which are used to describe the execution process of the computer readable instructions in the electronic device 1.

[0182] The memory 12 can be used to store the computer readable instructions and / or modules, and the processor 13 realizes various functions of the electronic device 1 by running or executing the computer readable instructions and / or modules stored in the memory 12, and calling the data stored in the memory 12. The memory 12 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application program required by a function (such as a sound playing function, an image playing function, etc.), and the like; and the data storage area can store data created according to the use of the electronic device, etc. The memory 12 can include non-volatile and volatile memories, for example: a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash storage device, or other storage devices.

[0183] The memory 12 can be an external memory and / or an internal memory of the electronic device 1. Further, the memory 12 can be a memory having a physical form, such as a memory stick, a TF card (Trans-flash Card), and the like.

[0184] The modules / units integrated in the electronic device 1, if implemented in the form of software function units and sold or used as independent products, can be stored in a computer readable storage medium. Based on such understanding, all or part of the processes in the above-mentioned embodiment methods can also be implemented through computer readable instructions for instructing related hardware, and the computer readable instructions can be stored in a computer readable storage medium. When the computer readable instructions are executed by a processor, the steps of the above-mentioned various method embodiments can be implemented.

[0185] The computer readable instructions include computer readable instruction codes, which can be in the form of source code, object code, executable files, or some intermediate forms, etc. The computer readable medium can include any entity or device capable of carrying the computer readable instruction codes, recording media, U disks, mobile hard disks, magnetic disks, optical disks, computer memories, read-only memories (ROMs), random access memories (RAMs).

[0186] The blockchain referred to in the present application is a new application mode of distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm and other computer technologies. The blockchain is essentially a decentralized database, which is a series of data blocks associated using cryptographic methods, each data block contains information of a batch of network transactions, used to verify the validity (anti-fake) of the information and generate the next block. The blockchain can include a blockchain underlying platform, a platform product service layer, and an application service layer, etc.

[0187] In combination Figure 1 The memory 12 in the electronic device 1 stores computer readable instructions to implement a loss assessment and claim settlement method, and the processor 13 can execute the computer readable instructions to implement:

[0188] obtain an initial label of a loss assessment object, initial label information corresponding to the initial label, and a claim form image, and use a pre-trained text recognition model to recognize the claim form image to obtain a keyword label of the claim form image and a text keyword; generate keyword label information corresponding to the keyword label based on the label position and the text position corresponding to the keyword label and the text keyword in the claim form image; input the initial label, the initial label information, the keyword label, and the keyword label information into a preset general coefficient model and a preset feature coefficient model respectively to calculate a general coefficient of a preset loss item and a feature coefficient of the preset loss item; adjust the general coefficient based on the feature coefficient to obtain a target coefficient of the preset loss item, and correct the target coefficient based on a preset loss assessment standard library to obtain a loss assessment coefficient of the preset loss item; and calculate claim settlement data of the loss assessment object according to a plurality of the preset loss items and the loss assessment coefficient corresponding to each preset loss item.

[0189] Specifically, the specific implementation method of the processor 13 on the above computer readable instructions can refer to the description of the related steps in the corresponding embodiments, which will not be described here. Figure 1 The corresponding embodiments will not be described here.

[0190] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the above-described device embodiments are only schematic, for example, the division of the modules is only a logical function division, and there can be another division manner in actual implementation.

[0191] The computer readable storage medium stores computer readable instructions, wherein the computer readable instructions are executed by the processor 13 to implement the following steps:

[0192] The initial label of the loss assessment object, initial label information corresponding to the initial label, and a claim form image are acquired, and a pre-trained character recognition model is used to recognize the claim form image to obtain a keyword label of the claim form image and a text keyword; keyword label information corresponding to the keyword label is generated based on a label position and a text position corresponding to the keyword label and the text keyword in the claim form image; the initial label, the initial label information, the keyword label, and the keyword label information are input into a preset general coefficient model and a preset feature coefficient model, respectively, to calculate a general coefficient of a preset loss item and a feature coefficient of the preset loss item; the general coefficient is adjusted based on the feature coefficient to obtain a target coefficient of the preset loss item, and the target coefficient is corrected based on a preset loss assessment standard library to obtain a loss assessment coefficient of the preset loss item; and claim data of the loss assessment object is calculated according to a plurality of the preset loss items and the loss assessment coefficient corresponding to each preset loss item.

[0193] The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical units, i.e., they may be located in one place, or distributed on multiple network units. Some or all of the modules can be selected to achieve the purpose of the embodiment to implement the embodiment.

[0194] In addition, each functional module in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present alone, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of hardware plus software functional modules.

[0195] Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting, and the scope of the present application is defined by the appended claims rather than the above description, and therefore all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present application. Any additional reference signs in the claims should not be considered as limiting the claims involved.

[0196] In addition, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. The plurality of units or devices described can also be implemented by one unit or device through software or hardware. The words first, second, etc. are used to indicate names and do not represent any particular order.

[0197] Finally, it should be noted that the above examples are only used to illustrate the technical solutions of the present application and are not limiting. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present application.

Claims

1. A method of loss assessment and claim settlement, characterized in that, The loss settlement and claim method comprises: Obtaining an initial label of a loss settlement object, initial label information corresponding to the initial label, and a claim form image, and using a pre-trained character recognition model to recognize the claim form image to obtain a keyword label and a text keyword of the claim form image; Generating keyword label information corresponding to the keyword label based on a label position and a text position corresponding to the keyword label and the text keyword in the claim form image; Obtaining a preset adversarial neural network, obtaining location information to which the loss settlement object belongs and training data corresponding to the location information, selecting preset feature data from the training data based on a preset feature keyword, training the adversarial neural network based on the preset feature data to obtain a feature coefficient model; Inputting the initial label, the initial label information, the keyword label and the keyword label information into a preset general coefficient model and the feature coefficient model respectively to calculate a general coefficient of a preset loss item and a feature coefficient of the preset loss item; the general coefficient comprises a general disability coefficient, a general base and a general month, and the feature coefficient comprises a feature base, a feature month and a feature disability coefficient; the general coefficient model comprises a plurality of preset general disability grades and a general corresponding relationship corresponding to each preset general disability grade, and the general corresponding relationship corresponding to each preset general disability grade comprises a general base label and a calculation method of a general base corresponding to the general base label, a general month label and a calculation method of a general month corresponding to the general month label; Adjusting the general coefficient based on the feature coefficient to obtain a target coefficient of the preset loss item, and correcting the target coefficient based on a preset loss settlement standard library to obtain a loss settlement coefficient of the preset loss item; Calculating claim data of the loss settlement object according to a plurality of preset loss items and a loss settlement coefficient corresponding to each preset loss item.

2. The method of claim 1, wherein, The pre-trained character recognition model comprises a recurrent sequence generation network and a decoding network, and the use of the pre-trained character recognition model to recognize the claim form image to obtain the keyword label and the text keyword of the claim form image comprises: Positioning a position of text information in the claim form image based on a pixel value of a pixel point in the claim form image to obtain a text position; Performing feature extraction on the text information based on the text position to obtain a feature sequence; Inputting the feature sequence into the recurrent sequence generation network to obtain a recurrent sequence; Decoding the recurrent sequence based on the decoding network to obtain text information of each loss settlement image; Determining a first preset keyword in the text information as the keyword label and determining a second preset keyword in the text information as the text keyword.

3. The method of claim 2, wherein, The generation of the keyword label information corresponding to the keyword label based on the label position and the text position corresponding to the keyword label and the text keyword in the claim form image comprises: detecting whether a label rectangular region corresponding to the keyword label and a text rectangular region corresponding to the text keyword intersect according to the label position and the text position; if the label rectangular region intersects with the text rectangular region, calculating an intersection region of the label rectangular region and each intersected text rectangular region; calculating a first area ratio of the intersection region on the label rectangular region and a second area ratio of the intersection region on the text rectangular region; selecting a text rectangular region with both the first area ratio and the second area ratio greater than a preset threshold as a target rectangular region, and generating keyword label information corresponding to the keyword label based on a quantity of the target rectangular regions and text keywords corresponding to the target rectangular regions.

4. The loss assessment and claim method of claim 3 wherein, The generating keyword label information corresponding to the keyword label based on the quantity of the target rectangular regions and the text keywords corresponding to the target rectangular regions comprises: if the quantity of the target rectangular regions is single, determining a text keyword corresponding to the target rectangular region as the keyword label information corresponding to the keyword label; or if the quantity of the target rectangular regions is multiple, performing weighted sum operation on each first area ratio and corresponding second area ratio to obtain a final score value of each target rectangular region, and selecting a text keyword corresponding to a target rectangular region with a maximum final score value as the keyword label information corresponding to the keyword label.

5. The method of claim 1, wherein, The inputting the initial label, the initial label information, the keyword label and the keyword label information into a preset universal coefficient model to calculate a preset loss term universal coefficient comprises: selecting a disability label from the initial label and the keyword label based on a preset disability keyword, and taking a value corresponding to the disability label in the claim form image as an initial disability grade of the loss-determining object; determining a universal proportion coefficient corresponding to a preset universal disability grade same as the initial disability grade in the universal coefficient model as the universal disability coefficient, and determining a universal corresponding relationship corresponding to the preset universal disability grade in the universal coefficient model as a target corresponding relationship; identifying an initial base label corresponding to a universal base label in the target corresponding relationship and an initial month label corresponding to a universal month label in the target corresponding relationship from the initial label and the keyword label; calculating the universal base based on an initial base value of the initial base label and a universal base calculation mode in the target corresponding relationship, and calculating the universal month based on an initial month value of the initial month label and a universal month calculation mode in the target corresponding relationship.

6. The method of claim 1, wherein, The adjusting the universal coefficient based on the feature coefficient to obtain a target coefficient of the preset loss term comprises: detecting whether the feature coefficient corresponds to each universal coefficient; if the feature coefficient does not correspond to each universal coefficient, determining the universal coefficient as the target coefficient; or If there is at least one feature coefficient corresponding to the general coefficient, the general coefficient is replaced by the corresponding feature coefficient, and the replaced general coefficient is determined as the target coefficient.

7. A loss adjusting and claims handling apparatus for implementing the loss adjusting and claims handling method according to any one of claims 1 to 6, characterized in that The loss settlement and claim device comprises: An acquisition unit is configured to acquire an initial label of a loss settlement object, initial label information corresponding to the initial label, and a claim form image, and use a pre-trained character recognition model to recognize the claim form image to obtain a keyword label and a text keyword of the claim form image. A generation unit is configured to generate keyword label information corresponding to the keyword label based on a label position and a text position corresponding to the keyword label and the text keyword in the claim form image. An input unit is configured to input the initial label, the initial label information, the keyword label, and the keyword label information into a pre-set general coefficient model and a pre-set feature coefficient model to calculate a general coefficient of a pre-set loss item and a feature coefficient of the pre-set loss item. An adjustment unit is configured to adjust the general coefficient based on the feature coefficient to obtain a target coefficient of the pre-set loss item, and correct the target coefficient based on a pre-set loss settlement standard library to obtain a loss settlement coefficient of the pre-set loss item. A calculation unit is configured to calculate claim data of the loss settlement object according to a plurality of pre-set loss items and a loss settlement coefficient corresponding to each pre-set loss item.

8. An electronic device, comprising: The electronic device comprises: a memory storing computer readable instructions; and a processor executing the computer readable instructions stored in the memory to implement the loss settlement and claim method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer readable storage medium stores computer readable instructions, and the computer readable instructions are executed by a processor in an electronic device to implement the loss settlement and claim method according to any one of claims 1 to 6.

Images