Health risk assessment method and assessment system, computer equipment, storage medium

Abstract

The embodiment of the present application provides a health risk assessment method and assessment system, computer equipment, and storage medium, which belongs to the field of artificial intelligence technology. The method includes: obtaining a health report image of a target object; performing text recognition on the health report image using a pre-trained text recognition model to obtain a first text information; performing text matching on the first text information according to pre-set risk association dictionary information to obtain multiple groups of keyword data information; performing risk calculation based on multiple groups of keyword data information to obtain risk calculation results; performing risk assessment on the target object according to pre-set risk assessment rules and risk calculation results to obtain risk assessment results. The embodiment of the present application can improve the efficiency and accuracy of health risk assessment.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a health risk assessment method and assessment system, computer equipment, and storage medium. Background Art

[0002] At present, when reviewing the health status of insured customers, insurance companies will pay more and more attention to the assessment of health risks. When conducting health risk assessments on customers, underwriters mainly rely on personal experience and manual operations. In actual assessments, when using a risk calculator to conduct health risk assessments, a large amount of health report information needs to be input, which means that underwriters need to repeatedly review the customer's relevant information and then enter and fill it in, which reduces the efficiency of information assessment. At the same time, this assessment method is easily affected by human factors, resulting in a low accuracy rate in risk assessment. Therefore, how to improve the efficiency and accuracy of health risk assessment has become a technical problem that needs to be solved urgently. Summary of the invention

[0003] The main purpose of the embodiments of the present application is to propose a health risk assessment method and assessment system, computer equipment, and storage medium that can improve the efficiency and accuracy of health risk assessment.

[0004] To achieve the above objectives, a first aspect of an embodiment of the present application proposes a health risk assessment method, the method comprising:

[0005] Obtaining a health report image of a target subject;

[0006] Performing text recognition on the health report image using a pre-trained text recognition model to obtain first text information;

[0007] Performing text matching on the first text information according to preset risk association dictionary information to obtain multiple groups of keyword data information;

[0008] Perform risk calculation based on the multiple sets of keyword data information to obtain risk calculation results;

[0009] A risk assessment is performed on the target object according to the preset risk assessment rules and the risk calculation result to obtain a risk assessment result.

[0010] In some embodiments, the first text information is text-matched according to the preset risk association dictionary information to obtain multiple groups of keyword data information, including:

[0011] Extracting keywords from the first text information to obtain a plurality of initial text keyword information;

[0012] Determine corresponding target value information according to the first text information and each of the initial text keyword information;

[0013] Acquire preset risk association dictionary information, wherein the risk association dictionary information includes a plurality of candidate dimension information and a plurality of candidate keyword information corresponding to each of the candidate dimension information;

[0014] Perform similarity matching calculation based on the candidate keyword information and the initial text keyword information to determine target text keyword information that matches the initial text keyword information;

[0015] Determine target dimension information according to the risk association dictionary information and the target text keyword information;

[0016] A plurality of groups of keyword data information are constructed according to the initial text keyword information, the target numerical information and the target dimension information.

[0017] In some embodiments, extracting keywords from the first text information to obtain a plurality of initial text keyword information includes:

[0018] Performing part-of-speech tagging on the first text information to obtain second text information, where the second text information is tagged with part-of-speech information of each word in the first text information;

[0019] Keywords are extracted from the second text information according to the part-of-speech information to obtain a plurality of initial text keyword information.

[0020] In some embodiments, performing similarity matching calculation based on the candidate keyword information and the initial text keyword information to determine the target text keyword information matching the initial text keyword information includes:

[0021] Performing vectorization processing on the initial text keyword information to obtain a keyword vector;

[0022] Vectorize the candidate keyword information to obtain a candidate vector;

[0023] A similarity matching calculation is performed based on the keyword vector and the candidate vector to determine target text keyword information that matches the initial text keyword information.

[0024] In some embodiments, after performing risk assessment on the target object according to the preset risk assessment rule and the risk calculation result to obtain the risk assessment result, the method further includes:

[0025] Determining the health risk level of the target object according to the risk assessment result;

[0026] A risk treatment plan for the target object is determined according to the health risk level.

[0027] In some embodiments, the pre-trained text recognition model is trained by the following method:

[0028] Constructing a training sample set, wherein the training sample set includes a plurality of training samples, each training sample includes a historical report image and historical text data information;

[0029] Build an initial neural network model;

[0030] Inputting each of the historical report images into the initial neural network model for model training to obtain target text data information;

[0031] Determine a loss value according to the target text data information and the historical text data information;

[0032] The model parameters of the initial neural network model are adjusted according to the target text data information and the historical text data information, and the adjusted model is continuously trained based on the training sample set until the loss value meets the preset training end condition to obtain the text recognition model.

[0033] In some embodiments, the initial neural network model includes a convolution layer, a recurrent layer, and a transcription layer, and inputting each of the historical report images into the initial neural network model for model training to obtain target text data information includes:

[0034] Performing text line detection on each of the historical report images to obtain at least one set of text line coordinate information;

[0035] Performing feature extraction on the historical report image through the convolution layer and each group of text line coordinate information to obtain a convolution feature vector;

[0036] Performing vector label prediction on the convolution feature vector through the circulation layer to obtain a text sequence feature vector;

[0037] The text sequence feature vector is integrated through the transcription layer to obtain target text data information.

[0038] To achieve the above-mentioned purpose, a second aspect of an embodiment of the present application proposes a health risk assessment system, the system comprising:

[0039] An image acquisition module, used to acquire a health report image of a target object;

[0040] A text recognition module, used to perform text recognition on the health report image using a pre-trained text recognition model to obtain first text information;

[0041] A text matching module, used for performing text matching on the first text information according to preset risk association dictionary information to obtain multiple groups of keyword data information;

[0042] A risk calculation module, used to perform risk calculation based on the multiple sets of keyword data information to obtain risk calculation results;

[0043] The risk assessment module is used to perform risk assessment on the target object according to preset risk assessment rules and the risk calculation result to obtain a risk assessment result.

[0044] To achieve the above objective, a third aspect of an embodiment of the present application provides a computer device, including:

[0045] at least one memory;

[0046] at least one processor;

[0047] at least one computer program;

[0048] The at least one computer program is stored in the at least one memory, and the at least one processor executes the at least one computer program to implement the method described in the first aspect above.

[0049] To achieve the above-mentioned purpose, the fourth aspect of an embodiment of the present application proposes a storage medium, which is a computer-readable storage medium. The computer-readable storage medium stores a computer program, and the computer program is used to enable a computer to execute the method described in the first aspect above.

[0050] The embodiments of the present application propose a health risk assessment method and assessment system, computer equipment, and storage medium. By acquiring a health report image of a target object, in order to recognize the text information in the health report image, a pre-trained text recognition model is used to perform text recognition on the health report image to obtain first text information. In order to improve the efficiency of health risk assessment, text matching is performed on the first text information according to preset risk-related dictionary information to obtain multiple groups of keyword data information. In order to improve the accuracy of health risk assessment, risk calculation is performed based on multiple groups of keyword data information to obtain risk calculation results, and risk assessment is performed on the target object according to preset risk assessment rules and risk calculation results to obtain risk assessment results. The embodiments of the present application can improve the efficiency and accuracy of health risk assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 is a first flow chart of the health risk assessment method provided in an embodiment of the present application;

[0052] Figure 2 yes Figure 1 Flow chart of step S130 in FIG.

[0053] Figure 3 yes Figure 2 Flow chart of step S210 in FIG.

[0054] Figure 4 yes Figure 2 Flow chart of step S240 in FIG.

[0055] Figure 5 is a second flow chart of the health risk assessment method provided in an embodiment of the present application;

[0056] Figure 6 is a training flow chart of a text recognition model provided in an embodiment of the present application;

[0057] Figure 7 yes Figure 6 Flowchart of step S630 in FIG.

[0058] Figure 8 is a structural diagram of a health risk assessment system provided in an embodiment of the present application;

[0059] Fig. 9 It is a schematic diagram of the hardware structure of the computer device provided in the embodiment of the present application. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0061] It should be noted that, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification, claims and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0062] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0063] First, some nouns involved in this application are analyzed:

[0064] Artificial Intelligence (AI): It is a new technical science that studies and develops theories, methods, technologies and application systems for simulating, extending and expanding human intelligence. AI is a branch of computer science. AI attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a similar way to human intelligence. Research in this field includes robots, language recognition, image recognition, natural language processing and expert systems. AI can simulate the information process of human consciousness and thinking. AI is also a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0065] Natural Language Processing (NLP): It is the use of computers to process, understand and apply human languages ​​(such as Chinese, English, etc.). NLP is a branch of artificial intelligence and an interdisciplinary subject between computer science and linguistics. It is often referred to as computational linguistics. Natural language processing includes grammatical analysis, semantic analysis, and text understanding. Natural language processing is often used in technical fields such as machine translation, handwritten and printed character recognition, speech recognition and text-to-speech conversion, information retrieval, information extraction and filtering, text classification and clustering, public opinion analysis and opinion mining. It involves data mining related to language processing, machine learning, knowledge acquisition, knowledge engineering, artificial intelligence research, and linguistic research related to language computing.

[0066] Robotic Process Automation (RPA): is an intelligent software that achieves automation in the workflow by simulating and enhancing the interaction between humans and computers. RPA has the characteristics of little impact on the existing enterprise system, basically no coding, short implementation cycle, and friendliness to non-technical business personnel. RPA can not only simulate humans, but also use and integrate existing technologies such as optical character recognition, speech recognition, virtual assistants, advanced analysis, machine learning and artificial intelligence to achieve its automation goals. This technology can effectively improve service speed, reduce costs and release employee capabilities.

[0067] Euclidean distance: also known as Euclidean metric, refers to the real distance between two points in m-dimensional space, or the natural length of a vector. The Euclidean distance in two-dimensional and three-dimensional space is the actual distance between two points. Euclidean distance can reflect the absolute difference between individual values ​​and can be used for analysis that needs to reflect the difference from the numerical size of the dimension.

[0068] Convolutional Recurrent Neural Network (CRNN): It is a network structure that is mainly used for end-to-end recognition of text sequences of indefinite length. It does not require cutting individual characters first, but converts text recognition into a time-dependent sequence learning problem, that is, image-based sequence recognition.

[0069] Convolutional Neural Network (CNN): It is a feedforward neural network that performs convolution calculations. It is inspired by the biological receptive field mechanism and is translationally invariant. It maximizes the use of local information and retains planar structural information.

[0070] Long Short Term Memory (LSTM): It is a time recurrent neural network that not only solves the problem that RNN cannot handle long-distance dependencies, but also solves common problems in neural networks such as gradient explosion or gradient disappearance. It is very effective in processing sequence data.

[0071] Connectionist Temporal Classification (CTC): It is a loss function in the sequence labeling problem. Traditional sequence labeling algorithms require that the input and output symbols are completely aligned at each moment. CTC expands the label set and adds empty elements. After the sequence is labeled with the expanded label set, all predicted sequences that can be converted to real sequences through the mapping function are correct prediction results, that is, the predicted sequence can be obtained without data alignment processing.

[0072] At present, when reviewing the health status of insured customers, insurance companies will pay more and more attention to the assessment of health risks. When conducting health risk assessments on customers, underwriters mainly rely on personal experience and manual operations. In actual assessments, when using a risk calculator to conduct health risk assessments, a large amount of health report information needs to be input, which means that underwriters need to repeatedly review the customer's relevant information and then enter and fill it in, which reduces the efficiency of information assessment. At the same time, this assessment method is easily affected by human factors, resulting in a low accuracy rate in risk assessment. Therefore, how to improve the efficiency and accuracy of health risk assessment has become a technical problem that needs to be solved urgently.

[0073] Based on this, the embodiments of the present application provide a health risk assessment method and assessment system, a computer device, and a storage medium, which can improve the efficiency and accuracy of health risk assessment.

[0074] The health risk assessment method and assessment system, computer device, and storage medium provided in the embodiments of the present application are specifically illustrated through the following embodiments. First, a health risk assessment method in the embodiments of the present application is described.

[0075] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0076] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0077] The health risk assessment method provided in the embodiment of the present application relates to the field of artificial intelligence technology. The health risk assessment method provided in the embodiment of the present application can be applied to the terminal, can also be applied to the server side, and can also be software running in the terminal or the server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or it can be configured as a server cluster or a distributed system composed of multiple physical servers, and can also be configured to provide cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDN) and big data and artificial intelligence platforms and other basic cloud computing services cloud server; the software can be an application that implements the health risk assessment method, etc., but is not limited to the above forms.

[0078] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0079] See also Figure 1 , Figure 1 is an optional flow chart of the health risk assessment method provided in the embodiment of the present application. Figure 1 The method may specifically include but is not limited to steps S110 to S150.

[0080] Step S110, obtaining a health report image of the target object;

[0081] Step S120, performing text recognition on the health report image using a pre-trained text recognition model to obtain first text information;

[0082] Step S130, performing text matching on the first text information according to preset risk association dictionary information to obtain multiple groups of keyword data information;

[0083] Step S140, performing risk calculation based on the multiple sets of keyword data information to obtain a risk calculation result;

[0084] Step S150, performing risk assessment on the target object according to the preset risk assessment rules and risk calculation results to obtain a risk assessment result.

[0085] In step S110 of some embodiments, in order to implement health risk assessment of the target object, first, a health report image of the target object is obtained. Specifically, a document image such as a physical examination report or a checkup report sheet of the target object used for health risk assessment is obtained.

[0086] It should be noted that the health report image of the target object can be obtained by writing a web crawler or script program to crawl data in a targeted manner.

[0087] It should be noted that before step S110, when the physical examination report, inspection report and other materials uploaded by the target object for health risk assessment are in any non-image file format such as portable document format (PDF), electronic document, spreadsheet, etc., the non-image file materials uploaded by the target object can first be converted into an image to obtain the health report image of the target object.

[0088] It should be noted that in order to improve the accuracy of health risk assessment and avoid the influence of background information in the health report image on risk assessment, after obtaining the initial health report image of the target object, the initial image can be first captured to obtain the health report image to be identified.

[0089] In step S120 of some embodiments, in order to realize the recognition of text information in the health report image, a pre-trained text recognition model is used to perform text recognition on the acquired health report image to obtain the first text information. Specifically, the text recognition model is used to perform text recognition on the text area in the acquired health report image, and the recognized multiple text information are sorted to obtain the first text information corresponding to the health report image.

[0090] In step S130 of some embodiments, since the health information required to be associated with the risk judgment of some diseases in the health risk assessment is different, in order to improve the efficiency of the health risk assessment, the first text information is text matched according to the preset risk association dictionary information to obtain multiple groups of keyword data information.

[0091] It should be noted that during health risk assessment, the risk-associated dictionary information corresponding to the assessment of different disease risks is different. When it is necessary to conduct risk assessment of multiple different disease risks of the target object, the first text information can be matched separately according to the pre-set multiple risk-associated dictionary information to obtain multiple groups of keyword data information. The multiple groups of keyword data information are used to determine whether the target object has symptom information corresponding to the current disease risk.

[0092] In step S140 of some embodiments, in order to improve the accuracy of health risk assessment, risk calculation is performed based on multiple groups of keyword data information to obtain risk calculation results. At the same time, in order to avoid the impact of manually input keyword data information on the risk assessment results, the multiple groups of keyword data information obtained can be automatically filled into a pre-set risk calculator by means of RPA. For example, the health risk assessment method of the present application is applied to premium calculation in the insurance field. When a cardiovascular health risk assessment is required for a target object, the multiple groups of keyword data information obtained are automatically filled into the cardiovascular metabolic risk calculator according to a pre-set cardiovascular metabolic risk calculator to obtain the risk calculation result of the target object under the cardiovascular risk assessment, that is, the risk coefficient of the target object is obtained, and the risk coefficient is used to evaluate subsequent risk premium calculations.

[0093] In step S150 of some embodiments, after the risk calculation result is obtained, the target object is subjected to risk assessment according to the preset risk assessment rules and the risk calculation result to obtain the risk assessment result. It should be noted that the risk assessment rule is a risk assessment calculation rule set according to actual needs. For example, the health risk assessment method of the present application is applied to the premium calculation in the insurance field. When the target object needs to be subjected to risk assessment, the corresponding age interval is first determined according to the age of the target object, and then the corresponding risk assessment rule is determined. Then, the risk assessment calculation is performed according to the risk calculation result and the risk assessment rule to obtain the risk assessment result corresponding to the target object, which is used to characterize the health risk situation of the target object. In a specific embodiment, for example, three risk assessment intervals are set according to the age of the insured object, that is, the target object, specifically, the age of the insured object is divided into: greater than or equal to fifty years old, less than fifty years old and greater than or equal to forty years old, less than forty years old, and the corresponding risk assessment rules are set according to the three age division intervals.

[0094] It should be noted that in each specific implementation of the present application, when it comes to the need to perform relevant processing based on data related to user identity or characteristics such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first, and the collection, use, and processing of these data will comply with the relevant laws, regulations, and standards of the relevant countries and regions. In addition, when the embodiment of the present application needs to obtain the user's sensitive personal information, the user's separate permission or consent will be obtained through a pop-up window or by jumping to a confirmation page. After clearly obtaining the user's separate permission or consent, the necessary user-related data for the normal operation of the embodiment of the present application will be obtained.

[0095] See also Figure 2 , Figure 2is a flowchart of a specific method of step S130 in some embodiments of the present application. In some embodiments of the present application, step S130 specifically includes but is not limited to steps S210 to S260. Figure 2 These six steps are introduced in detail.

[0096] Step S210, extracting keywords from the first text information to obtain a plurality of initial text keyword information;

[0097] Step S220, determining corresponding target value information according to the first text information and each initial text keyword information;

[0098] Step S230, obtaining preset risk association dictionary information, the risk association dictionary information including a plurality of candidate dimension information and a plurality of candidate keyword information corresponding to each candidate dimension information;

[0099] Step S240, performing similarity matching calculation based on the candidate keyword information and the initial text keyword information to determine the target text keyword information that matches the initial text keyword information;

[0100] Step S250, determining target dimension information according to risk association dictionary information and target text keyword information;

[0101] Step S260, constructing multiple groups of keyword data information according to the initial text keyword information, target numerical information and target dimension information.

[0102] In step S210 and step S220 of some embodiments, in order to determine whether the identified first text information contains associated health information corresponding to the disease risk to be assessed, keyword extraction is performed on the first text information to obtain multiple initial text keyword information, and the corresponding target numerical information is determined based on the first text information and each initial text keyword information. Specifically, after obtaining multiple initial text keyword information in the first text information, each initial text keyword information is used as a benchmark, and the first group of numerical parts on the right side are used as the target numerical information of the initial text keyword information. It should be noted that when there is no numerical value on the right side of the initial text keyword information, the numerical part starting from the next line or the numerical part on the lower side of the initial text keyword information can be set as the target numerical information according to actual conditions, and this is not limited here.

[0103] In step S230 of some embodiments, in order to determine whether the extracted initial text keyword information is the associated health information corresponding to the disease risk to be evaluated, the preset risk association dictionary information is obtained, and the risk association dictionary information includes multiple candidate dimension information and multiple candidate keyword information corresponding to each candidate dimension information. Specifically, since the health information associated with different disease risks is different when evaluating different disease risks, the risk association dictionary information corresponding to different disease risks is also different. The risk association dictionary information includes multiple candidate dimension information and multiple candidate keyword information corresponding to each candidate dimension information. The candidate dimension information and the candidate keyword information are used to determine whether the extracted initial text keyword information is the associated health information corresponding to the disease risk to be evaluated.

[0104] It should be noted that the risk-related dictionary information includes multiple candidate dimension information and multiple candidate keyword information corresponding to each candidate dimension information. For example, when a health risk assessment is performed on cardiovascular disease risk, the health information associated with the cardiovascular disease risk includes six dimension information, and the corresponding risk-related dictionary information includes six candidate dimension information, specifically including the target object's basic information, physical information, blood pressure information, lipid information, glucose metabolism information and liver function information. Among them, each candidate dimension information corresponds to multiple candidate keyword information. Specifically, the target object's basic information includes age, gender, region and other information, physical information includes height, weight, waist circumference and other information, blood pressure information includes systolic pressure, diastolic pressure and other information, lipid information includes total cholesterol, high-density lipoprotein, triglycerides and other information, glucose metabolism information includes glycated hemoglobin, plasma glucose, diabetes status and other information, liver function information includes serum aspartate aminotransferase, serum alanine aminotransferase, serum glutamyl transferase, hepatitis B status, hepatitis C status and other information.

[0105] In steps S240 to S260 of some embodiments, in order to determine the target dimension information corresponding to each initial text keyword information, a similarity matching calculation is performed based on the candidate keyword information and the initial text keyword information to determine the target text keyword information that matches the initial text keyword information. And the candidate dimension information corresponding to the target text keyword information is determined based on the risk association dictionary information, and the candidate dimension information is the target dimension information corresponding to the initial text keyword information. In order to improve the evaluation efficiency of health risk assessment, multiple groups of keyword data information are constructed based on the initial text keyword information, the target numerical information and the target dimension information, that is, keyword data information in the form of "target dimension information-initial text keyword information-target numerical information" is obtained.

[0106] See also Figure 3 , Figure 3is a flowchart of a specific method of step S210 in some embodiments of the present application. In some embodiments of the present application, step S210 specifically includes but is not limited to step S310 and step S320. Figure 3 These two steps are introduced in detail.

[0107] Step S310, performing part-of-speech tagging on the first text information to obtain second text information, where the second text information tags the part-of-speech information of each word in the first text information;

[0108] Step S320: extract keywords from the second text information according to the part-of-speech information to obtain a plurality of initial text keyword information.

[0109] In step S310 of some embodiments, in order to more accurately extract keywords related to health risk assessment from the first text information, the first text information is tagged with parts of speech to obtain second text information, and the second text information marks the part of speech information of each word in the first text information. Specifically, in the first text information corresponding to the target object, since the description of the disease may contain some descriptions of time, disease severity, etc. that are not related to the disease name, thereby affecting the accuracy of keyword extraction, the first text information is tagged with parts of speech using a part of speech tagging algorithm according to the position rules set for different parts of speech to obtain the second text information. For example, the name of the disease is a noun, the description of the degree of the disease is an adverb, and the numerical information corresponding to the disease name is a quantifier.

[0110] It should be noted that the part-of-speech tagging algorithm may adopt a hidden Markov model (HMM) or a conditional random field (CRFs), etc., which is not limited here.

[0111] In step S320 of some embodiments, keywords are extracted from the second text information according to the part-of-speech information to obtain multiple initial text keyword information. Specifically, after the part-of-speech information of each word in the first text information is tagged, keywords are extracted from the second text information according to the part-of-speech information to obtain multiple initial text keyword information, and the initial text key information is a keyword related to health risk assessment. For example, when a text in the first text information is "suffering from severe X2 years", by performing part-of-speech tagging on the text, it is obtained that "severe" is an adverb of degree, "X" is a noun, and "2 years" is a quantifier, then "X" is the name of the disease to be extracted, that is, after keyword extraction, the initial text keyword information "X" is obtained.

[0112] See also Figure 4 , Figure 4is a flowchart of a specific method of step S240 in some embodiments of the present application. In some embodiments of the present application, step S240 specifically includes but is not limited to steps S410 to S430. Figure 4 These three steps are introduced in detail.

[0113] Step S410, vectorizing the initial text keyword information to obtain a keyword vector;

[0114] Step S420, vectorizing the candidate keyword information to obtain a candidate vector;

[0115] Step S430 , performing similarity matching calculation based on the keyword vector and the candidate vector to determine the target text keyword information that matches the initial text keyword information.

[0116] In step S410 and step S420 of some embodiments, in order to more accurately compare the similarity between the candidate keyword information and the initial text keyword information, the initial text keyword information is vectorized to obtain a keyword vector, and the candidate keyword information is vectorized to obtain a candidate vector, that is, the candidate keyword information and the initial text keyword information are vectorized to improve the text similarity matching calculation. Specifically, the initial text keyword information is vectorized. First, an initial text dictionary can be constructed based on the initial text keyword information. For example, the initial text keywords are "height" and "blood pressure". When the constructed initial text dictionary is set to sequential numbering, {"body": 1, "height": 2, "blood": 3, "pressure": 4}, then "height" and "blood pressure" can be represented as [1,1,0,0] and [0,0,1,1] after vectorization.

[0117] In step S430 of some embodiments, after all the initial text keyword information and candidate keyword information are vectorized, the obtained keyword vector and candidate vector are subjected to similarity matching calculation to determine the target text keyword information that matches the initial text keyword information. Specifically, the keyword vector and the candidate vector can be subjected to similarity matching calculation using the Euclidean distance calculation method to obtain a similarity calculation result. In order to ensure the accuracy of the similarity matching, the candidate keyword information corresponding to the candidate vector with the largest similarity calculation result is selected as the target text keyword information, and then, according to the risk association dictionary information, the candidate dimension information corresponding to the candidate keyword information is the target dimension information.

[0118] It should be noted that the embodiment of the present application may also use vector calculation methods such as cosine similarity calculation and Manhattan distance calculation to perform similarity matching calculation on keyword vectors and candidate vectors, which is not limited here.

[0119] See also Figure 5 , Figure 5 is another optional flow chart of the health risk assessment method provided in the embodiment of the present application. In some embodiments of the present application, after step S150, the health risk assessment method specifically further includes but is not limited to step S510 and step S520. Figure 5 These two steps are introduced in detail.

[0120] Step S510, determining the health risk level of the target object according to the risk assessment result;

[0121] Step S520, determining a risk treatment plan for the target object according to the health risk level.

[0122] In step S510 and step S520 of some embodiments, in order to realize the application of health risk assessment in actual business scenarios, after the risk assessment result of the target object is obtained, the health risk level of the target object can be determined according to the risk assessment result. For example, when performing risk assessment, three risk assessment intervals are set according to the age of the insured object, that is, the target object, specifically, the age of the insured object is divided into: greater than or equal to fifty years old, less than fifty years old and greater than or equal to forty years old, and less than forty years old, and the corresponding risk assessment rules are set according to the three age division intervals, and the three risk scoring rules set correspond to the three health risk levels of the target object. Determine the risk treatment plan of the target object according to the health risk level. In a specific embodiment, the health risk assessment method of the present application is applied to the premium calculation in the insurance field. When the target object needs to be evaluated for cardiovascular health risks, the health risk level of the target object is determined according to the risk assessment results and the insured age of the target object, so that the risk premium calculation plan of the target object can be determined according to the health risk level. The embodiment of the present application can improve the efficiency and accuracy of health risk assessment.

[0123] See also Figure 6 , Figure 6 The training flow chart of the pre-trained text recognition model provided in the embodiment of the present application. In some embodiments of the present application, the training process of the text recognition model specifically includes but is not limited to steps S610 to S650. Figure 6 These five steps are introduced in detail.

[0124] Step S610, constructing a training sample set, the training sample set includes a plurality of training samples, each training sample includes a historical report image and historical text data information;

[0125] Step S620, constructing an initial neural network model;

[0126] Step S630, inputting each historical report image into the initial neural network model for model training to obtain target text data information;

[0127] Step S640, determining the loss value according to the target text data information and the historical text data information;

[0128] Step S650, adjusting the model parameters of the initial neural network model according to the target text data information and the historical text data information, and continuing to train the adjusted model based on the training sample set until the loss value meets the preset training end condition to obtain a text recognition model.

[0129] In steps S610 to S650 of some embodiments, in order to obtain a text recognition model, first, a training sample set is constructed, the training sample set includes multiple training samples, and each training sample includes a historical report image and historical text data information. By constructing an initial neural network model, each historical report image is input into the initial neural network model for model training to obtain the target text data information. In order to improve the accuracy of the obtained text recognition model for recognizing health report images, a loss value is determined according to the target text data information and the historical text data information. Specifically, the loss function preset in the initial neural network model is used to calculate the loss of the target text data information and the historical text data information to obtain the loss value. The parameters of the initial neural network model are adjusted according to the loss value, and the target text data information obtained after the parameter adjustment is continuously calculated until the loss value obtained according to the target text data information and the historical text data information can meet the preset training end condition, and the text recognition model and the target model parameters corresponding to the text recognition model are obtained.

[0130] It should be noted that the loss function can be any one of the 0-1 loss function, absolute value loss function, logarithmic loss function, exponential loss function, Hinge loss function, etc., which can be selected and adjusted according to actual needs and will not be repeated here.

[0131] It should be noted that the preset training end condition may be when the loss value is less than a preset loss value threshold, or when the recognition accuracy of the text recognition model is greater than or equal to a preset accuracy threshold.

[0132] See also Figure 7 , Figure 7 Flowchart of the specific method of step S630 in some embodiments of the present application. In some embodiments of the present application, the initial neural network model includes a convolutional layer, a recurrent layer and a transcription layer. Step S630 specifically includes but is not limited to steps S710 to S740. Figure 7 These four steps are introduced in detail.

[0133] Step S710, performing text line detection on each historical report image to obtain at least one set of text line coordinate information;

[0134] Step S720, extracting features of the historical report image through the convolution layer and each set of text line coordinate information to obtain a convolution feature vector;

[0135] Step S730, performing vector label prediction on the convolution feature vector through a recurrent layer to obtain a text sequence feature vector;

[0136] Step S740, integrating the text sequence feature vectors through the transcription layer to obtain target text data information.

[0137] In steps S710 to S740 of some embodiments, in order to improve the recognition accuracy of the text recognition model, firstly, each historical report image is subjected to text line detection to obtain at least one set of text line coordinate information, and an initial neural network model based on the CRNN network structure is used to perform text recognition on each historical report image. Specifically, the initial neural network model includes a convolutional layer, a circulation layer, and a transcription layer. The historical report image is feature extracted by the CNN of the convolutional layer and each set of text line coordinate information to obtain a convolutional feature vector, thereby achieving feature dimension reduction and reduction of the amount of calculation for the corresponding area in the historical report image. Then, the convolutional feature vector is vector-label predicted by the LSTM of the circulation layer to obtain a text sequence feature vector. Finally, the text sequence feature vector is feature-integrated by the CTC of the transcription layer to obtain the target text data information, which can solve the problem that characters cannot be aligned during training and improve the recognition accuracy of the model. The embodiment of the present application obtains a text recognition model by combining CRNN and CTC training, and uses a pre-trained text recognition model to perform text recognition on health report images, which can improve the accuracy of health risk assessment.

[0138] The health risk assessment method provided in the embodiment of the present application obtains the health report image of the target object, and in order to realize the recognition of the text information in the health report image, the pre-trained text recognition model is used to perform text recognition on the health report image to obtain the first text information. In order to improve the efficiency of health risk assessment, the first text information is tagged with parts of speech to obtain the second text information, and the second text information marks the part of speech information of each word in the first text information, and the keywords in the second text information are extracted according to the part of speech information to obtain multiple initial text keyword information, and then the corresponding target numerical information is determined according to the first text information and each initial text keyword information. Similarity matching calculation is performed according to the candidate keyword information and the initial text keyword information to determine the target text keyword information matching the initial text keyword information, and the target dimension information is determined according to the preset risk association dictionary information and the target text keyword information, and then multiple groups of keyword data information are constructed according to the initial text keyword information, the target numerical information and the target dimension information. In order to improve the accuracy of health risk assessment, risk calculation is performed according to multiple groups of keyword data information to obtain risk calculation results, and risk assessment is performed on the target object according to the preset risk assessment rules and risk calculation results to obtain risk assessment results. In practical applications, the health risk level of the target object is determined according to the risk assessment results, and the risk treatment plan of the target object is determined according to the health risk level. The embodiment of the present application obtains a text recognition model by combining CRNN and CTC training, and uses a pre-trained text recognition model to perform text recognition on health report images. At the same time, by combining RPA to avoid the influence of manually input keyword data information on risk assessment results, the efficiency and accuracy of health risk assessment are improved.

[0139] See also Figure 8 The embodiment of the present application also provides a health risk assessment system that can implement the above-mentioned health risk assessment method. The system includes an image acquisition module 810, a text recognition module 820, a text matching module 830, a risk calculation module 840 and a risk assessment module 850.

[0140] An image acquisition module 810 is used to acquire a health report image of a target object;

[0141] A text recognition module 820 is used to perform text recognition on the health report image using a pre-trained text recognition model to obtain first text information;

[0142] A text matching module 830 is used to perform text matching on the first text information according to preset risk association dictionary information to obtain multiple groups of keyword data information;

[0143] The risk calculation module 840 is used to perform risk calculation based on multiple sets of keyword data information to obtain risk calculation results;

[0144] The risk assessment module 850 is used to perform risk assessment on the target object according to preset risk assessment rules and risk calculation results to obtain risk assessment results.

[0145] It should be noted that the health risk assessment system of the embodiment of the present application is used to implement the above-mentioned health risk assessment method. The health risk assessment system of the embodiment of the present application corresponds to the aforementioned health risk assessment method. Please refer to the aforementioned health risk assessment method for the specific processing process, which will not be repeated here.

[0146] The embodiment of the present application also provides a computer device, which includes: at least one memory, at least one processor, and at least one computer program, wherein the at least one computer program is stored in the at least one memory, and the at least one processor executes the at least one computer program to implement any of the health risk assessment methods in the above embodiments. The computer device can be any intelligent terminal including a tablet computer, a vehicle-mounted computer, etc.

[0147] See also Fig. 9 , Fig. 9 The hardware structure of a computer device according to another embodiment is illustrated. The computer device includes:

[0148] The processor 910 may be implemented by a general-purpose central processing unit (CPU), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0149] The memory 920 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 920 can store an operating system and other application programs. When the technical solution provided in the embodiment of this specification is implemented by software or firmware, the relevant program code is stored in the memory 920, and the processor 910 calls and executes the health risk assessment method of the embodiment of this application;

[0150] Input / output interface 930, used to implement information input and output;

[0151] Communication interface 940, used to realize communication interaction between the device and other devices, which can be realized through wired mode (such as USB, network cable, etc.) or wireless mode (such as mobile network, WIFI, Bluetooth, etc.);

[0152] bus 950 , which transmits information between the various components of the device (e.g., processor 910 , memory 920 , input / output interface 930 , and communication interface 940 );

[0153] The processor 910 , the memory 920 , the input / output interface 930 , and the communication interface 940 are connected to each other in communication within the device via a bus 950 .

[0154] An embodiment of the present application also provides a storage medium, which is a computer-readable storage medium. The computer-readable storage medium stores a computer program, and the computer program is used to enable a computer to execute any one of the health risk assessment methods in the above embodiments.

[0155] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0156] The embodiments described in the embodiments of the present application are intended to more clearly illustrate the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0157] It can be understood by those skilled in the art that Figures 1 to 7 The technical solutions shown in the figure do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than those shown in the figure, or a combination of certain steps, or different steps.

[0158] The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0159] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0160] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0161] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0162] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0163] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0164] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0165] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store programs.

[0166] The preferred embodiments of the present invention are described with reference to the accompanying drawings, but the scope of the present invention is not limited by the above. Any modification, equivalent substitution and improvement made by a person skilled in the art without departing from the scope and essence of the present invention should be within the scope of the present invention.

Claims

1. A health risk assessment method, characterized in that: The method comprises: Obtaining a health report image of a target subject; Performing text recognition on the health report image using a pre-trained text recognition model to obtain first text information; Performing text matching on the first text information according to preset risk-related dictionary information to obtain multiple groups of keyword data information; performing text matching on the first text information according to preset risk-related dictionary information to obtain multiple groups of keyword data information includes: performing keyword extraction on the first text information to obtain multiple initial text keyword information; determining corresponding target numerical information according to the first text information and each of the initial text keyword information; obtaining preset risk-related dictionary information, the risk-related dictionary information including multiple candidate dimension information and multiple candidate keyword information corresponding to each of the candidate dimension information; performing similarity matching calculation according to the candidate keyword information and the initial text keyword information to determine the target text keyword information that matches the initial text keyword information; determining the target dimension information according to the risk-related dictionary information and the target text keyword information; constructing multiple groups of keyword data information according to the initial text keyword information, the target numerical information and the target dimension information; Perform risk calculation based on the multiple sets of keyword data information to obtain risk calculation results; A risk assessment is performed on the target object according to the preset risk assessment rules and the risk calculation result to obtain a risk assessment result.

2. The method according to claim 1, characterized in that The keyword extraction is performed on the first text information to obtain a plurality of initial text keyword information, including: Performing part-of-speech tagging on the first text information to obtain second text information, where the second text information is tagged with part-of-speech information of each word in the first text information; Keywords are extracted from the second text information according to the part-of-speech information to obtain a plurality of initial text keyword information.

3. The method according to claim 1, characterized in that The performing similarity matching calculation based on the candidate keyword information and the initial text keyword information to determine the target text keyword information matching the initial text keyword information includes: Performing vectorization processing on the initial text keyword information to obtain a keyword vector; Vectorize the candidate keyword information to obtain a candidate vector; A similarity matching calculation is performed based on the keyword vector and the candidate vector to determine target text keyword information that matches the initial text keyword information.

4. The method according to claim 1, characterized in that After performing risk assessment on the target object according to the preset risk assessment rule and the risk calculation result to obtain the risk assessment result, the method further includes: Determining the health risk level of the target object according to the risk assessment result; A risk treatment plan for the target object is determined according to the health risk level.

5. The method according to any one of claims 1 to 4, characterized in that: The pre-trained text recognition model is trained by the following method: Constructing a training sample set, wherein the training sample set includes a plurality of training samples, each training sample includes a historical report image and historical text data information; Build an initial neural network model; Inputting each of the historical report images into the initial neural network model for model training to obtain target text data information; Determine a loss value according to the target text data information and the historical text data information; The model parameters of the initial neural network model are adjusted according to the target text data information and the historical text data information, and the adjusted model is continuously trained based on the training sample set until the loss value meets the preset training end condition to obtain the text recognition model.

6. The method according to claim 5, characterized in that The initial neural network model includes a convolution layer, a recurrent layer and a transcription layer. Each of the historical report images is input into the initial neural network model for model training to obtain target text data information, including: Performing text line detection on each of the historical report images to obtain at least one set of text line coordinate information; Performing feature extraction on the historical report image through the convolution layer and each group of text line coordinate information to obtain a convolution feature vector; Performing vector label prediction on the convolution feature vector through the circulation layer to obtain a text sequence feature vector; The text sequence feature vector is integrated through the transcription layer to obtain target text data information.

7. A health risk assessment system, characterized in that: The system comprises: An image acquisition module, used to acquire a health report image of a target object; A text recognition module, used to perform text recognition on the health report image using a pre-trained text recognition model to obtain first text information; A text matching module is used to perform text matching on the first text information according to preset risk-related dictionary information to obtain multiple groups of keyword data information; the text matching on the first text information according to the preset risk-related dictionary information to obtain multiple groups of keyword data information includes: extracting keywords from the first text information to obtain multiple initial text keyword information; determining corresponding target numerical information according to the first text information and each of the initial text keyword information; obtaining preset risk-related dictionary information, the risk-related dictionary information includes multiple candidate dimension information and multiple candidate keyword information corresponding to each of the candidate dimension information; performing similarity matching calculation according to the candidate keyword information and the initial text keyword information to determine the target text keyword information that matches the initial text keyword information; determining the target dimension information according to the risk-related dictionary information and the target text keyword information; constructing multiple groups of keyword data information according to the initial text keyword information, the target numerical information and the target dimension information; A risk calculation module, used to perform risk calculation based on the multiple sets of keyword data information to obtain risk calculation results; The risk assessment module is used to perform risk assessment on the target object according to preset risk assessment rules and the risk calculation result to obtain a risk assessment result.

8. A computer device, characterized in that: include: at least one memory; at least one processor; at least one computer program; The at least one computer program is stored in the at least one memory, and the at least one processor executes the at least one computer program to implement: The method according to any one of claims 1 to 6.

9. A storage medium, the storage medium being a computer-readable storage medium, characterized in that: The computer readable storage medium stores a computer program, and the computer program is used to cause a computer to execute: The method according to any one of claims 1 to 6.

Images