CN108877928B

CN108877928B - Patient information acquisition method and device, computer equipment and storage medium

Info

Publication number: CN108877928B
Application number: CN201810546481.4A
Authority: CN
Inventors: 励超磨; 魏海彬; 苟永亮; 翁志龙
Original assignee: Ping An Health Cloud Co Ltd; Ping An Healthcare Technology Co Ltd
Current assignee: Ping An Health Cloud Co Ltd; Ping An Healthcare Technology Co Ltd
Priority date: 2018-05-31
Filing date: 2018-05-31
Publication date: 2023-04-07
Anticipated expiration: 2038-05-31
Also published as: CN108877928A

Abstract

The application relates to a patient information acquisition method, a patient information acquisition device, computer equipment and a storage medium. The method comprises the following steps: extracting tongue picture attributes from the tongue picture image of the patient; acquiring a tongue picture definition matched with the tongue picture attribute; extracting symptom characteristics from the patient's chief complaint data and tongue picture paraphrases; searching an information collection template matched with the symptom characteristics, and loading template data corresponding to the information collection template; generating and outputting a node acquisition problem according to node data in the template data, and acquiring patient reply data corresponding to the node acquisition problem; and generating patient acquisition information according to the patient reply data. By adopting the method, the accuracy of the patient information recording can be improved.

Description

Patient information acquisition method and device, computer equipment and storage medium

Technical Field

The present application relates to the field of computer technologies, and in particular, to a method and an apparatus for acquiring patient information, a computer device, and a storage medium.

Background

When a doctor of traditional Chinese medicine asks a patient, the doctor needs to ask the patient and ask the patient for a plurality of physical symptoms, life habits and other problems.

When inquiring about the patient, the doctor of traditional Chinese medicine needs to collate and record the result of the inquiry and the information of the patient's answer. Therefore, when recording, doctors of traditional Chinese medicine often simply record key information which is felt to be important by themselves only by experience, and cannot accurately and comprehensively record the information of patients due to personal reasons such as the speed of speech, pronunciation and the like of the patients.

Disclosure of Invention

In view of the above, it is necessary to provide a patient information collecting method, an apparatus, a computer device and a storage medium capable of accurately recording patient information.

A method of patient information acquisition, the method comprising:

acquiring patient chief complaint data and a patient tongue picture image;

extracting tongue picture attributes from the tongue picture image of the patient;

acquiring a tongue picture definition matched with the tongue picture attribute;

extracting symptom characteristics from the patient complaint data and the tongue picture paraphrases;

searching an information collection template matched with the symptom characteristics, and loading template data corresponding to the information collection template;

generating and outputting a node acquisition problem according to node data in the template data, and acquiring patient reply data corresponding to the node acquisition problem;

and generating patient acquisition information according to the patient reply data.

In one embodiment, generating and outputting a node acquisition question according to node data in the template data, and acquiring patient response data corresponding to the node acquisition question includes:

acquiring an initial node in the template data and initial node data corresponding to the initial node;

generating and outputting an initial acquisition problem according to the initial node data, and acquiring initial patient reply data corresponding to the initial acquisition problem;

and acquiring a connecting node corresponding to the starting node from the template data, selecting a first node corresponding to the starting patient reply data from the connecting nodes, taking the first node as a current node, and continuously generating and outputting a current acquisition problem according to the current node data corresponding to the current node until the current node corresponding to the acquired current patient reply data is a template tip node.

In one embodiment, loading template data corresponding to the information collection template includes:

acquiring the template structure type of the information collection template;

when the template structure type is the nested template, loading the mother template data in the information collection template from a template database to a local cache;

before continuing to generate and output the current acquisition problem according to the current node data corresponding to the current node, the method further comprises the following steps:

judging whether the first node is a nested template skip node or not;

and when the first node is judged to be the nested template skip node, acquiring a nested sub-template to which the first node belongs, and loading sub-template data corresponding to the nested sub-template from a database to a local cache.

In one embodiment, a sub-template modification instruction and sub-template upgrading data are received;

reading a sub-template identifier in the sub-template modification instruction, and searching sub-template data to be upgraded corresponding to the sub-template identifier;

comparing the upgrading sub-template data with the sub-template data to be upgraded to generate template change data;

searching an author identification corresponding to the associated template corresponding to the sub-template identification;

and generating an associated template upgrading prompt according to the template change data, and sending the associated template upgrading prompt to the author terminal corresponding to the author identifier.

In one embodiment, obtaining a tongue definition matching the tongue attribute comprises:

inputting the tongue picture attributes into a preset neural network classifier to obtain a first paraphrase matching probability of each tongue picture paraphrase;

inputting the tongue picture attributes into a preset Bayes classifier to obtain second paraphrase matching probabilities of the tongue picture paraphrases;

and obtaining the paraphrase matching rate of each tongue picture paraphrase according to the first paraphrase matching probability and the second paraphrase matching probability, and extracting the tongue picture paraphrase with the highest paraphrase matching rate as the tongue picture paraphrase matched with the tongue picture attribute.

In one embodiment, after obtaining the tongue picture paraphrase matching the tongue picture attribute, the method further comprises:

acquiring doctor data corresponding to an online doctor, and extracting duty cycle data from the doctor data;

screening out idle doctors from the online doctors according to the duty cycle data;

sending the patient tongue picture image and the tongue picture paraphrase to a doctor terminal corresponding to the idle doctor;

the extracting of symptom features from the patient complaint data and the tongue picture paraphrases comprises:

and when receiving a paraphrase confirmation notice which is returned by the doctor terminal and corresponds to the tongue picture paraphrases, extracting symptom characteristics from the patient complaint data and the tongue picture paraphrases.

In one embodiment, the method further comprises the following steps:

when receiving a paraphrase confirmation notice and a corrected tongue picture paraphrase which are returned by the doctor terminal and correspond to the tongue picture paraphrase, extracting symptom characteristics from the patient complaint data and the corrected tongue picture paraphrase;

adding the tongue picture attribute and the corrected tongue picture paraphrase association to a tongue picture corrected sample set;

and adjusting the neuron weight of the preset neural network classifier and the probability distribution of the preset Bayes classifier according to the corrected sample set.

A patient information acquisition device, the device comprising:

the patient data acquisition module is used for acquiring patient chief complaint data and a patient tongue picture image;

the tongue picture attribute extraction module is used for extracting tongue picture attributes from the tongue picture image of the patient;

the paraphrase acquisition module is used for acquiring the tongue picture paraphrases matched with the tongue picture attributes;

the characteristic extraction module is used for extracting symptom characteristics from the patient complaint data and the tongue picture paraphrases;

the template searching module is used for searching an information collecting template matched with the symptom characteristics and loading template data corresponding to the information collecting template;

the data acquisition module is used for generating and outputting a node acquisition problem according to node data in the template data and acquiring patient reply data corresponding to the node acquisition problem;

and the information generation module is used for generating patient acquisition information according to the patient reply data.

A computer device comprising a memory storing a computer program and a processor implementing the steps of the above method when executing the computer program.

A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the above-mentioned method.

According to the patient information acquisition method, the patient information acquisition device, the computer equipment and the storage medium, after patient chief complaint data and a patient tongue picture are obtained, image identification and attribute extraction are carried out on the tongue picture, a tongue picture paraphrase matched with the extracted attribute is searched, an information collection template matched with the tongue picture paraphrase is searched according to symptom features contained in the patient chief complaint data and the tongue picture paraphrase, and corresponding patient information is acquired according to node data in the template data, so that automatic paraphrasing of the tongue picture can be realized, automatic acquisition of patient information is realized according to the information collection template, and comprehensive and accurate patient information can be obtained.

Drawings

FIG. 1 is a diagram illustrating an exemplary embodiment of a method for collecting patient information;

FIG. 2 is a schematic flow chart diagram of a method for patient information acquisition in one embodiment;

FIG. 3 is a flow diagram illustrating the paraphrase matching step in one embodiment;

FIG. 4 is a block diagram of a patient information acquisition device in one embodiment;

FIG. 5 is a diagram of the internal structure of a computer device in one embodiment.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.

The patient information acquisition method provided by the application can be applied to the application environment shown in fig. 1. Wherein the terminal communicates with the server via a network. The method comprises the steps that a terminal obtains patient chief complaint data and a patient tongue picture image, extracts tongue picture attributes from the patient tongue picture image, obtains tongue picture paraphrases matched with the tongue picture attributes, extracts symptom features from the patient chief complaint data and the tongue picture paraphrases, searches for information collection templates matched with the symptom features, sends a template data loading request to a server, the server sends template data corresponding to the information collection templates to the terminal according to the template data loading request, and the terminal generates and outputs a node collection problem according to node data in the received template data sent by the server to obtain patient reply data corresponding to the node collection problem; and generating patient acquisition information according to the patient reply data. The terminal can be, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers and portable wearable devices, and the server can be implemented by an independent server or a server cluster formed by a plurality of servers.

In one embodiment, as shown in fig. 2, a method for acquiring patient information is provided, which is described by taking the method as an example for being applied to the terminal in fig. 1, and includes the following steps:

and step 210, acquiring patient complaint data and tongue picture images of the patient.

The patient complaint data is description data of the patient on the health condition of the patient, the patient complaint data can comprise description data of the physical condition, the suffered disease and the suffering degree of the patient, the suffering symptoms and the severity of the symptoms, and the like of the patient, and the tongue image of the patient is an image of the tongue of the patient.

The terminal can prompt the user to input the chief complaint data, for example, the user can be prompted in a voice or text mode to "please describe your problems in detail, including physical conditions, diseases, symptoms and the like", the user can input the patient chief complaint data to the terminal through voice or text, and the terminal obtains the patient chief complaint data input by the user. The user can upload the pre-stored tongue image of the patient, and can also directly acquire the tongue image of the patient through the image acquisition equipment of the terminal.

Step 220, extracting tongue picture attributes from the tongue picture image of the patient.

The terminal performs preprocessing operation on the acquired tongue image of the patient, and the preprocessing operation may include operations of correcting image colors, extracting a tongue part from the image, performing tongue part segmentation, removing a shadow region of the image, and the like.

Further, the terminal detects the quality of the tongue image before preprocessing the tongue image, and judges whether the tongue image meets the image quality requirement according to the quality detection result. The quality detection of the tongue image can include detecting image parameter data such as resolution, contrast, brightness and the like of the tongue image, comparing the detected image parameter data with a preset standard parameter range of a corresponding parameter, judging whether the detected image parameter data is in the preset standard parameter range, and passing the quality detection of the image parameter data when the detected image parameter data is in the preset standard parameter range; and when the image parameters which are not in the preset standard parameter range exist, the quality detection of the image parameter data fails, a prompt with unqualified image quality is generated, and the prompt is displayed, so that the user can upload the tongue image of the patient again according to the prompt.

The attributes of the tongue image can include the color, the color and the quality of the tongue coating. Wherein, the specific attribute values of the tongue color attribute can include magenta, cyan, purple, dry white, white and so on; the specific attribute values of the tongue fur color attribute can comprise white tongue fur, yellow tongue fur, grey tongue fur, black tongue fur and the like; the tongue coating can be thick and thin, moist and dry, sticky, greasy, peeled, rooted or unrooted, and the like, and the specific attribute values of the tongue coating can comprise thick tongue coating, thin tongue coating, moist tongue coating, greasy tongue coating, and the like. And the terminal performs image analysis and feature detection on the preprocessed tongue image of the patient, and extracts corresponding tongue image attributes according to feature detection results.

Step 230, obtain the tongue picture paraphrases matching the tongue picture attributes.

The terminal stores a plurality of tongue picture definitions in advance, and the tongue picture definitions are analysis results of tongue picture images, such as 'red tongue, tooth marks at the edge and thick and greasy fur'. The attribute label marking is carried out on the tongue picture paraphrases stored in the terminal in advance, the terminal matches the extracted tongue picture attributes with the attribute labels of the tongue picture paraphrases, and the tongue picture paraphrases with the attribute labels completely consistent with the tongue picture attributes are found out.

Step 240, extracting symptom characteristics from the patient's chief complaint data and tongue picture paraphrases.

The symptom characteristics are used for representing physical state characteristics of the patient and can comprise symptom expression characteristics such as cough, excessive phlegm, cold sweat and the like, symptom degree characteristics such as severe pain, slight flatulence and the like, physical characteristics such as obesity, emaciation and tongue image attribute characteristics such as red tongue, pale tongue, thick and greasy coating and the like.

The terminal obtains a symptom characteristic word list, the symptom characteristic word list comprises symptom characteristic words summarized from a plurality of clinical case data and a Chinese and western medicine dictionary, and the similar words of all the symptom characteristic words are also added to the symptom characteristic word list in an associated mode. And the terminal extracts the symptom characteristics matched with the symptom characteristic words in the symptom characteristic word list from the patient main complaint data and the tongue picture paraphrases.

And step 250, searching an information collection template matched with the symptom characteristics, and loading template data corresponding to the information collection template.

The information collection template is a template which is prepared for a certain disease or symptom and is used for collecting the health condition of the patient. The information collection template carries out symptom characteristic labels in advance, and the symptom characteristic labels are set according to symptom characteristic words in the symptom characteristic word list. An information gathering template may be tagged with one or more symptom signatures, i.e., patient health information may be collected for one or more diseases or symptoms belonging to the same category. And the terminal matches the extracted symptom characteristics with marked symptom characteristic labels of all the information collecting templates, and if the marked symptom characteristic labels completely contain the extracted symptom characteristics, the matching is successful. And the terminal acquires the template identification of the successfully matched information collection template, the template identification is used for uniquely identifying the information collection template, and the terminal searches and loads template data corresponding to the template identification. If the symptom characteristics extracted by the terminal comprise phlegm dampness, qi deficiency, obesity and pale tongue, and the matched information collection template corresponding to the extracted symptom characteristics is 'universal adult spleen qi deficiency constitution pale tongue'.

And step 260, generating and outputting a node acquisition problem according to the node data in the template data, and acquiring the patient reply data corresponding to the node acquisition problem.

The data in the information collection template comprises node data of a plurality of information collection nodes and connection relations among the plurality of information collection nodes, and each information collection node is used for collecting accompanying performance information of a certain health characteristic of a patient. The information collection nodes are connected with each other, some information collection nodes positioned at the upstream of the template can be connected with a plurality of information collection nodes positioned at the downstream of the template, and the patient information collected by the upstream information collection nodes determines the trend selection of the downstream information collection nodes. Therefore, a plurality of information acquisition paths can be formed in the information acquisition template according to the specific patient information acquired by each information acquisition node.

The node data of each information collecting node can include, but is not limited to, a node name, a collecting problem title, a collecting problem type, collecting problem content, a node connection condition relation and the like. The collection problem type may include a single-term or multi-term selection type problem, or may be an open-type problem or the like. The terminal generates the current acquisition problem according to the acquisition problem content in the current node data, and displays the generated current acquisition problem, such as through various forms of characters, pictures or voice. After receiving the current collected questions, the user can input question answers to the terminal, the terminal obtains the question answers input by the user, and patient reply data is generated according to the question answers.

Step 270, generating patient collection information according to the patient reply data.

And the terminal performs data integration and data processing on the node information of each node passing through the template execution path and the correspondingly acquired patient reply data and generates patient acquisition information.

In the patient information acquisition method, after the terminal acquires the patient chief complaint data and the tongue picture image of the patient, the tongue picture image is subjected to image identification and attribute extraction, a tongue picture paraphrase matched with the extracted attribute is searched, an information collection template matched with the tongue picture paraphrase is searched according to symptom features contained in the patient chief complaint data and the tongue picture paraphrase, and corresponding patient information is acquired according to node data in the template data, so that automatic paraphrasing of the tongue picture image can be realized, automatic acquisition of the patient information is realized according to the information collection template, and comprehensive and accurate patient information can be obtained.

In one embodiment, generating and outputting the node acquisition question according to the node data in the template data, and acquiring the patient reply data corresponding to the node acquisition question may include: acquiring an initial node in the template data and initial node data corresponding to the initial node; generating and outputting an initial acquisition problem according to the initial node data, and acquiring initial patient reply data corresponding to the initial acquisition problem; and acquiring a connecting node corresponding to the starting node from the template data, selecting a first node corresponding to the starting patient reply data from the connecting nodes, taking the first node as a current node, and continuing to generate and output a current acquisition problem according to the current node data corresponding to the current node until the current node corresponding to the acquired starting patient reply data is a template tip node.

The initial node is the first information collection node in the information collection template, the terminal obtains node data corresponding to the initial node, the terminal extracts node acquisition problem data from the node data, generates an acquisition problem according to the node acquisition problem data, and outputs the generated acquisition problem.

For example, a node name of an information collecting node is "night sweat or not", and for a one-item selection type question, the corresponding question content may include a question title "do you sweat frequently? ", the question options are" make "and" not ". The node connection condition relation comprises a node selection condition judgment logic, wherein the judgment logic is used for judging whether the collected data of the upstream connection node meets the selection regulation of the node or not, if so, the node is executed, and if not, the execution is refused. If the upstream node with the node name of sleep dreaminess condition is 'night sweat or not', the selection condition judgment logic of the node with the sleep dreaminess condition is met when the information acquired by the node 'night sweat or not' is 'yes'.

After receiving the collected questions, the user can input question answers to the terminal, the terminal obtains the question answers input by the user, and patient reply data is generated according to the question answers. When the answer to the question is the option, the answer selected by the user can be directly used as the patient reply data, when the answer to the question is the open answer of the user, the terminal extracts or converts the information of the answer to the question, specific extraction or conversion rules are set in advance according to specific questions of the node, and the terminal uses the extracted information as the patient reply data. If the answer to the question collected by the node of sleep time is sleep for 6 hours, the terminal extracts the number 6 from the answer as the data replied by the patient.

The method comprises the steps that a terminal obtains a connecting node which is located at the downstream of an initial node and has a connection relation with the initial node from template data, whether collected patient reply data of the initial node meets a node selection condition of the connecting node or not is judged, a first node meeting the node selection condition is found from the connecting node and is used as a current node, the current collection problem is generated and output according to current node data corresponding to the current node in a continuous and circulating mode, the patient reply data corresponding to the current collection problem is collected and obtained until the current node corresponding to the collected patient reply data is a template tip node, and the template tip node is an information collection node without the downstream connecting node in an information collection template.

In one embodiment, loading template data corresponding to the information collection template may include: acquiring the template structure type of an information collection template; when the template structure type is the nested template, loading the mother template data in the information collection template from the template database to a local cache; before continuing to generate and output the current acquisition problem according to the current node data corresponding to the current node, the method may further include: judging whether the first node is a nested template skip node or not; and when the first node is judged to be the nested template skip node, acquiring a nested sub-template to which the first node belongs, and loading sub-template data corresponding to the nested sub-template from the database to a local cache.

The template structure types of the information collection template comprise an independent template and a nested template, wherein the independent template is formed by a single template, and the nested template is formed by mutually connecting or nesting a mother template and a plurality of nested sub templates. The nested sub-template can be regarded as a node in the mother template, the template name of the nested sub-template is adopted to name the corresponding node in the mother template, and the connecting node between different templates is a nested template skip node.

And the terminal acquires the template structure type of the searched information collecting template, and loads the mother template data in the information collecting template from the server where the template database is located when the template structure type is judged to be the nested template, wherein the mother template data only comprises identification data such as node names corresponding to the nested sub-templates and does not comprise specific template data of the specific nested sub-templates.

When the terminal collects information according to the nested template and before continuing to collect patient reply data corresponding to the current node, whether the first node is a nested template skip node or not is judged, when the first node is judged to be the nested template skip node, the terminal obtains a first node identifier of the first node, searches for a nested sub-template corresponding to the first node identifier, and loads sub-template data corresponding to the searched nested sub-template into a local cache from a server where the template database is located.

In one embodiment, before loading the mother template data and the nested sub-template data, the terminal searches whether the template data exists in the mother template identifier or the template data corresponding to the nested sub-template identifier from a local cache, and if the template data exists, the template data does not need to be loaded from the database, and if the template data does not exist, the template data is loaded from the template database. And the terminal regularly counts the execution frequency of the template data loaded in the local cache, deletes the template data with the execution frequency lower than the preset frequency from the cache and removes the local redundant data.

The information of patients needing to be collected for some diseases or symptoms is very much, the number of nodes and the data amount of node data included in the corresponding information collection template are very huge, some information collection templates may contain dozens of layers of hundreds of nodes and dozens of nested sub-templates, if all template data are loaded into a local cache, a large amount of storage space is occupied, the processing efficiency is affected, and only the template data of a few sub-templates may be executed according to the selection path of data replied by a user, so that data redundancy and storage space waste are caused. Therefore, in the embodiment, only when jumping to the related nested sub-template, the data of the sub-template is loaded from the database to the local cache, so that data redundancy can be reduced, and the processing efficiency can be improved.

In one embodiment, the patient information acquisition method may further include: receiving a sub-template modification instruction and upgrading sub-template data; reading a sub-template identifier in the sub-template modification instruction, and searching sub-template data to be upgraded corresponding to the sub-template identifier; comparing the upgrading sub-template data with the sub-template data to be upgraded to generate template change data; searching author identification corresponding to the associated template corresponding to the sub-template identification; and generating an associated template upgrading prompt according to the template change data, and sending the associated template upgrading prompt to the author terminal corresponding to the author identifier.

The user can modify the template data of each information collection template through the terminal, and the modification of the template data comprises the modification of node data of each node, such as node names, node problems, node titles and the like, or the deletion, addition and other operations of the node, or the modification of node attributes of the node, such as node connection relation, or the modification of other template data. And when the template data of the nested sub-template is modified and confirmed by a user, generating a sub-template modification instruction, wherein the sub-template modification instruction carries the template identification of the modified nested sub-template.

And the terminal receives the sub-template modification instruction and acquires the updated sub-template data modified by the sub-template data by the user. And the terminal reads the sub-template identification in the sub-template modification instruction, searches the sub-template data to be upgraded corresponding to the sub-template identification, and the sub-template data to be upgraded is the template data of the initial sub-template corresponding to the upgrading sub-template. And the terminal compares the upgrading sub-template data with the sub-template data to be upgraded, finds out the difference data between the upgrading sub-template data and the sub-template data to be upgraded, and generates template change data according to the difference data.

And the terminal searches for an associated template corresponding to the sub-template identifier, wherein the associated template is an information collection template of the sub-template to be upgraded corresponding to the sub-template identifier contained in the template data, namely the sub-template to be upgraded is nested in the associated template. And the terminal acquires the author identification corresponding to the searched associated template, wherein the author identification is used for identifying the user identity of the user creating the associated template. The number of the associated templates corresponding to the sub-template identifier searched by the terminal may be one or more, and thus, the number of the corresponding author identifiers may also be one or more.

The terminal generates a related template upgrading prompt according to the template changing data, the terminal can add a link of the template changing data in the related template upgrading prompt, and a user can check the changing condition of the sub-template data by clicking the link. The prompt terms in the associated template upgrade prompt may be "XX child template for upgrade, do it need to upgrade the parent template? "and the like. And the terminal sends the generated associated template upgrading prompt to an author terminal corresponding to the author identifier so that a creation user of the associated template can determine whether to upgrade the associated template.

In one embodiment, as shown in FIG. 3, the paraphrase matching step of obtaining tongue picture paraphrases matching the tongue picture attributes may include:

step 232, inputting the tongue picture attributes into a preset neural network classifier to obtain a first paraphrase matching probability of each tongue picture paraphrase.

All possible values corresponding to the tongue picture attributes are collected in the server in advance, and all possible values are assigned to a binarization vector. The server searches the extracted binary vectors corresponding to the tongue picture attributes, and generates a feature vector combination according to all the obtained binary vectors, so that the tongue picture attributes are converted into digital vectors from text data.

If the tongue color attribute is the tongue color attribute, the feature vector corresponding to the red tongue is 00000010, the feature vector corresponding to the purple tongue is 00000001, and if the tongue color attribute is the tongue coating color attribute, the feature vector corresponding to the gray coating is 00001001, and the feature vector corresponding to the black coating is 00001010.

The first paraphrase matching probability is based on the matching probability of the feature vector calculated by the preset neural network classifier and all preset tongue picture paraphrases. The preset neural network classifier is a neural network classification model which is trained by a large number of samples in advance, and the neural network classification model is used for classifying tongue picture attributes by tongue picture paraphrasing.

Specifically, the sample data used for training the neural network classification model includes feature vectors corresponding to tongue image attributes of each inquiry case and a correct tongue image definition corresponding to the tongue image of the patient, and the tongue image definition can be a binary vector code of the tongue image definition, and the like. The preset neural network classifier takes the feature vector combination corresponding to the tongue picture attribute extracted by the terminal as the input of the classifier, the matching probability of all preset tongue picture paraphrases and the input tongue picture attribute combination is output, and the matching probability is used for reflecting the matching degree of the tongue picture image of the patient and each preset tongue picture paraphrase.

In one embodiment, the generation method of the preset neural network classifier may include: constructing an initial neural network model, collecting historical diagnosis and treatment sample data, and extracting tongue picture attributes and corresponding tongue picture definitions from the historical diagnosis and treatment sample data; setting the tongue picture attribute as input data of an initial neural network model, setting the tongue picture definition as target data of an output layer of the initial neural network model, and training and adjusting initial weighting weights among neurons in a hidden layer of the initial neural network model to obtain optimal weighting weights; and generating a preset neural network classifier according to the initial neural network model and the optimal weighting weight.

The constructed initial neural network model is composed of an input layer, an output layer and a plurality of hidden layers. The hidden layer is each layer composed of a plurality of neurons and links between the input layer and the output layer, each layer in the hidden layer may include a plurality of convolution layers, pooling layers, connection layers, dropout layers, and the like, the number of layers in each layer in the hidden layer, the type of the adopted activation function, and the initial weighting weight between each neuron may be set by a worker according to experience, wherein the activation function adopted in the hidden layer may include a sigmoid function, a ReLu function, a tanh function, and the like.

And step 234, inputting the tongue picture attributes into a preset Bayes classifier to obtain a second paraphrase matching probability of each tongue picture paraphrase.

The second paraphrase matching probability is the matching probability of the feature vector and all the preset tongue picture paraphrases calculated according to the preset Bayesian classifier. The preset Bayes classifier is a Bayes classification model which is trained by a large number of samples in advance, and the Bayes classification model is used for carrying out tongue picture paraphrase classification on the tongue picture image of the patient.

Specifically, the sample data used for training the bayesian classification model includes feature vectors corresponding to the tongue picture attributes of each inquiry case and a correct tongue picture definition corresponding to the tongue picture image of the patient, and the tongue picture definition can be a binary vector code of the tongue picture definition, and the like. The preset neural network classifier takes the feature vector combination corresponding to the tongue picture attribute extracted by the terminal as the input of the classifier, the matching probability of all preset tongue picture paraphrases and the input feature vector combination is output, and the matching probability is used for reflecting the matching degree of the tongue picture image of the patient and each preset tongue picture paraphrase.

In one embodiment, the generation method of the preset bayesian classifier can comprise the following steps: constructing an initial Bayesian classification model, collecting historical diagnosis and treatment sample data, and extracting tongue picture attributes and corresponding tongue picture definitions from the historical diagnosis and treatment sample data; extracting tongue picture attributes from the diagnosis and treatment data of the patient and generating attribute feature vector combinations according to the tongue picture attributes; counting the tongue picture paraphrase probability distribution of each attribute feature vector combination according to the tongue picture paraphrases; and generating a preset Bayesian classifier according to the initial Bayesian classification model and the tongue picture paraphrase probability distribution. In this embodiment, a naive bayes probability model can be employed to construct the initial bayes classification model, such as a polynomial naive bayes classification model, a bernoulli naive bayes classification model, and the like.

Step 236, obtaining the paraphrase matching rate of each tongue paraphrase according to the first paraphrase matching probability and the second paraphrase matching probability, and extracting the tongue paraphrase with the highest paraphrase matching rate as the tongue paraphrase matched with the tongue attribute.

And the terminal performs comprehensive analysis on the first paraphrase matching probability of each tongue picture paraphrase calculated according to a preset neural network classifier and the second paraphrase matching probability of each tongue picture paraphrase calculated according to a preset Bayes classifier according to the result of the analysis, obtains the comprehensive matching probability of each tongue picture paraphrase according to the result of the analysis, and screens out the tongue picture paraphrase corresponding to the comprehensive matching probability with the highest value.

In one embodiment, deriving a paraphrase matching ratio for each tongue paraphrase based on the first paraphrase matching probability and the second paraphrase matching probability may comprise: acquiring a first initial weight corresponding to the first paraphrase matching probability and a second initial weight corresponding to the second paraphrase matching probability; and calculating the paraphrase weighted matching probability according to the first paraphrase matching probability, the corresponding first initial weight, the second paraphrase matching probability and the corresponding second initial weight.

The terminal sets initial probability weights in advance for a first paraphrase matching probability calculated according to a preset neural network model and a second paraphrase matching probability calculated according to a preset Bayes classifier, and the initial probability weights are set as a first initial weight and a second initial weight respectively. The first initial weight and the second initial weight are set according to the accuracy of the classification result of the preset neural network classifier on the historical sample data and the accuracy of the classification result of the preset Bayes classifier on the historical sample data. Specifically, the accuracy of the classification result may be a statistical result of the accuracy of all sample data obtained by different classifiers, at this time, the first initial weights corresponding to all tongue definitions may be set to the same value, the second initial weights corresponding to all tongue definitions may also be set to the same value, and the sum of the first initial weight and the second initial weight is 1.

Furthermore, the accuracy of the classification result of the two classifiers classifying the historical sample data corresponding to each tongue picture paraphrase can be counted, and the first initial weight and the second initial weight of each paraphrase are respectively set according to the calculated accuracy of each paraphrase, that is, the values of the first initial weights set by different paraphrases may be different, the values of the second initial weights set by different paraphrases may be different, but the sum of the first initial weight and the second initial weight corresponding to the same paraphrase is 1.

In one embodiment, after obtaining the tongue definition matching the tongue attribute, the method may further include: acquiring doctor data corresponding to an online doctor, and extracting duty cycle data from the doctor data; screening out idle doctors from online doctors according to the duty cycle data; sending the tongue picture image and the tongue picture paraphrase of the patient to a doctor terminal corresponding to an idle doctor; extracting symptom characteristics from patient complaint data and tongue picture paraphrases, comprising: and when a paraphrase confirmation notice corresponding to the tongue picture paraphrases returned by the doctor terminal is received, extracting symptom characteristics from the patient complaint data and the tongue picture paraphrases.

And after acquiring the tongue picture paraphrases matched with the tongue picture attributes, the terminal acquires corresponding doctor data of the current online doctor. The online doctor is a doctor who can provide online remote inquiry at present, and the doctor data can comprise doctor information of a hospital where the doctor belongs, department information of the doctor, doctor-performing experience information, current duty cycle information and the like.

The terminal extracts the busy degree data from the acquired doctor data, wherein the busy degree data is data capable of reflecting the current inquiry busy degree of a doctor, and specifically includes data such as the number of currently queued patients and the average inquiry time of the doctor to the patients, the terminal can evaluate the expected waiting time of the doctor according to the number of the currently queued patients and the average inquiry time of the doctor, and calculate the corresponding busy degree according to the expected waiting time of each doctor, generally, the corresponding busy degree with longer expected waiting time is higher, and the corresponding busy degree with lower expected waiting time is lower. A high duty cycle indicates that the doctor is busy, and a low duty cycle indicates that the doctor is idle.

The terminal screens the doctor in the current visit with the lowest duty cycle as an idle doctor, and acquires a doctor terminal identifier corresponding to the idle doctor, wherein the doctor terminal identifier is used for uniquely identifying the terminal used by the doctor in the current visit. The terminal sends the tongue picture image and the tongue picture paraphrase together to a doctor terminal corresponding to an idle doctor. The doctor in the doctor terminal can check and analyze the received tongue picture image, judge whether the tongue picture paraphrase is consistent with the tongue picture image uploaded by the user, when the doctor judges that the tongue picture paraphrase is consistent with the tongue picture image, the doctor can confirm the tongue picture paraphrase through the doctor terminal, and when the doctor terminal detects that the doctor confirms the tongue picture paraphrase, a paraphrase confirmation notice is generated and returned to the terminal. And after the terminal receives the paraphrase confirmation notice returned by the doctor terminal, continuing to execute the step of extracting the symptom characteristics from the patient's chief complaint data and the tongue picture paraphrases.

In one embodiment, the terminal can set a doctor feedback time threshold, calculate the predicted feedback time of the doctor according to the current time and the doctor feedback time threshold, and when the predicted feedback time of the doctor is reached and no paraphrase feedback sent by the doctor terminal is received, produce a paraphrase confirmation reminder and send the paraphrase confirmation reminder to the doctor terminal to prompt the doctor to preferentially process a paraphrase confirmation task.

In this embodiment, after the matched tongue picture paraphrases are found by the terminal, the matched tongue picture paraphrases are sent to the doctor for confirmation, and the doctor performs the following intelligent inquiry operation after confirmation, so that the problem that the time is wasted by the patient due to invalid information acquisition caused by paraphrase matching errors is avoided.

In one embodiment, the patient information acquisition method may further include: when receiving a paraphrase confirmation notice and a corrected tongue picture paraphrase which are returned by a doctor terminal and correspond to the tongue picture paraphrase, extracting symptom characteristics from the patient's chief complaint data and the corrected tongue picture paraphrase; adding the tongue picture attribute and the corrected tongue picture paraphrase into a tongue picture correction sample set in a correlated manner; and adjusting the neuron weight of the preset neural network classifier and the probability distribution of the preset Bayes classifier according to the corrected sample set.

When the doctor judges that the tongue picture paraphrases do not accord with the tongue picture image, the doctor can modify the received tongue picture paraphrases through the doctor terminal, and after the doctor confirms the input modified tongue picture paraphrases and the doctor terminal detects the modification operation of the doctor, the modified tongue picture paraphrases of the doctor are obtained and a paraphrase confirmation notice is generated, and the paraphrase confirmation notice and the modified tongue picture paraphrases are returned to the terminal.

And after the terminal receives the paraphrase confirmation notice and the corrected tongue picture paraphrases, extracting symptom features from the patient main complaint data and the corrected tongue picture paraphrases, and continuously executing the step of searching an information collection template matched with the symptom features.

And after the terminal correlates the corrected tongue picture paraphrases with the tongue picture attributes, adding the corrected tongue picture paraphrases into the tongue picture correction sample set. The terminal periodically extracts tongue picture correction sample data from the tongue picture correction sample set, extracts a preset number of tongue picture attributes and corresponding corrected tongue picture paraphrases from the tongue picture correction sample data, uses the extracted tongue picture attributes as input data of a preset neural network classifier, outputs the corrected tongue picture paraphrases as targets of the preset neural network classifier, and continuously adjusts and optimizes the neuron weight of neurons in a hidden layer of the preset neural network classifier. And the terminal re-counts the matching probability of each preset tongue picture definition and the feature vector combination of the tongue picture attribute according to the tongue picture correction sample data and the historical sample data, and adjusts the probability distribution of the preset Bayes classifier.

It should be understood that although the various steps in the flow charts of fig. 2-3 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 2-3 may include multiple sub-steps or multiple stages that are not necessarily performed at the same time, but may be performed at different times, and the order of performance of the sub-steps or stages is not necessarily sequential, but may be performed in turn or alternating with other steps or at least some of the sub-steps or stages of other steps.

In one embodiment, as shown in fig. 4, there is provided a patient information acquisition apparatus including: a patient data acquisition module 410, a tongue image attribute extraction module 420, a paraphrase acquisition module 430, a feature extraction module 440, a template lookup module 450, a data acquisition module 460, and an information generation module 470, wherein:

a patient data acquiring module 410 for acquiring the patient complaint data and the tongue image of the patient.

A tongue attribute extracting module 420, configured to extract tongue attributes from the tongue image of the patient.

And a paraphrase acquiring module 430 for acquiring the tongue picture paraphrases matched with the tongue picture attributes.

And the feature extraction module 440 is used for extracting symptom features from the patient chief complaint data and the tongue picture paraphrases.

The template searching module 450 is used for searching an information collecting template matched with the symptom characteristics and loading template data corresponding to the information collecting template; .

And the data acquisition module 460 is configured to generate and output a node acquisition problem according to the node data in the template data, and acquire patient response data corresponding to the node acquisition problem.

And an information generating module 470, configured to generate the patient collecting information according to the patient reply data.

In one embodiment, the data collection module 460 may include:

and the current node extraction module is used for acquiring the initial node in the template data and the initial node data corresponding to the initial node.

And the reply data acquisition module is used for generating and outputting an initial acquisition problem according to the initial node data and acquiring initial patient reply data corresponding to the initial acquisition problem.

And the template execution module is used for acquiring a connecting node corresponding to the starting node from the template data, selecting a first node corresponding to the starting patient reply data from the connecting nodes, taking the first node as a current node, and continuously generating and outputting a current acquisition problem according to the current node data corresponding to the current node until the current node corresponding to the acquired current patient reply data is a template tip node.

In one embodiment, the template lookup module 450 may include:

and the structure type acquisition module is used for acquiring the template structure type of the information collection template.

And the master template loading module is used for loading the master template data in the information collection template from the template database to the local cache when the template structure type is the nested template.

The patient information acquisition apparatus may further include:

and the skip node judging module is used for judging whether the first node is a nested template skip node.

And the sub-template loading module is used for acquiring the nested sub-template to which the first node belongs and loading the sub-template data corresponding to the nested sub-template from the database to the local cache when the first node is judged to be the nested template skip node.

In one embodiment, the patient information acquisition apparatus may further include:

and the modification instruction receiving module is used for receiving the sub-template modification instruction and the upgrading sub-template data.

And the sub-template data searching module is used for reading the sub-template identification in the sub-template modification instruction and searching the sub-template data to be upgraded corresponding to the sub-template identification.

And the change data generation module is used for comparing the upgrade sub-template data with the sub-template data to be upgraded to generate template change data.

And the author identification searching module is used for searching the author identification corresponding to the associated template corresponding to the sub-template identification.

And the upgrade prompt generation module is used for generating an associated template upgrade prompt according to the template change data and sending the associated template upgrade prompt to the author terminal corresponding to the author identifier.

In one embodiment, the paraphrase acquisition module 430 may include:

and the first probability obtaining module is used for inputting the tongue picture attributes into a preset neural network classifier to obtain the first paraphrase matching probability of each tongue picture paraphrase.

And the second probability obtaining module is used for inputting the tongue picture attributes into a preset Bayes classifier to obtain second paraphrase matching probability of each tongue picture paraphrase.

And the paraphrase extraction module is used for obtaining the paraphrase matching rate of each tongue paraphrase according to the first paraphrase matching probability and the second paraphrase matching probability, and extracting the tongue paraphrase with the highest paraphrase matching rate as the tongue paraphrase matched with the tongue attribute.

and the doctor data acquisition module is used for acquiring doctor data corresponding to the online doctor and extracting the duty cycle data from the doctor data.

And the idle doctor screening module is used for screening idle doctors from online doctors according to the duty cycle data.

And the tongue image data sending module is used for sending the tongue image and the tongue picture definition of the patient to a doctor terminal corresponding to an idle doctor.

The feature extraction module 440 is further configured to extract symptom features from the patient complaint data and the tongue picture paraphrases when receiving a paraphrase confirmation notification corresponding to the tongue picture paraphrases returned by the doctor terminal.

and the corrected paraphrase receiving module is used for extracting symptom characteristics from the patient chief complaint data and the corrected tongue picture paraphrases when receiving paraphrase confirmation notice and the corrected tongue picture paraphrases which are returned by the doctor terminal and correspond to the tongue picture paraphrases.

And the tongue picture data adding module is used for adding the tongue picture attribute and the corrected tongue picture paraphrase into the tongue picture corrected sample set in an associated manner.

And the classifier optimization module is used for adjusting the neuron weight of the preset neural network classifier and the probability distribution of the preset Bayesian classifier according to the corrected sample set.

For the specific definition of the patient information collecting device, reference may be made to the above definition of the patient information collecting method, which is not described herein again. The various modules in the patient information acquisition device described above may be implemented in whole or in part by software, hardware, and combinations thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.

In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as shown in fig. 5. The computer device includes a processor, a memory, a network interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a patient information acquisition method. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on a shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.

It will be appreciated by those skilled in the art that the configuration shown in fig. 5 is a block diagram of only a portion of the configuration associated with the present application, and is not intended to limit the computing device to which the present application may be applied, and that a particular computing device may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.

In one embodiment, there is provided a computer device comprising a memory storing a computer program and a processor implementing the following steps when the processor executes the computer program: acquiring patient complaint data and a patient tongue picture image; extracting tongue picture attributes from the tongue picture image of the patient; acquiring a tongue picture definition matched with the tongue picture attribute; extracting symptom characteristics from the patient's chief complaint data and tongue picture paraphrases; searching an information collection template matched with the symptom characteristics, and loading template data corresponding to the information collection template; generating and outputting a node acquisition problem according to node data in the template data, and acquiring patient reply data corresponding to the node acquisition problem; and generating patient acquisition information according to the patient reply data.

In one embodiment, the processor executes the computer program to generate and output a node acquisition question according to the node data in the template data, and the step of acquiring the patient reply data corresponding to the node acquisition question is further configured to: acquiring an initial node in the template data and initial node data corresponding to the initial node, generating a node acquisition problem according to the node data in the template data and outputting the node acquisition problem, wherein acquiring patient response data corresponding to the node acquisition problem may include: acquiring an initial node in the template data and initial node data corresponding to the initial node; generating and outputting an initial acquisition problem according to the initial node data, and acquiring initial patient reply data corresponding to the initial acquisition problem; and acquiring a connecting node corresponding to the starting node from the template data, selecting a first node corresponding to the starting patient reply data from the connecting nodes, taking the first node as a current node, and continuing to generate and output a current acquisition problem according to the current node data corresponding to the current node until the current node corresponding to the acquired starting patient reply data is a template tip node.

In one embodiment, the processor, when executing the computer program, further performs the step of loading template data corresponding to the information collection template to: acquiring a template structure type of an information collection template; when the template structure type is the nested template, loading the mother template data in the information collection template from the template database to a local cache; the following steps are also implemented: judging whether the first node is a nested template skip node or not; and when the first node is judged to be the nested template skip node, acquiring a nested sub-template to which the first node belongs, and loading sub-template data corresponding to the nested sub-template from the database to a local cache.

In one embodiment, the processor when executing the computer program further performs the steps of: receiving a sub-template modification instruction and upgrading sub-template data; reading a sub-template identifier in the sub-template modification instruction, and searching sub-template data to be upgraded corresponding to the sub-template identifier; comparing the upgrading sub-template data with the sub-template data to be upgraded to generate template change data; searching author identification corresponding to the associated template corresponding to the sub-template identification; and generating an associated template upgrading prompt according to the template change data, and sending the associated template upgrading prompt to the author terminal corresponding to the author identifier.

In one embodiment, the processor when executing the computer program to perform the step of obtaining a tongue paraphrase that matches the tongue attribute is further operable to: inputting the tongue picture attribute into a preset neural network classifier to obtain a first paraphrase matching probability of each tongue picture paraphrase; inputting the tongue picture attribute into a preset Bayes classifier to obtain a second paraphrase matching probability of each tongue picture paraphrase; and obtaining the paraphrase matching rate of each tongue picture paraphrase according to the first paraphrase matching probability and the second paraphrase matching probability, and extracting the tongue picture paraphrase with the highest paraphrase matching rate as the tongue picture paraphrase matched with the tongue picture attribute.

In one embodiment, the processor, when executing the computer program, further performs the steps of: acquiring doctor data corresponding to an online doctor, and extracting duty cycle data from the doctor data; screening out idle doctors from online doctors according to the duty cycle data; sending the tongue picture image and the tongue picture paraphrase of the patient to a doctor terminal corresponding to an idle doctor; the step of extracting the symptom characteristics from the patient complaint data and the tongue picture paraphrases is also used for extracting the symptom characteristics from the patient complaint data and the tongue picture paraphrases when receiving a paraphrase confirmation notice which is returned by the doctor terminal and corresponds to the tongue picture paraphrases.

In one embodiment, the processor, when executing the computer program, further performs the steps of: when receiving a paraphrase confirmation notice and a corrected tongue picture paraphrase which are returned by a doctor terminal and correspond to the tongue picture paraphrase, extracting symptom characteristics from the patient's chief complaint data and the corrected tongue picture paraphrase; adding the tongue picture attribute and the corrected tongue picture paraphrase into a tongue picture correction sample set in a correlated manner; and adjusting the neuron weight of the preset neural network classifier and the probability distribution of the preset Bayes classifier according to the corrected sample set.

In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of: acquiring patient complaint data and a patient tongue picture image; extracting tongue picture attributes from the tongue picture image of the patient; acquiring a tongue picture definition matched with the tongue picture attribute; extracting symptom characteristics from the patient's chief complaint data and tongue picture paraphrases; searching an information collection template matched with the symptom characteristics, and loading template data corresponding to the information collection template; generating and outputting a node acquisition problem according to node data in the template data, and acquiring patient reply data corresponding to the node acquisition problem; and generating patient acquisition information according to the patient reply data.

In one embodiment, the computer program is executed by the processor to generate and output a node acquisition question according to the node data in the template data, and the step of acquiring the patient reply data corresponding to the node acquisition question is further configured to: acquiring a start node in the template data and start node data corresponding to the start node, generating a node acquisition problem according to the node data in the template data and outputting the node acquisition problem, wherein acquiring patient reply data corresponding to the node acquisition problem may include: acquiring an initial node in the template data and initial node data corresponding to the initial node; generating and outputting an initial acquisition problem according to the initial node data, and acquiring initial patient reply data corresponding to the initial acquisition problem; and acquiring a connecting node corresponding to the starting node from the template data, selecting a first node corresponding to the starting patient reply data from the connecting nodes, taking the first node as a current node, and continuing to generate and output a current acquisition problem according to the current node data corresponding to the current node until the current node corresponding to the acquired starting patient reply data is a template tip node.

In one embodiment, the computer program when executed by the processor further performs the step of loading template data corresponding to the information collection template for: acquiring the template structure type of an information collection template; when the template structure type is the nested template, loading the mother template data in the information collection template from the template database to a local cache; the following steps are also implemented: judging whether the first node is a nested template skip node or not; and when the first node is judged to be the nested template skip node, acquiring a nested sub-template to which the first node belongs, and loading sub-template data corresponding to the nested sub-template from the database to a local cache.

In one embodiment, the computer program when executed by the processor further performs the steps of: receiving a sub-template modification instruction and upgrading sub-template data; reading a sub-template identifier in the sub-template modification instruction, and searching sub-template data to be upgraded corresponding to the sub-template identifier; comparing the upgrading sub-template data with the sub-template data to be upgraded to generate template change data; searching author identification corresponding to the associated template corresponding to the sub-template identification; and generating an associated template upgrading prompt according to the template change data, and sending the associated template upgrading prompt to an author terminal corresponding to the author identifier.

In one embodiment, the computer program when executed by the processor is further operable to perform the step of obtaining a tongue paraphrase that matches tongue attributes: inputting the tongue picture attributes into a preset neural network classifier to obtain a first paraphrase matching probability of each tongue picture paraphrase; inputting the tongue picture attributes into a preset Bayes classifier to obtain second paraphrase matching probability of each tongue picture paraphrase; and obtaining the paraphrase matching rate of each tongue picture paraphrase according to the first paraphrase matching probability and the second paraphrase matching probability, and extracting the tongue picture paraphrase with the highest paraphrase matching rate as the tongue picture paraphrase matched with the tongue picture attribute.

In one embodiment, the computer program when executed by the processor further performs the steps of: acquiring doctor data corresponding to an online doctor, and extracting duty cycle data from the doctor data; screening out idle doctors from online doctors according to the duty cycle data; sending the tongue picture image and the tongue picture paraphrase of the patient to a doctor terminal corresponding to an idle doctor; the step of extracting the symptom characteristics from the patient complaint data and the tongue picture paraphrases is also used for extracting the symptom characteristics from the patient complaint data and the tongue picture paraphrases when receiving a paraphrase confirmation notice which is returned by the doctor terminal and corresponds to the tongue picture paraphrases.

In one embodiment, the computer program when executed by the processor further performs the steps of: when receiving a paraphrase confirmation notice and a corrected tongue picture paraphrase which are returned by a doctor terminal and correspond to the tongue picture paraphrase, extracting symptom characteristics from patient chief complaint data and the corrected tongue picture paraphrase; adding the tongue picture attribute and the corrected tongue picture paraphrase into a tongue picture correction sample set in a correlated manner; and adjusting the neuron weight of the preset neural network classifier and the probability distribution of the preset Bayes classifier according to the corrected sample set.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), rambus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

All possible combinations of the technical features in the above embodiments may not be described for the sake of brevity, but should be considered as being within the scope of the present disclosure as long as there is no contradiction between the combinations of the technical features.

The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.

Claims

1. A method of patient information acquisition, the method comprising:

acquiring patient complaint data and a patient tongue picture image;

extracting symptom characteristics from the patient's chief complaint data and the tongue picture paraphrases;

generating patient acquisition information according to the patient reply data;

the acquiring of the tongue picture paraphrases matched with the tongue picture attributes comprises the following steps:

inputting the tongue picture attributes into a preset neural network classifier to obtain a first paraphrase matching probability of each tongue picture paraphrase; the preset neural network classifier is used for carrying out tongue picture paraphrase classification on the tongue picture attributes;

inputting the tongue picture attributes into a preset Bayes classifier to obtain second paraphrase matching probabilities of the tongue picture paraphrases; the preset Bayes classifier is used for carrying out tongue interpretation classification on the tongue picture image of the patient;

2. The method according to claim 1, wherein generating and outputting a node acquisition question according to the node data in the template data, and acquiring the patient reply data corresponding to the node acquisition question comprises:

and acquiring a connecting node corresponding to the starting node from the template data, selecting a first node corresponding to the starting patient reply data from the connecting nodes, taking the first node as a current node, and continuing to generate and output a current acquisition problem according to the current node data corresponding to the current node until the current node corresponding to the acquired current patient reply data is a template tip node.

3. The method according to claim 2, wherein the loading the template data corresponding to the information collecting template comprises:

acquiring the template structure type of the information collection template;

before continuing to generate and output the current acquisition problem according to the current node data corresponding to the current node, the method further comprises:

judging whether the first node is a nested template skip node or not;

4. The method of claim 3, further comprising:

receiving a sub-template modification instruction and upgrading sub-template data;

and generating an associated template upgrading prompt according to the template change data, and sending the associated template upgrading prompt to an author terminal corresponding to the author identifier.

5. The method of claim 4, wherein after obtaining the tongue picture paraphrase matching the tongue picture attribute, further comprising:

sending the tongue picture image and the tongue picture paraphrase of the patient to a doctor terminal corresponding to the idle doctor;

the method for extracting symptom characteristics from the patient complaint data and the tongue picture paraphrases comprises the following steps:

and when a paraphrase confirmation notice which is returned by the doctor terminal and corresponds to the tongue picture paraphrase is received, extracting symptom characteristics from the patient complaint data and the tongue picture paraphrase.

6. The method of claim 5, further comprising:

when receiving a paraphrase confirmation notice and a corrected tongue picture paraphrase which are returned by the doctor terminal and correspond to the tongue picture paraphrase, extracting symptom characteristics from the patient's chief complaint data and the corrected tongue picture paraphrase;

adding the tongue picture attribute and the corrected tongue picture paraphrase association to a tongue picture correction sample set;

7. A patient information acquisition device, the device comprising:

the paraphrase acquisition module is used for inputting the tongue picture attributes into a preset neural network classifier to obtain a first paraphrase matching probability of each tongue picture paraphrase; inputting the tongue picture attributes into a preset Bayes classifier to obtain second paraphrase matching probability of each tongue picture paraphrase; obtaining the paraphrase matching rate of each tongue picture paraphrase according to the first paraphrase matching probability and the second paraphrase matching probability, and extracting the tongue picture paraphrase with the highest paraphrase matching rate as the tongue picture paraphrase matched with the tongue picture attributes; the preset neural network classifier is used for carrying out tongue picture paraphrase classification on the tongue picture attributes; the preset Bayes classifier is used for carrying out tongue interpretation classification on the tongue picture image of the patient;

and the information generating module is used for generating patient acquisition information according to the patient reply data.

8. The apparatus of claim 7, wherein the data acquisition module comprises:

the current node extraction module is used for acquiring an initial node in the template data and initial node data corresponding to the initial node;

the reply data acquisition module is used for generating and outputting an initial acquisition problem according to the initial node data and acquiring initial patient reply data corresponding to the initial acquisition problem;

and the template execution module is used for acquiring the connecting nodes corresponding to the starting node from the template data, selecting the first nodes corresponding to the starting patient reply data from the connecting nodes, taking the first nodes as current nodes, and continuously generating and outputting current acquisition problems according to the current node data corresponding to the current nodes until the current nodes corresponding to the acquired current patient reply data are template peripheral nodes.

9. A computer device comprising a memory and a processor, the memory storing a computer program, wherein the processor when executing the computer program performs the steps of the method according to any of claims 1 to 6.

10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 6.