Electronic medical record information extraction method, system and equipment and storage medium

Abstract

The invention discloses an electronic medical record information extraction method, system and device and a storage medium, and generally relates to the technical field of data processing. The method comprises the following steps: extracting a first character feature through a first sub-model in an electronic medical record information extraction model, and extracting a first picture feature through the first sub-model; processing the character features in the first electronic medical record information and the image features in the first electronic medical record information through a second sub-model in the electronic medical record information extraction model, and identifying a second character feature in the first character features and a second image feature in the first image features; identifying the second character feature and the second picture feature through third sub-model identification in the electronic medical record information extraction model, and obtaining entity information corresponding to the first electronic medical record information; and processing the entity information according to a predefined structured format through a fourth sub-model in the electronic medical record information extraction model to generate structured information. The diagnosis duration of a doctor can be shortened, and the diagnosis efficiency of the doctor is improved.

Description

Technical Field

[0001] The present disclosure generally relates to the field of data processing technology, and more particularly to a method, system, device and storage medium for extracting electronic medical record information. Background Art

[0002] With the continuous development of electronic technology, it has become very common in various fields to analyze and calculate data in their field through model training.

[0003] In the medical field, electronic medical records have long been popularized to ensure the accuracy of patient diagnosis. Based on its easy-to-preserve and non-lost characteristics, electronic medical records can retain all the diagnostic process information for each patient in detail. For any patient, in order to ensure the accuracy of the diagnosis, doctors may need the support of multiple examination data for each diagnosis. Therefore, for patients, when they finally face the doctor and the doctor makes a diagnosis for them, the patient's electronic medical record may already contain a large amount of data, and there are many pages that require the doctor to constantly click on the display terminal (such as a computer) to flip through, and the doctor needs to check for a period of time before making a diagnosis decision. Summary of the invention

[0004] In view of the above-mentioned defects or deficiencies in the relevant technology, it is desired to provide a method, system, device and medium for extracting electronic medical record information, which can solve the problem that electronic medical records often have a lot of content, and in the process of manual review, it is necessary to review the electronic medical records for a long time before a diagnosis decision can be made, which reduces the diagnosis efficiency. Through the electronic medical record information extraction method of the present application, the doctor's diagnosis time can be greatly shortened and the doctor's diagnosis efficiency can be improved.

[0005] In a first aspect, a method for extracting electronic medical record information is provided, the method comprising: Extracting a first text feature in the first electronic medical record information through a first sub-model in the electronic medical record information extraction model, and extracting a first image feature in the first electronic medical record information through the first sub-model, wherein the first sub-model is a Word2Vec model and a normalization model; Processing the text features in the first electronic medical record information and the image features in the first electronic medical record information through the second sub-model in the electronic medical record information extraction model, identifying the second text features in the first text features and the second picture features in the first picture features, the second sub-model is a multi-task model, and the multi-task model includes a BERT model, an LSTM model, and a CNN model, the second text features are part of the text features or all of the text features in the first text features, and the second picture features are part of the picture features or all of the picture features in the first picture features; Recognize the second text feature and the second image feature through the third sub-model in the electronic medical record information extraction model to obtain entity information corresponding to the first electronic medical record information, wherein the entity information includes: user information, user status, and conclusive information of the user status, and the third sub-model includes a multi-channel self-attention model; The entity information is processed in a predefined structured format through the fourth sub-model in the electronic medical record information extraction model to generate structured information, and the structured information is used to display key information in the entity information in a preset format. The fourth sub-model is a structured model.

[0006] In the present application, the first text feature in the first electronic medical record information is extracted by the Word2Vec model and the normalization model in the electronic medical record information extraction model, and the first picture feature in the first electronic medical record information is extracted by the above-mentioned first sub-model; then, the text features in the above-mentioned first electronic medical record information and the image features in the first electronic medical record information are processed by the multi-task model in the above-mentioned electronic medical record information extraction model to identify the second text feature in the first text feature and the second picture feature in the first picture feature (the multi-task model includes the BERT model, the LSTM model, and the CNN model), and the second text feature is the first text feature. The second image feature is part of the text feature or all of the text features in the feature, and the second image feature is part of the image feature or all of the image features in the first image feature; then, the second text feature and the second image feature are recognized by the multi-channel self-attention model in the electronic medical record information extraction model to obtain the entity information corresponding to the first electronic medical record information, and the entity information includes: user information, user status, and conclusive information of the user status; finally, the entity information is processed according to the predefined structured format by the structured model in the electronic medical record information extraction model to generate structured information, and the structured information is used to display the key information in the entity information according to the preset format. In this way, the electronic medical record information extraction model can automatically extract all the valid information in the electronic medical record, and make structural adjustments, and finally display the structured information to the user without the user having to look through all the electronic medical records by themselves. In this way, the efficiency of doctors viewing electronic medical records can be greatly improved and the time for viewing electronic medical records can be shortened.

[0007] In a second aspect, an electronic medical record information extraction system is provided, the system comprising: an extraction unit, configured to extract a first text feature in the first electronic medical record information through a first sub-model in an electronic medical record information extraction model, and to extract a first image feature in the first electronic medical record information through the first sub-model, wherein the first sub-model is a Word2Vec model and a normalization model; an execution unit, configured to process text features in the first electronic medical record information and image features in the first electronic medical record information through a second sub-model in the electronic medical record information extraction model, and identify a second text feature in the first text feature and a second picture feature in the first picture feature, wherein the second sub-model is a multi-task model, and the multi-task model includes a BERT model, an LSTM model, and a CNN model, wherein the second text feature is a part of the text feature or all of the text features in the first text feature, and the second picture feature is a part of the picture feature or all of the picture features in the first picture feature; an acquisition unit, configured to recognize the second text feature and the second image feature through a third sub-model in the electronic medical record information extraction model, and acquire entity information corresponding to the first electronic medical record information, wherein the entity information includes: user information, user status, and conclusive information of the user status, and the third sub-model includes a multi-channel self-attention model; A generation unit is used to process the entity information in a predefined structured format through the fourth sub-model in the electronic medical record information extraction model to generate structured information, wherein the structured information is used to display key information in the entity information in a preset format, and the fourth sub-model is a structured model.

[0008] In a third aspect, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method described in the first aspect is implemented.

[0009] In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored, characterized in that when the program is executed by a processor, the method described in the first aspect is implemented.

[0010] In a fifth aspect, a computer program product is provided. The computer program product includes instructions, and the instructions are executed by a processor to implement the method described in the first aspect.

[0011] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings: Figure 1 A flowchart of a method for extracting electronic medical record information provided in an embodiment of the present application; Figure 2 A schematic diagram of the application process of the electronic medical record information extraction model provided in an embodiment of the present application is shown in FIG. Figure 3 A schematic diagram of the training process of the electronic medical record information extraction model provided in an embodiment of the present application; Figure 4 A schematic diagram of a process for obtaining structured information provided in an embodiment of the present application; Figure 5 A schematic diagram of the structure of a resource access device provided in an embodiment of the present application; Figure 6 A schematic diagram of the structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0013] The present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the invention are shown in the accompanying drawings.

[0014] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0015] Figure 1 FIG. 1 is a flow chart of a method for extracting electronic medical record information provided by an embodiment of the present application. Figure 1 As shown, the method includes the following steps 301 to 304: Step 301: extracting a first text feature in first electronic medical record information through a first sub-model in an electronic medical record information extraction model, and extracting a first image feature in the first electronic medical record information through the first sub-model.

[0016] In an embodiment of the present application, the first sub-model is a Word2Vec model and a normalized model.

[0017] It is understandable that in the embodiment of the present application, the first electronic medical record information may include text information and image information. Therefore, in order to ensure the integrity of the extracted first electronic medical record information, it is necessary to extract both the text features of the first electronic medical record information and the image features of the first electronic medical record information. In an embodiment of the present application, the Word2Vec model can be used to identify text information in the first electronic medical record and further extract the first text feature in the first electronic medical record, and the normalization model can be used to identify image information in the first electronic medical record and further extract the first image feature in the first electronic medical record.

[0018] In the embodiment of the present application, the above-mentioned first text feature is the feature information after the text information in the first electronic medical record is vectorized; correspondingly, the above-mentioned first image feature is the feature information after the image information in the first electronic medical record information is vectorized.

[0019] Step 302: Process the text features in the first electronic medical record information and the image features in the first electronic medical record information through the second sub-model in the electronic medical record information extraction model to identify the second text features in the first text features and the second image features in the first image features.

[0020] In an embodiment of the present application, the second sub-model is a multi-task model.

[0021] Exemplarily, the multi-task model includes a BERT model, a LSTM model, and a CNN model.

[0022] In the embodiment of the present application, the second text feature is part of the text features or all of the text features in the first text feature, and the second picture feature is part of the picture features or all of the picture features in the first picture feature; It can be understood that the above-mentioned first text feature is all the text features in the first electronic medical record information, and correspondingly, the first picture feature is also all the image features in the first electronic medical record information. For the present application, its purpose is to extract the effective features or key features in the first electronic medical record information. The effective features or key features may be part of the features or all of the features in the first electronic medical record information. Based on this, the second text feature is part of the features or all of the features of the first text feature, and correspondingly, the second picture feature is part of the features or all of the features of the first picture feature.

[0023] In an embodiment of the present application, the above-mentioned BERT model, LSTM model and CNN model are models that have been trained and can be used to identify and extract key features.

[0024] In an embodiment of the present application, the above-mentioned BERT model and LSTM model can be used to identify and extract the second text feature from the first text feature; the above-mentioned CNN model can be used to identify and extract the second image feature from the first image feature.

[0025] Step 303: Recognize the second text feature and the second image feature through the third sub-model in the electronic medical record information extraction model to obtain entity information corresponding to the first electronic medical record information.

[0026] In the embodiment of the present application, the above entity information includes: user information, user status, and conclusive information of the user status.

[0027] In an embodiment of the present application, the third sub-model includes a multi-channel self-attention model.

[0028] In the embodiment of the present application, after obtaining the key information in the first electronic medical record information, that is, the second text feature and the second image feature, it is necessary to further clarify the association between different key information, so as to enhance the readability of the key information subsequently displayed. Based on this, it is necessary to integrate the second text feature and the second image feature, and finally obtain the entity information corresponding to the first electronic medical record.

[0029] It can be understood that the user information in the above-mentioned entity information can be the user's own personal information, the above-mentioned user status can be the user's inspection process information, such as the user's indicator information, etc., and the conclusive information of the above-mentioned user status is the conclusive information corresponding to the user status.

[0030] Furthermore, the above entity information may include past information in the first electronic medical record, or past information and current information, wherein the current information includes user information and user status information.

[0031] Step 304: Process the entity information according to a predefined structured format through the fourth sub-model in the electronic medical record information extraction model to generate structured information.

[0032] In an embodiment of the present application, the above-mentioned structured information is used to display key information in the entity information according to a preset format.

[0033] In the embodiment of the present application, the fourth sub-model is a structured model.

[0034] It is understandable that after generating the above-mentioned entity information, it is necessary to process the entity information according to a predefined structured format through a structured model so that it can be finally presented to the user. For example, the doctor's information is structured information with strong readability and unified structure, thereby improving the user's efficiency in reading the first electronic medical record and obtaining effective information, and shortening the subsequent judgment time.

[0035] In the method provided in the embodiment of the present application, the first text feature in the first electronic medical record information is extracted by the Word2Vec model and the normalization model in the electronic medical record information extraction model, and the first picture feature in the first electronic medical record information is extracted by the above-mentioned first sub-model; then, the text features in the above-mentioned first electronic medical record information and the image features in the first electronic medical record information are processed by the multi-task model in the above-mentioned electronic medical record information extraction model to identify the second text feature in the first text feature and the second picture feature in the first picture feature (the multi-task model includes a BERT model, an LSTM model, and a CNN model), and the second text feature is The first text feature is part of the text feature or all of the text features, and the second image feature is part of the image feature or all of the image features in the first image feature; then, the second text feature and the second image feature are identified by the multi-channel self-attention model in the electronic medical record information extraction model to obtain the entity information corresponding to the first electronic medical record information, and the entity information includes: user information, user status, and conclusive information of the user status; finally, the entity information is processed according to the predefined structured format by the structured model in the electronic medical record information extraction model to generate structured information, and the structured information is used to display the key information in the entity information according to the preset format. In this way, the electronic medical record information extraction model can automatically extract all the effective information in the electronic medical record, and make structural adjustments, and finally display the structured information to the user without the user having to look through all the electronic medical records by himself, so that the efficiency of doctors viewing electronic medical records can be greatly improved and the time for viewing electronic medical records can be shortened.

[0036] In another embodiment of the present application, a specific implementation method for preprocessing the initial electronic medical record information is also provided. Exemplarily, before the "extracting the first text feature in the first electronic medical record information by the first sub-model in the electronic medical record information extraction model" mentioned above, its implementation method also includes: performing data preprocessing on the initial electronic medical record information according to a preset preprocessing method to obtain the above-mentioned first electronic medical record information.

[0037] Exemplarily, the above-mentioned preset preprocessing methods include: information segmentation, part-of-speech tagging, grayscale processing, and binarization processing.

[0038] It is understandable that after the initial electronic medical record information is obtained, it is necessary to pre-process the initial electronic medical record information so that the first text feature and the first image feature can be extracted clearly and quickly in the subsequent process.

[0039] For example, for the text information of the initial electronic medical record information, the text information should be pre-processed by information segmentation and part-of-speech tagging. After the initial electronic medical record information is segmented and part-of-speech tagged, a vocabulary list can be generated. Exemplarily, for the image information of the initial electronic medical record information, the image information should be pre-processed by grayscale processing and binarization processing.

[0040] Exemplarily, in the above-mentioned process of preprocessing the initial electronic medical record information according to a preset method, it is necessary to collect and purify the initial electronic medical record information, and eliminate irrelevant information and noise, so that in the subsequent process of extracting the first text feature, the first text feature required in the text information and the first image feature required in the image information can be more accurately extracted.

[0041] In another embodiment of the present application, a specific implementation method for extracting the first text feature is also provided. Exemplarily, the "extracting the first text feature in the first electronic medical record information through the first sub-model in the electronic medical record information extraction model" mentioned above includes: identifying the first text information in the first electronic medical record information through the Word2Vec model; obtaining all lexical features in the first text information, and processing the lexical features into lexical vectors, obtaining the association between different lexical vectors; generating the first text feature according to the lexical vector and the association between different lexical vectors.

[0042] Exemplarily, the first text feature is feature information formed by a text vector.

[0043] Exemplarily, the first text feature vector is generated by a trained Word2Vec model. The first text feature vector may be a word vector, which is used to capture the relationship between words by mapping each word to a vector space.

[0044] Exemplarily, the above Word2Vec model includes the calculation formula of the Skip-Gram model.

[0045] In the related art, the calculation formula of the Skip-Gram model is shown in formula (1): (1) in, is the vector of the target word, is the vector of the context word, is the total number of words in the vocabulary.

[0046] In an embodiment of the present application, the above-mentioned Word2Vec model is trained by a large amount of corpus in the medical field. Therefore, a large number of word vectors adapted to the medical field are introduced into the Word2Vec model in the embodiment of the present application, so that the word vectors are closer to the context of the identified first electronic medical record information, thereby improving the model's ability to understand medical terms.

[0047] In the embodiment of the present application, the Skip-Gram model calculation formula (2) in the embodiment of the present application is: (2) in is the vector of the target word, is the vector of the context word, is the bias term for domain adaptation, is the total number of words in the vocabulary.

[0048] In another embodiment of the present application, a specific implementation method for extracting the first image feature is also provided. Exemplarily, the above-mentioned "extracting the first image feature in the first electronic medical record information through the second sub-model" includes: identifying all image information in the first electronic medical record information through the normalization processing model; obtaining the light brightness information of each image information, the light brightness information includes the light brightness information of multiple local areas in each image information; performing normalization processing on the light brightness information of each image information to generate the first image feature.

[0049] Exemplarily, the above-mentioned one picture information includes the above-mentioned multiple local areas Exemplarily, the first image feature includes a plurality of first sub-image features in the first electronic medical record information, and one first sub-image feature corresponds to one image information.

[0050] Exemplarily, the image information may be normalized by a normalization model, thereby reducing the influence of illumination and contrast on the recognition process during the process of recognizing the image in the first electronic medical record.

[0051] For example, in an embodiment of the present application, in the process of normalizing image information, it is necessary to consider local normalization processing of the image information. The reason is that, in view of the non-uniform illumination problem that may exist in the image information in the first electronic medical record information, normalization using the mean and standard deviation of the local area can more effectively handle illumination changes in the image information.

[0052] The normalization processing calculation formula of the normalization model in the related art is as follows (3) (3) in, is the original image, is the mean of the image, is the standard deviation of the image.

[0053] The improved normalized calculation formula of the embodiment of the present application is as follows: (4) in is the original image, is the mean value of the local area of ​​the image, is the standard deviation of the local area of ​​the image.

[0054] In another embodiment of the present application, a specific implementation method for identifying the second text feature and the second image feature is also provided. Exemplarily, the above-mentioned "processing the first text feature in the first electronic medical record information and the first image feature in the first electronic medical record information through the second sub-model in the electronic medical record information extraction model, identifying the second text feature in the first text feature and the second image feature in the first image feature" includes: extracting the text sequence vector in the first text feature through the BERT model and the LSTM model, and processing the first text feature according to the text sequence vector to generate the second text feature; extracting the text sequence vector in the first image feature through the CNN model, and processing the first image feature according to the text sequence vector to generate the second image feature.

[0055] Illustratively, in the embodiment of the present application, the extracted first text feature and first image feature need to be further processed so as to identify key features therein.

[0056] Exemplarily, in the process of extracting the second text feature from the first text feature, the BERT model is used to encode the text data and capture the contextual relationships and semantic information in the text. Through its Transformer architecture, BERT is able to understand the meaning of words in different contexts, thereby extracting rich text features. The LSTM model is then used to process sequence data, such as time series data, to capture long-term dependencies. LSTM is particularly suitable for processing and predicting time series-based data. For example, we use BERT to identify disease names and drug names in the first text feature, while LSTM can be used to analyze the time series of treatment processes and disease progression in the first text feature. Combined in this way, the model is able to more comprehensively understand and process the information content of the first electronic medical record information.

[0057] In the related art, the calculation formula of the BERT model is as shown in the following formula (5): (5) in, is the input word vector sequence, It is the core architecture of the BERT model.

[0058] In the embodiment of the present application, the above BERT model can be fine-tuned specifically for its applicable scenarios in the medical field, that is, the BERT model is continuously trained and fine-tuned using electronic medical record data to better capture the language characteristics of the medical field. The calculation formula of the improved and adjusted BERT model (6) is: (6) in is the input word vector sequence, It is a parameter specific to the medical field.

[0059] For the image part, a CNN model is used.

[0060] In the related art, the calculation formula of the CNN model is as shown in the following formula (7): (7) in, is the input image, is the convolution kernel, represents the convolution operation, b is the bias term, and σ is the activation function.

[0061] Furthermore, in the embodiment of the present application, in order to characterize the different components of the second text feature and enhance the model's ability to recognize key information, an attention mechanism may also be introduced.

[0062] In the related art, the calculation formula of the attention mechanism is as shown in the following formula (8): (8) in, is the query matrix, is the key matrix, is the value matrix, is the dimension of the key vector.

[0063] In the embodiment of the present application, it is necessary to adjust the above multi-channel self-attention mechanism, by applying the attention mechanism in different channels and outputting a two-dimensional weight matrix to characterize the different components of the sentence and enhance the model's ability to recognize key information. The improved attention mechanism calculation formula in the embodiment of the present application is shown in the following formula (9): (9) in is the query matrix, is the key matrix, is the value matrix, is the interaction matrix between channels, is the dimension of the key vector.

[0064] In another embodiment of the present application, a specific implementation method for generating structured information is also provided. Exemplarily, the above-mentioned "processing the above-mentioned entity information according to a predefined structured format through the fourth sub-model in the above-mentioned electronic medical record information extraction model to generate structured information" includes: identifying the above-mentioned entity information and extracting the entity relationship in the entity information through the BiLSTM-CRF model in the structured model, and determining the entity boundary of the above-mentioned entity information; determining the key information in the above-mentioned entity information according to the above-mentioned entity information, the entity relationship in the above-mentioned entity information, and the entity boundary, and generating structured information.

[0065] Exemplarily, the above entity boundary is the boundary of the entity, that is, the starting and ending positions of the entity in the electronic medical record information.

[0066] In the embodiment of the present application, in the process of generating the above structured information, entity recognition and entity relationship organization need to be performed. In the process of entity recognition, named entity recognition and relationship extraction need to be performed, and the BiLSTM-CRF model is used to complete this process.

[0067] In the related technology, the BiLSTM-CRF model is used. The BiLSTM-CRF model is based on weighted position embedding, which enhances the feature expression ability of characters of different distances in the Transformer network and improves the model's recognition accuracy of entity boundaries.

[0068] The calculation formula of the BiLSTM-CRF model is as shown in the following formula (13) and formula (14): (13) (14) in, is the input sequence, is the label sequence, is the transfer matrix, is the emission matrix.

[0069] In the embodiment of the present application, the BiLSTM-CRF model is improved for the application scenario of electronic medical records. The calculation formula of the improved BiLSTM-CRF model is shown in formula (15) and formula (16): (15) (16) in is the input sequence, is the label sequence, is the transfer matrix, is the emission matrix, It is a weighted position embedding.

[0070] In an example, for the first electronic medical record information, the process of determining the entity information, extracting the entity relationship, and determining the entity boundary of the entity information is as follows: (1) Entity recognition: Use the BiLSTM-CRF (Bidirectional Long Short-Term Memory Network-Conditional Random Field) model to identify medical entities in the text, such as disease names, drug names, treatment methods, etc., and confirm entity boundaries. The following table shows the recognition of key information:

[0071] (2) Relationship extraction: After identifying entities, the model further analyzes the relationships between entities, such as the association between "patient" and "disease" and the connection between "drug" and "indication". The relationship extraction process is determined by analyzing the context and grammatical structure between entities.

[0072] (3) Structured information generation: The identified entities and their boundaries, as well as the relationships between entities, are organized in a predefined structured format (such as JSON or XML).

[0073] The structured information is as follows (JSON format): The contents contained in the above structured information are described as follows: PatientInfo: Contains basic information about the patient, such as name, gender, and age.

[0074] ChiefComplaint: Record the patient's main symptoms and problems.

[0075] MedicalHistory: Provides a summary of the patient's medical history, including past medical history and long-term medications.

[0076] TestResults: Detailed records of the patient's test results, including blood tests and blood sugar levels.

[0077] Diagnosis: Determine the patient's diagnosis, including the type and severity of the disease.

[0078] TreatmentPlan: Lists the treatment plan, including drug name, dosage, and treatment cycle.

[0079]

[0080] In this example, entities such as "Zhang San" (patient name), "drug-induced liver injury" (diagnosis), and "white moss skin" (drug name) are identified and labeled with entity boundaries. Relationships between entities, such as the impact of "white moss skin" on "abnormal liver function", are also extracted and structured. In this way, unstructured electronic medical record data is converted into structured information for easy storage, retrieval, and further data analysis.

[0081] It should be noted that in the embodiment of the present application, both named entity recognition (NER) and relation extraction use the BiLSTM-CRF model. It is understandable that in the embodiment of the present application, entity information can also be extracted from the first text information by using named entity recognition (NER) technology. Key areas can be extracted from the image using object detection or image segmentation technology. The extracted information can be organized according to a predefined structured format, such as JSON, XML, etc., and then the relationship between entities can be identified using relationship extraction technology, thereby obtaining the relationship between different vocabulary vectors, and structural processing can also be performed.

[0082] like Figure 2 As shown, Figure 2 The application process of the electronic medical record information extraction model is shown in FIG. The application process is trained in a graphical manner.

[0083] In another embodiment of the present application, a specific implementation method for generating an electronic medical record information processing model is also provided. Exemplarily, the implementation method includes: obtaining multiple sets of original electronic medical records and target data corresponding to the multiple sets of original electronic medical records; inputting the multiple sets of original electronic medical records into the initial electronic medical record information extraction model to generate test data corresponding to the original electronic medical records; comparing the multiple sets of test data with the multiple sets of target data, and determining the difference between the multiple sets of test data and the multiple sets of target data; When the difference is greater than a preset threshold, the initial electronic medical record information extraction model is adjusted according to the difference so that the difference between the test data and the target data generated by the electronic medical record information processing model is less than the preset threshold.

[0084] Exemplarily, the target data includes: target structured information, the structured information corresponding to the multiple groups of original electronic medical records is structured information in a target format corresponding one-to-one to each sub-medical record in the multiple groups of original electronic medical records, and the multiple groups of original electronic medical records include data information of at least two users.

[0085] Exemplarily, the test data includes test structured information Exemplarily, the above-mentioned initial electronic medical record information extraction model includes an initial Word2Vec model, a normalization model, a BERT model, a LSTM model, a CNN model, a multi-channel self-attention model, and a structured model.

[0086] It can be understood that in the embodiment of the present application, it is necessary to train the initial electronic medical record information extraction model multiple times using target data and original data corresponding to multiple groups of original electronic medical records, and compare the difference between each training result and the target result. The difference can be expressed by the above-mentioned difference. Only when the difference is less than the preset threshold can an electronic medical record information extraction model that can be put into practical use be obtained.

[0087] Furthermore, in the embodiments of the present application, the model is trained by the labeled data, that is, the target data, and the model performance can be evaluated and optimized by using methods such as cross-validation. Technologies such as transfer learning can be used to improve the generalization ability of the model.

[0088] It can be understood that in the embodiment of the present application, the above-mentioned electronic medical record information extraction model includes multiple sub-models, such as a first sub-model, a second sub-model, a third sub-model and a fourth sub-model, wherein the above-mentioned four sub-models are as described above, and the training process of the electronic medical record information extraction model can be completed by means of operations research optimization thereof.

[0089] In an embodiment of the present invention, operations research optimization is used in the process of training the electronic medical record information extraction model.

[0090] It is understandable that operations research optimization is mainly used to solve decision-making problems in the information extraction process, which can be formalized as an optimization problem, such as using linear programming or integer programming to optimize the information extraction strategy, which contains a series of decision variables, objective functions and constraints. The results of operations research optimization are fed back into the model training process to form an iterative loop to continuously optimize model performance. Operations research optimization may be closely integrated with the training process of deep learning models, and mathematical programming methods can be used to guide various aspects of model training, thereby improving the performance of the model in electronic medical record information extraction and structuring tasks, more finely controlling the model's learning process, and ensuring that the model can effectively generalize to new data sets while meeting specific performance indicators.

[0091] The operations research optimization process mainly includes the following aspects: 1. Feature selection and weight assignment: Operations research optimization is used to determine the weights of different features (such as text features, image features, etc.) in the model to improve the accuracy of information extraction. This involves the definition of decision variables, where each feature can be assigned a weight value, and these weights are determined by optimizing the objective function to achieve feature selection.

[0092] 2. Model parameter adjustment: Select the best model parameter configuration through operations research methods. For example, in a multi-task learning model, integer programming can be used to balance the processing intensity of text and image data to improve model performance.

[0093] 3. Information extraction strategy optimization: Use linear programming or integer programming to optimize the information extraction strategy to ensure that the extracted key information is both accurate and comprehensive. This includes setting objective functions and constraints to optimize the information extraction process.

[0094] 4. Resource allocation: When processing large-scale electronic medical record data, operations research can help optimize the allocation of computing resources. For example, in a cloud environment, computing nodes and storage resources can be reasonably allocated to improve processing efficiency.

[0095] The operations research optimization process is described below. The operations research optimization process includes linear programming and integer programming: In the related art, the calculation formula of the optimization target of the linear programming process is shown in the following formula (10): (10) in, is the cost vector, is the decision variable vector, is the constraint matrix, is the constraint value vector.

[0096] Linear programming and integer programming are combined with the penalty term of the category constraint matrix. Aiming at the uneven distribution of electronic medical record entity relationships, the penalty term based on the category constraint matrix is ​​derived, and the model is learned and trained together with the loss function. Based on this, in the embodiment of the present application, the improved linear programming calculation formula is shown in the following formula (11): (11) in, is the cost vector, is the decision variable vector, is the constraint matrix, is the constraint value vector. is the penalty term coefficient, is a penalty term based on the class constraint matrix.

[0097] (2) Integer programming: The calculation formula of the optimization target is shown in the following formula (12): subject to quad (12) in, is a vector of integer decision variables.

[0098] like Figure 3 As shown, Figure 3 The training process of the electronic medical record information extraction model is shown in FIG. In this process, model parameters need to be set in advance, and training is performed in the manner shown in the figure.

[0099] The following is a summary of the electronic medical record information extraction method in the embodiment of the present application. The electronic medical record information extraction method proposed in the embodiment of the present application and the electronic medical record information extraction model used in the method, which is mainly achieved through the combination of deep learning and operations research optimization technology. Significant improvement in the effect of electronic medical record information extraction. In the actual application process of the model, firstly through the data preprocessing stage, basic word segmentation, part-of-speech tagging and image graying and binarization processing are performed. At the same time, the domain-adapted Word2Vec model is introduced to generate word vectors that are closer to the medical context, and local normalization is used to more effectively cope with the illumination changes in the image. Specifically, the model is composed of a plurality of sub-models and belongs to a multi-task learning model. The model combines BERT and LSTM to process text data, and uses CNN to process image data. At the same time, a multi-channel self-attention mechanism is introduced to strengthen the model's ability to recognize key information. In addition, the operations research optimization technology is integrated to optimize the information extraction process by introducing a penalty term based on a category constraint matrix, especially in dealing with the problem of uneven distribution of electronic medical record entity relationships. In the information extraction and structuring stage, this method uses an improved BiLSTM-CRF model and introduces weight-based position embedding to enhance the recognition accuracy of entity boundaries. These optimization measures not only improve the accuracy and efficiency of the model, but also enhance the model's ability to handle the complexity and diversity of electronic medical record data, making the present invention innovative and practical in the field of electronic medical record information extraction technology.

[0100] The above-mentioned process of obtaining structured information is summarized as follows through a flowchart: Figure 4 As shown: When the electronic medical record information extraction model obtains the electronic medical record, the data of the original electronic medical record will be preprocessed first (S101). The data preprocessing mainly involves word segmentation, part-of-speech tagging, grayscale conversion, and binarization of the unstructured data such as medical records, examination reports, and medication records in the electronic medical record to obtain the processed first electronic medical record; then, feature extraction is performed (S102). Specifically, models such as Word2Vec are used to extract text features of word vectors in the text, and image features are extracted through a normalization model such as a CNN model; then, text features and image features are fused through a multi-task learning model (S103). Specifically, the BERT model and LSTM model are used to process text features, and the CNN model is used to extract image features. The model processes the image features to obtain key information in the text features and key information in the image features; then, the multi-channel self-attention model of the attention mechanism automatically identifies and focuses on the key information in the above text features and key information in the image features (S104), and uses the BiLSTM-CRF model to identify entity information from the text (S105), such as disease names, drug names, corresponding plans, etc., and performs relationship extraction (S106) to identify the relationship between entities; finally, according to predefined structured formats, such as JSON and XML, the extracted information and relationships are organized into structured information, and the final structured information is output (S107), which can be used for further data analysis, clinical decision support and other applications.

[0101] It can be understood that in the above process, the model will identify key medical record information in the first electronic medical record.

[0102] The application process of this application in a specific application scenario is described below, as shown below: In actual application, suppose there is an original electronic medical record of Zhang San, which records the basic information, chief complaint, medical history summary, examination results, diagnosis and corresponding treatment plan of the patient Zhang San. The original electronic medical record contains key information such as Zhang San’s current symptoms, medication use records, examination test results and clinical diagnosis and treatment activities.

[0103] The information includes: 1. Patient information: Zhang San, male, 45 years old; 2. Chief complaint: Weight loss of 5 kg in the past month, accompanied by intermittent fever; 3. Medical history summary: The patient has a history of diabetes and has been taking hypoglycemic drugs for a long time; 4. Examination results: Routine blood test showed high white blood cell count and blood sugar level of 8.5mmol / L.

[0104] 5. Diagnosis: Drug-induced liver injury, acute, hepatocellular damage type, RUCAM score 10 points.

[0105] The original electronic medical record processing flow requires the use of multiple different sub-models in the electronic medical record information extraction model, including Word2Vec model, normalization model, BERT model, LSTM model, CNN model, multi-channel self-attention model, structured model, etc. The specific processing process includes: 1. Data preprocessing 1) Segment and tag the text information in the original electronic medical record, and generate word vectors using the domain-adapted Word2Vec model 2) Perform local normalization on image information (such as X-rays, MRI scans) to improve the stability of image features.

[0106] 2. Extract features through the model and build the model 1) Use the BERT model to process text information data and extract text features (that is, the first text feature mentioned above).

[0107] 2) Use the CNN model to process image information data and extract image features (i.e. the first image feature mentioned above).

[0108] 3) Combining the text features and the image features, according to the pre-set multi-task learning model, further identifying the key text features (that is, the second text features mentioned above) and the key image features (that is, the second picture features mentioned above).

[0109] For example, the key text features and key image features identified for the above information are: Patient information: 1) Name: Zhang San 2) Gender: Male 3) Age: 45 Main complaint: 1) Weight loss of 5 kg 2) With intermittent fever Medical history summary: 1) History of diabetes 2) Long-term use of hypoglycemic drugs Examination results: 1) Blood routine: high white blood cell count 2) Blood sugar: 8.5 mmol / L Diagnosis: 1) Drug-induced liver injury, acute, hepatocellular injury type 2) RUCAM score: 10 points Treatment options: 1) Insulin injection once a day Antibiotic treatment for 7 days 3. Apply the attention mechanism By using a multi-channel self-attention mechanism, the entity information in the above key text features and key image features is identified.

[0110] In this example, three recognition results are required: medical record information recognition, drug information recognition, and clinical information recognition. Specifically, medical record information recognition includes: identifying the patient's basic information, chief complaint, and medical history summary; drug information recognition includes: identifying the hypoglycemic drugs and antibiotics used by the patient; and corresponding plan recognition includes: identifying the corresponding plan for the patient's symptoms. 4. Information extraction and structuring 1) Use the BiLSTM-CRF model for named entity recognition and confirm entity information.

[0111] 2) Perform relationship extraction to identify the relationships between entities.

[0112] 3) Organize the extracted information and relationships into structured data according to a predefined structured format.

[0113] In this example, the structured data is the "Structured information is as follows (JSON format)" in the above content. It should be noted that the above-mentioned electronic medical record information extraction model also needs to be pre-trained, and the training process can be added with the process of operations research optimization. Specifically, the operations research optimization can optimize the information extraction process through the linear programming and integer programming models of the training model to ensure the accuracy and completeness of the information extracted by the model.

[0114] The present application also provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the training rule determination method described in the present application is implemented. For example, the following can be executed: Figure 1 The individual steps of the method are shown.

[0115] The present invention provides a computer program product, which includes instructions that are executed by a processor when the processor executes the instructions. Figure 1 The individual steps of the method are shown.

[0116] It should be noted that although the operations of the method of the present invention are described in a particular order in the drawings, this does not require or imply that the operations must be performed in this particular order or that all illustrated operations must be performed to achieve desired results.

[0117] Figure 5 This is a block diagram of an electronic medical record information extraction system according to an embodiment of the present application. Figure 5 The system includes an extracting unit 601, an executing unit 602, an acquiring unit 603 and a generating unit 604.

[0118] An extraction unit 601 is used to extract a first text feature in the first electronic medical record information through a first sub-model in the electronic medical record information extraction model, and to extract a first image feature in the first electronic medical record information through the first sub-model, wherein the first sub-model is a Word2Vec model and a normalization model; An execution unit 602 is used to process the text features in the first electronic medical record information and the image features in the first electronic medical record information through the second sub-model in the electronic medical record information extraction model to identify the second text features in the first text features and the second picture features in the first picture features, wherein the second sub-model is a multi-task model, and the multi-task model includes a BERT model, an LSTM model, and a CNN model, the second text features are part of the text features or all of the text features in the first text features, and the second picture features are part of the picture features or all of the picture features in the first picture features; An acquisition unit 603 is used to recognize the second text feature and the second image feature through a third sub-model in the electronic medical record information extraction model to acquire entity information corresponding to the first electronic medical record information, wherein the entity information includes: user information, user status, and conclusive information of the user status, and the third sub-model includes a multi-channel self-attention model; The generation unit 604 is used to process the entity information in a predefined structured format through the fourth sub-model in the electronic medical record information extraction model to generate structured information, wherein the structured information is used to display key information in the entity information in a preset format, and the fourth sub-model is a structured model.

[0119] In one embodiment, the execution unit 602 is also used to perform data preprocessing on the initial electronic medical record information according to a preset preprocessing method to obtain the first electronic medical record information; the preset preprocessing method includes: information segmentation, part-of-speech tagging, grayscale processing, and binarization processing.

[0120] In one embodiment, the extraction unit 601 is specifically used to identify the first text information in the first electronic medical record information through the Word2Vec model; obtain all lexical features in the first text information, and process the lexical features into lexical vectors to obtain the association between different lexical vectors; generate a first text feature based on the lexical vector and the association between different lexical vectors; wherein the first text feature is feature information composed of the text vector.

[0121] In one embodiment, the extraction unit 601 is specifically used to identify all image information in the first electronic medical record information through the normalization processing model; obtain the light brightness information of each image information, the light brightness information includes the light brightness information of multiple local areas in each image information, and the one image information includes the multiple local areas; perform normalization processing on the light brightness information of each image information to generate a first image feature, the first image feature includes multiple first sub-image features in the first electronic medical record information, and one first sub-image feature corresponds to one image information.

[0122] In one embodiment, the acquisition unit 603 is specifically used to extract the text sequence vector in the first text feature through the BERT model and the LSTM model, and process the first text feature according to the text sequence vector to generate a second text feature; extract the text sequence vector in the first image feature through the CNN model, and process the first image feature according to the text sequence vector to generate a second image feature.

[0123] In one embodiment, the production unit 604 is specifically used to identify the entity information and extract the entity relationships in the entity information through the BiLSTM-CRF model in the structured model, and determine the entity boundaries of the entity information; determine the key information in the entity information based on the entity information, the entity relationships in the entity information, and the entity boundaries, and generate structured information.

[0124] In one embodiment, the execution unit 602 is also used to obtain multiple groups of original electronic medical records and target data corresponding to the multiple groups of original electronic medical records, the target data including: target structured information, the structured information corresponding to the multiple groups of original electronic medical records is structured information in a target format corresponding to each sub-medical record in the multiple groups of original electronic medical records, and the multiple groups of original electronic medical records include data information of at least two users; input the multiple groups of original electronic medical records into an initial electronic medical record information extraction model to generate test data corresponding to the original electronic medical records, the test data including test structured information; compare the multiple groups of test data with the multiple groups of target data, and determine the difference between the multiple groups of test data and the multiple groups of target data; when the difference is greater than a preset threshold, adjust the initial electronic medical record information extraction model according to the difference, so that the difference between the test data generated by the electronic medical record information processing model and the target data is less than the preset threshold, and the initial electronic medical record information extraction model includes an initial Word2Vec model, a normalization model, a BERT model, an LSTM model, a CNN model, a multi-channel self-attention model, and a structured model.

[0125] It should be understood that the units recorded in the electronic medical record information extraction system correspond to the various steps in the method described in the accompanying drawings. Therefore, the operations and features described above for the method are also applicable to the electronic medical record information extraction system and the units contained therein, and will not be repeated here. The electronic medical record information extraction system can be pre-implemented in a browser or other security application of a computer device, or can be loaded into a browser or its security application of a computer device by downloading or the like. The corresponding units in the electronic medical record information extraction system can cooperate with the units in the computer device to implement the solution of the embodiment of the present application.

[0126] For the several modules or units mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into multiple modules or units to be embodied.

[0127] It should be noted that for details not disclosed in the electronic medical record information extraction system of the embodiment of the present application, please refer to the details disclosed in the above embodiments of the present application, which will not be repeated here.

[0128] Reference below Figure 6 , Figure 6 FIG. 1 is a schematic diagram showing a structure of a computer device suitable for implementing an embodiment of the present application. Figure 6 As shown, the computer system 1700 includes a central processing unit (CPU) 1701, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 1702 or the program loaded from the storage part 1708 to the random access memory (RAM) 1703. In the RAM 1703, various programs and data required for the operation instructions of the system are also stored. The CPU 1701, the ROM 1702, and the RAM 1703 are connected to each other through a bus 1704. An input / output (I / O) interface 1705 is also connected to the bus 1704.

[0129] The following components are connected to the I / O interface 1705: an input section 1706 including a keyboard, a mouse, etc.; an output section 1707 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 1708 including a hard disk, etc.; and a communication section 1709 including a network interface card such as a LAN card, a modem, etc. The communication section 1709 performs communication processing via a network such as the Internet. A drive 1710 is also connected to the I / O interface 1705 as needed. A removable medium 1711, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 1710 as needed, so that a computer program read therefrom is installed into the storage section 1708 as needed.

[0130] In particular, according to an embodiment of the present application, the above reference flow chart Figure 1 The described process can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 1709, and / or installed from a removable medium 1711. When the computer program is executed by the central processing unit (CPU) 1701, the above-mentioned functions defined in the system of the present application are executed.

[0131] It should be noted that the computer-readable medium shown in the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution unit, a device or a device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium such as a computer-readable storage medium that can send, propagate or transmit a program for use by or in conjunction with an instruction execution unit, device or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0132] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, functions and operating instructions of the system, method and computer program product according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the aforementioned module, program segment or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the box can also occur in a sequence different from that marked in the accompanying drawings. For example, the boxes represented by two connections can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operating instruction, or can be implemented with a combination of dedicated hardware and computer instructions.

[0133] The units or modules involved in the embodiments described in the present application may be implemented by software or hardware. The units or modules described may also be set in a processor. For example, it may be described as: a processor includes a first receiving module, a second receiving module, and a sending module. The names of these units or modules do not, in some cases, constitute limitations on the units or modules themselves.

[0134] As another aspect, the present application further provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiment, or may exist independently without being assembled into the electronic device. The above computer-readable storage medium stores one or more programs, and when the above programs are used by one or more processors to execute the method for extracting electronic medical record information described in the present application.

[0135] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the aforementioned disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present application (but not limited to) by each other to form a technical solution.

Claims

1. A method for extracting electronic medical record information, characterized in that: include: Extracting a first text feature in the first electronic medical record information through a first sub-model in the electronic medical record information extraction model, and extracting a first image feature in the first electronic medical record information through the first sub-model, wherein the first sub-model is a Word2Vec model and a normalization model; Processing the text features in the first electronic medical record information and the image features in the first electronic medical record information through the second sub-model in the electronic medical record information extraction model, identifying the second text features in the first text features and the second picture features in the first picture features, the second sub-model is a multi-task model, and the multi-task model includes a BERT model, an LSTM model, and a CNN model, the second text features are part of the text features or all of the text features in the first text features, and the second picture features are part of the picture features or all of the picture features in the first picture features; Recognize the second text feature and the second image feature through the third sub-model in the electronic medical record information extraction model to obtain entity information corresponding to the first electronic medical record information, wherein the entity information includes: user information, user status, and conclusive information of the user status, and the third sub-model includes a multi-channel self-attention model; The entity information is processed in a predefined structured format through the fourth sub-model in the electronic medical record information extraction model to generate structured information, and the structured information is used to display key information in the entity information in a preset format. The fourth sub-model is a structured model.

2. The method according to claim 1, characterized in that Before extracting the first text feature in the first electronic medical record information by using the first sub-model in the electronic medical record information extraction model, the method includes: Performing data preprocessing on the initial electronic medical record information according to a preset preprocessing method to obtain the first electronic medical record information; The preset preprocessing methods include: information segmentation, part-of-speech tagging, grayscale processing, and binarization processing.

3. The method according to claim 1, characterized in that: The extracting the first text feature in the first electronic medical record information by using the first sub-model in the electronic medical record information extraction model includes: Identify first text information in the first electronic medical record information by using a Word2Vec model; Acquire all lexical features in the first text information, process the lexical features into lexical vectors, and acquire correlations between different lexical vectors; generating a first text feature according to the vocabulary vector and the association relationship between different vocabulary vectors; The first text feature is feature information formed by a text vector.

4. The method according to claim 1, characterized in that: The extracting the first picture feature in the first electronic medical record information by using the second sub-model includes: Identify all image information in the first electronic medical record information through a normalization processing model; Acquire light brightness information of each picture information, wherein the light brightness information includes light brightness information of a plurality of local areas in each picture information, and one picture information includes the plurality of local areas; Normalization processing of light illuminance information is performed on each of the picture information to generate a first picture feature, wherein the first picture feature includes a plurality of first sub-picture features in the first electronic medical record information, and one first sub-picture feature corresponds to one picture information.

5. The method according to claim 1, characterized in that The processing of the first text feature in the first electronic medical record information and the image feature in the first electronic medical record information by the second sub-model in the electronic medical record information extraction model to identify the second text feature in the first text feature and the second picture feature in the first picture feature includes: Extracting a text sequence vector from the first text feature through a BERT model and an LSTM model, and processing the first text feature according to the text sequence vector to generate a second text feature; The text sequence vector in the first image feature is extracted through the CNN model, and the first image feature is processed according to the text sequence vector to generate a second image feature.

6. The method according to claim 1, characterized in that The fourth sub-model in the electronic medical record information extraction model processes the entity information according to a predefined structured format to generate structured information, including: By using the BiLSTM-CRF model in the structured model, the entity information is identified and the entity relationship in the entity information is extracted, and the entity boundary of the entity information is determined; According to the entity information, entity relationships in the entity information, and entity boundaries, key information in the entity information is determined, and structured information is generated.

7. The method according to claim 1, characterized in that The method further comprises: Acquire multiple sets of original electronic medical records and target data corresponding to the multiple sets of original electronic medical records, wherein the target data includes: target structured information, wherein the structured information corresponding to the multiple sets of original electronic medical records is structured information in a target format corresponding one-to-one to each sub-medical record in the multiple sets of original electronic medical records, and the multiple sets of original electronic medical records include data information of at least two users; Inputting the plurality of groups of original electronic medical records into an initial electronic medical record information extraction model to generate test data corresponding to the original electronic medical records, wherein the test data includes test structured information; Comparing the multiple sets of test data with the multiple sets of target data, and determining differences between the multiple sets of test data and the multiple sets of target data; When the difference is greater than a preset threshold, the initial electronic medical record information extraction model is adjusted according to the difference so that the difference between the test data and the target data generated by the electronic medical record information processing model is less than the preset threshold. The initial electronic medical record information extraction model includes an initial Word2Vec model, a normalized model, a BERT model, an LSTM model, a CNN model, a multi-channel self-attention model, and a structured model.

8. An electronic medical record information extraction system, characterized in that: include: an extraction unit, configured to extract a first text feature in the first electronic medical record information through a first sub-model in an electronic medical record information extraction model, and to extract a first image feature in the first electronic medical record information through the first sub-model, wherein the first sub-model is a Word2Vec model and a normalization model; an execution unit, configured to process text features in the first electronic medical record information and image features in the first electronic medical record information through a second sub-model in the electronic medical record information extraction model, and identify a second text feature in the first text feature and a second picture feature in the first picture feature, wherein the second sub-model is a multi-task model, and the multi-task model includes a BERT model, an LSTM model, and a CNN model, wherein the second text feature is a part of the text feature or all of the text features in the first text feature, and the second picture feature is a part of the picture feature or all of the picture features in the first picture feature; an acquisition unit, configured to recognize the second text feature and the second image feature through a third sub-model in the electronic medical record information extraction model, and acquire entity information corresponding to the first electronic medical record information, wherein the entity information includes: user information, user status, and conclusive information of the user status, and the third sub-model includes a multi-channel self-attention model; A generation unit is used to process the entity information in a predefined structured format through the fourth sub-model in the electronic medical record information extraction model to generate structured information, wherein the structured information is used to display key information in the entity information in a preset format, and the fourth sub-model is a structured model.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images