Intelligent electronic medical record management method and system based on big data

Abstract

The invention discloses an intelligent electronic medical record management method and system based on big data, and the method comprises the steps: obtaining medical image data of a patient, eliminating image noise in the medical image data through a Gaussian filter, and carrying out the feature extraction of a medical image after noise elimination through a convolutional neural network; detecting the medical image after feature extraction to obtain a focus area, generating annotation data according to the focus area, and associating with a medical record database through the annotation data; according to the association relationship between the annotation data and the medical record database, summarizing the annotation data and the patient information in the medical record database, filling a report template, converting the report template into report characters through a natural language processing model, issuing the report characters, and updating the medical record of the patient; according to the method, the Gaussian pyramid is established, the focus area in the image is detected, the annotation data is generated, and the annotation data is associated with the medical record database, so that the image processing efficiency and the focus detection precision are improved, and the electronic medical record management efficiency is enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent data management, and in particular to an intelligent electronic medical record management method and system based on big data. Background Art

[0002] With the rapid development of medical informatization, the Electronic Medical Records (EMR) system has gradually become an important part of modern medical services. Its core role is to reduce the use of paper records and improve the efficiency and quality of medical services by digitally storing and managing patients' diagnosis and treatment information. Especially in the context of the continuous advancement of technologies such as big data and artificial intelligence, data-driven intelligent electronic medical record systems are gradually moving into people's real lives, assisting doctors in making more accurate medical decisions.

[0003] At present, many hospitals have deployed intelligent systems based on medical image analysis and automated lesion detection. These systems combine advanced technologies such as deep learning and natural language processing, and can analyze patients' medical images through automated means and generate relevant diagnostic information. Although the existing technology has made significant progress in electronic medical record management and medical image analysis, there are still some technical bottlenecks that limit its effectiveness in practical applications. First, although the traditional medical image denoising method can effectively remove noise, it often sacrifices some detail information, resulting in a decrease in image quality, which in turn affects the accuracy of image feature extraction and lesion detection. Secondly, most of the existing lesion annotation and classification technologies rely on static image processing algorithms, lack dynamic processing of multi-scale and multi-resolution images, and cannot fully capture subtle changes in the image. This limitation leads to insufficient accuracy in lesion detection, especially when dealing with medical images with complex structures, it is often difficult to accurately locate and annotate the lesion area. In addition, when generating reports, the existing electronic medical record management system relies on natural language generation technology with preset templates and rules, lacks flexibility and personalized processing of report generation, resulting in the lack of professionalism and accuracy of the generated diagnostic reports. Especially when dealing with complex cases, the workload of doctors is still heavy. Summary of the invention

[0004] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0005] In view of the above existing problems, the present invention is proposed. Therefore, the present invention provides an intelligent electronic medical record management method based on big data to solve the problems mentioned in the background technology.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides an intelligent electronic medical record management method based on big data, comprising:

[0008] Acquire medical image data of the patient, use a Gaussian filter to eliminate image noise in the medical image data, and perform feature extraction on the medical image after noise elimination through a convolutional neural network;

[0009] Detecting the medical image after feature extraction to obtain the lesion area, generating annotation data according to the lesion area, and associating the annotation data with the medical record database;

[0010] According to the association between the annotated data and the medical record database, the annotated data and the patient information in the medical record database are summarized and filled in the report template. The report template is converted into report text through the natural language processing model for publication, and the patient's medical record is updated.

[0011] As a preferred solution of the intelligent electronic medical record management method based on big data described in the present invention, wherein: using a Gaussian filter to eliminate image noise in the medical image data includes:

[0012] Suppose the pixel coordinates of the medical image are (x, y), and the pixel value of the medical image data at the pixel coordinates (x, y) is I(x, y). The Gaussian filter is used to perform weighted averaging on each pixel in the medical image and its surrounding neighboring pixels, that is, the output value of each pixel is the weighted sum of itself and the surrounding pixels.

[0013] As a preferred solution of the intelligent electronic medical record management method based on big data described in the present invention, the feature extraction of the medical image after noise elimination is performed by a convolutional neural network, including:

[0014] The medical images after noise elimination are divided into a training set and a validation set, wherein the training set contains the medical image set I T (n);

[0015] The medical image set I T (n) input into the convolutional neural network, convolve each medical image in the medical image set through the convolution layer in the convolutional neural network, generate a corresponding medical image feature map by the convolution kernel, and flatten the medical image feature map to obtain a feature vector F of the medical image.T (n);

[0016] Where n represents the nth training image.

[0017] As a preferred solution of the intelligent electronic medical record management method based on big data described in the present invention, it also includes:

[0018] The medical image feature map extracted by the convolution kernel in the convolution layer is represented as F(I);

[0019] Calculate the extracted medical image feature map F(I) and the feature vector F of the medical image T (n), and according to the cosine similarity, the medical image with the highest cosine similarity is obtained as the medical image set I T (n) category field, repeat the cosine similarity calculation process until the medical image set I T (n) All medical images are classified;

[0020] A Gaussian pyramid is established to enhance the features of each classified medical image, and multi-resolution processing is performed on each layer of the Gaussian pyramid to generate medical images of different scales.

[0021] As a preferred solution of the intelligent electronic medical record management method based on big data described in the present invention, the medical image after feature extraction is detected to obtain the lesion area, and the annotation data is generated according to the lesion area, including:

[0022] Based on medical images of different scales, multiple candidate regions are generated, a confidence score and a confidence threshold are set for each candidate region, a bounding box B(x1, y1, x2, y2) is generated for each candidate region, and the bounding box is adjusted by a regression loss function;

[0023] When the confidence score exceeds the confidence threshold, the candidate region is a lesion region, and the position of the lesion region is marked to generate annotation data;

[0024] Among them, x1, y1 are the coordinates of the upper left corner, and x2, y2 are the coordinates of the lower right corner.

[0025] As a preferred solution of the intelligent electronic medical record management method based on big data described in the present invention, wherein: associating the annotated data with the medical record database includes:

[0026] The annotated data are aggregated as a sub-table, and each annotated data has a unique annotated ID, the medical record database is used as the main table, and an association relationship between the main table and the sub-table is established;

[0027] The association relationship between the main table and the sub-table includes:

[0028] The patient ID in the medical record database is used as the primary key, and the summarized annotation data, that is, the annotation ID in the child table, is used as the foreign key;

[0029] The primary key of the main table is connected to the foreign key of the sub-table to obtain a one-to-many relationship between the primary keys of the main table and the sub-table.

[0030] As a preferred solution of the intelligent electronic medical record management method based on big data described in the present invention, wherein: according to the association relationship between the marked data and the medical record database, the marked data and the patient information in the medical record database are summarized and filled in the report template, and the report template is converted into report text through the natural language processing model for publication, and the patient medical record is updated, including:

[0031] Before filling in the report template, sensitive information about the patient's condition is obtained based on the patient's medical records in the medical record database and the patient's current lesions in the annotated data;

[0032] Upload the sensitive information to the doctor's review list, and the doctor will decide whether to fill in the report template with the sensitive information;

[0033] When the report is released, the newly generated report will be updated to the medical record database, and the report will be compared with the historical medical records in the medical record database to generate a trend chart of the disease change for the doctor's reference.

[0034] In a second aspect, the present invention provides an intelligent electronic medical record management system based on big data, which includes:

[0035] The medical image preprocessing and feature extraction module is configured to obtain medical image data of the patient, eliminate image noise in the medical image data using a Gaussian filter, and perform feature extraction on the medical image after the noise is eliminated using a convolutional neural network;

[0036] The lesion detection and annotation generation module is configured to detect the medical image after feature extraction, obtain the lesion area, generate annotation data according to the lesion area, and associate the annotation data with the medical record database;

[0037] The report generation and medical record update module is configured to summarize the patient information in the annotated data and the medical record database according to the association between the annotated data and the medical record database, fill in the report template, convert the report template into report text through the natural language processing model for publication, and update the patient's medical record.

[0038] In a third aspect, the present invention provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, wherein: the processor implements any step of the above method when executing the computer program.

[0039] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: the computer program implements any step of the above method when executed by a processor.

[0040] Compared with the prior art, the invention has the following beneficial effects:

[0041] 1. The present invention uses Gaussian filters and convolutional neural networks (CNNs) to not only effectively remove noise, but also retain key details of the image, greatly improving the quality of the processed image, making subsequent feature extraction more accurate, solving the problem of sacrificing image details for image denoising in the prior art, and providing a more reliable imaging basis for the accuracy of lesion detection;

[0042] 2. Gaussian pyramids are established to enhance the features of images of different scales, so that the system can detect the features of lesions at all levels. At the same time, combined with the candidate region generation and confidence scoring mechanism, the system can dynamically adjust the bounding box according to the regression loss function, making the positioning of the lesion area more accurate, effectively solving the problem of inaccurate lesion positioning in the existing technology, especially when processing multi-scale images, making the system more robust;

[0043] 3. Automatically associate the annotated data with the medical record database. Combined with the one-to-many relationship between the main table (medical record database) and the sub-table (annotated data), the annotated information and medical records can be automatically summarized, allowing doctors to easily view relevant image annotated data in the electronic medical record, avoiding the inefficiency of manual data matching in the existing system and improving the efficiency and accuracy of information processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0045] Figure 1 This is an overall flow chart of an intelligent electronic medical record management method based on big data according to an embodiment of the present invention;

[0046] Figure 2 A three-dimensional schematic diagram of flattening processing of an intelligent electronic medical record management method based on big data according to an embodiment of the present invention;

[0047] Figure 3 A loss change curve diagram of the candidate area before and after adjustment of the intelligent electronic medical record management method based on big data according to an embodiment of the present invention;

[0048] Figure 4 This is a relationship diagram between the confidence and loss function of the intelligent electronic medical record management method based on big data described in one embodiment of the present invention. DETAILED DESCRIPTION

[0049] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0050] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0051] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0052] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0053] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0054] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0055] Example 1

[0056] Reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, and provides an intelligent electronic medical record management method based on big data, including:

[0057] S1. Obtain the patient's medical imaging data, use a Gaussian filter to eliminate image noise in the medical imaging data, and perform feature extraction on the medical imaging data after noise elimination through a convolutional neural network;

[0058] Specifically, there are three ways to obtain medical imaging data of patients. The first is direct acquisition, which directly generates imaging data through medical imaging equipment (such as CT machines, MRI equipment). That is, during the examination, the equipment will record the patient's scan results in the form of images and store them in DICOM (Digital Imaging and Communications) format. The second is manual upload. If the imaging data comes from other medical institutions or is stored in the form of CDs, USB flash drives, hard drives, etc., the doctor will import these imaging data into the system through manual upload. The third is remote acquisition. The doctor uses the telemedicine system to input the patient's identity information and obtain imaging data from external medical institutions through a secure network.

[0059] It should be explained that Gaussian filter is an image smoothing technology based on normal distribution (Gaussian distribution), which is mainly used to remove noise in images; in the processing of medical images, noise often appears as mutations or random interference that are inconsistent with surrounding pixels;

[0060] Furthermore, let the pixel coordinates of the medical image be (x, y), and the pixel value of the medical image data at the pixel coordinates (x, y) be I(x, y). Each pixel in the medical image and its surrounding neighboring pixels are weighted averaged through a Gaussian filter, that is, the output value of each pixel is the weighted sum of itself and the surrounding pixels.

[0061] Specifically, the Gaussian filter applies the weighted formula to each pixel (x, y) as follows:

[0062]

[0063] Where I′(x,y) is the pixel value after Gaussian filtering, I(x+i,y+j) is the pixel value in the vicinity of the pixel point (x,y), G(i,j) is the weight value of the Gaussian kernel, and i,j are the pixel offsets in the horizontal (horizontal axis) and vertical (vertical axis) directions, respectively, and k is the radius of the filter (i.e., the range of neighboring pixels). For example, a 3×3 filter means k=1, a 5×5 filter means k=2, and so on;

[0064] Specifically, if i=0, it means that the pixel is in the same column as the current pixel, if i=-1, it means the adjacent pixel on the left, and if i=1, it means the adjacent pixel on the right; if j=0, it means that the pixel is in the same row as the current pixel, if j=-1, it means the adjacent pixel above, and if j=1, it means the adjacent pixel below;

[0065] It should be noted that the weights of the weighted formula are determined by the weight value of the Gaussian kernel;

[0066] Furthermore, the weight value G(i,j) of the Gaussian kernel is expressed as:

[0067]

[0068] Among them, i 2 +j 2 It represents the square of the distance from the center pixel. The farther the distance, the smaller the weight. σ represents the width of the Gaussian kernel. The larger the width, the more obvious the smoothing effect of the filter.

[0069] Furthermore, the medical images after noise elimination are divided into a training set and a validation set. The training set contains the medical image set I T (n);

[0070] Furthermore, the medical image collection T (n) Input into the convolutional neural network, convolve each medical image in the medical image set through the convolution layer in the convolutional neural network, generate the corresponding medical image feature map by the convolution kernel, and flatten the medical image feature map (refer to Figure 2 ), and obtain the feature vector F of the medical image T (n);

[0071] It should be noted that at this time, the feature vector F of the medical image T (n) describes the key features of the image, such as edge, texture, and shape;

[0072] Where n represents the nth training image;

[0073] Furthermore, the medical image feature map extracted by the convolution kernel in the convolution layer is denoted as F(I);

[0074] Furthermore, the extracted medical image feature map F(I) and the feature vector F of the medical image are calculated. T According to the cosine similarity, the medical image with the highest cosine similarity is obtained as the medical image set I T (n) category field, repeat the cosine similarity calculation process until the medical image set I T (n) All medical images are classified;

[0075] Specifically, the cosine similarity SL(F(I),F T The calculation formula of (n)) is expressed as:

[0076]

[0077] Among them, ‖·‖ represents the norm of the feature vector or feature map (i.e., the length of the vector or the size of the feature map);

[0078] It should be noted that, through the obtained cosine similarity, the system will select the image with the highest similarity as the most likely category of the input image, and the medical image set I can be obtained by repeating the process. T (n) all category fields, thereby achieving the classification of medical images;

[0079] Furthermore, a Gaussian pyramid is established to enhance the features of each classified medical image, and multi-resolution processing is performed at each layer of the Gaussian pyramid to generate medical images of different scales;

[0080] Specifically, let G n (I′)=I′ is the nth level Gaussian pyramid image, defined as:

[0081]

[0082] Among them, w i,j is the weight matrix;

[0083] Furthermore, for each layer of G n (I′), the horizontal edge G is extracted by Sobel filter x (x,y) and the vertical edge G y (x,y), we get:

[0084]

[0085] Among them, S x and S y They are Sobel operators in horizontal and vertical directions respectively;

[0086] It should be noted that the extracted horizontal and vertical edges can be reorganized into a medical image feature map, so that the system can capture the global structure and focus on local details, making the enhanced image features clearer;

[0087] S2, detecting the medical image after feature extraction to obtain the lesion area, generating annotation data according to the lesion area, and associating the annotation data with the medical record database;

[0088] Furthermore, based on medical images of different scales, multiple candidate regions are generated, the confidence score and confidence threshold of each candidate region are set, a bounding box B(x1, y1, x2, y2) is generated for each candidate region, and the bounding box is adjusted by the regression loss function;

[0089] Specifically, the regression loss function is defined as follows:

[0090]

[0091] Furthermore, when the confidence score exceeds the confidence threshold, the candidate region is a lesion region, and the location of the lesion region is marked to generate annotation data;

[0092] Among them, x1, y1 are the coordinates of the upper left corner, x2, y2 are the coordinates of the lower right corner, and the confidence range is [0, 1];

[0093] Furthermore, after lesion detection and labeling, the system evaluates the similarity of different medical images and updates the Gaussian pyramid by calculating the Euclidean distance between the medical feature vector and the medical feature map;

[0094] Specifically, the Euclidean distance d(F(I),F) between the medical feature vector and the medical feature graph is calculated. T The formula for (n) is:

[0095]

[0096] Among them, F(I) k represents the kth eigenvalue in the medical feature vector F(I), F T (n) k The feature vector F representing the nth training image T (n) is the kth eigenvalue in (n); m represents the dimension of the eigenvector (i.e., the number of features);

[0097] It should be noted that calculating the Euclidean distance between the medical feature vector and the medical feature map helps the system filter out irrelevant false positive detection results and improve the accuracy of detection. In addition, for small or blurred lesion areas, evaluating the similarity of different medical images can assist the system in locating the real lesions and reduce missed reports.

[0098] Furthermore, the annotated data is summarized as a sub-table, and each annotated data has a unique annotation ID. The medical record database is used as the main table, and the association relationship between the main table and the sub-table is established;

[0099] Specifically, the relationship between the main table and the sub-table includes:

[0100] The patient ID in the medical record database is used as the primary key, and the summarized annotation data, that is, the annotation ID in the child table, is used as the foreign key;

[0101] Connect the primary key of the main table with the foreign key of the subtable to obtain a one-to-many relationship between the primary keys of the main table and the subtable;

[0102] For example, assume that the patient's medical record information is stored in a medical record database, the medical image data and related annotations are stored in an annotation data table, a patient has undergone multiple image scans, and each scan contains multiple annotation data, wherein Table 1 and Table 2 represent the main table (medical record database) and the sub-table (annotation database), respectively;

[0103] Table 1 Main table (medical record database)

[0104] Patient ID Patient Name Date of Birth Diagnostic Information 101 Zhang San 1980-02-15 Early stage lung cancer 102 Li Si 1975-09-12 Pulmonary nodules

[0105] Table 2 Subtable (annotation database)

[0106] Label ID Patient ID Image ID Lesion location Confidence Date 1 101 CT001 Left lower lobe 0.95 2024-10-01 2 101 CT002 Right middle lobe 0.87 2024-10-05 3 102 CT003 Left upper lobe 0.92 2024-10-02

[0107] As can be seen from Tables 1 and 2, patient Zhang San (patient ID is 101) underwent two imaging scans on October 1 and October 5, 2024, and the lesions in the left lower lobe and right middle lobe were annotated respectively; the system established an association through the patient ID (primary key) and the annotation ID (foreign key), so that the doctor only needs to query all the image annotation data of the patient through the medical record database (primary table);

[0108] It should be noted that the interconnection between databases not only improves the query efficiency, but also ensures the integrity and reliability of medical record data and annotation data;

[0109] S3. According to the association between the annotated data and the medical record database, the annotated data and the patient information in the medical record database are summarized and filled in the report template. The report template is converted into report text through the natural language processing model for publication, and the patient medical record is updated;

[0110] Furthermore, before filling in the report template, sensitive information of the patient's condition is obtained based on the patient's medical record in the medical record database and the patient's current lesion in the annotated data;

[0111] Specifically, sensitive information includes infectious diseases (such as hepatitis B), psychological and mental illnesses (such as depression), genetic diseases (such as BRCA gene mutations), and cancers (such as liver cancer).

[0112] Furthermore, sensitive information is uploaded to the doctor's review list, and the doctor decides whether to fill in the report template;

[0113] Furthermore, when the report is released, the newly generated report is updated to the medical record database, and the report is compared with the historical medical records in the medical record database to generate a chart of the trend of the disease change for the doctor's reference;

[0114] It should be noted that the report review mechanism allows doctors to control patients' sensitive information and ensure the reasonable release of information. It also provides an intuitive tool for visualizing changes in the patient's condition by combining the patient's historical medical records with the latest generated reports. It not only provides doctors with a more intuitive way to monitor the condition, but also enables doctors to more quickly judge the patient's condition evolution trend, helping to make clinical decisions more scientific and efficient.

[0115] Furthermore, this embodiment also provides an intelligent electronic medical record management system based on big data, including:

[0116] The medical image preprocessing and feature extraction module is configured to obtain medical image data of the patient, eliminate image noise in the medical image data using a Gaussian filter, and perform feature extraction on the medical image after the noise is eliminated using a convolutional neural network;

[0117] The lesion detection and annotation generation module is configured to detect the medical image after feature extraction, obtain the lesion area, generate annotation data according to the lesion area, and associate the annotation data with the medical record database;

[0118] The report generation and medical record update module is configured to summarize the patient information in the annotated data and the medical record database according to the association between the annotated data and the medical record database, fill in the report template, convert the report template into report text through the natural language processing model for publication, and update the patient's medical record.

[0119] This embodiment also provides a computer device, which is applicable to the case of an intelligent electronic medical record management method based on big data, including:

[0120] Memory and processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the intelligent electronic medical record management method based on big data as proposed in the above embodiment.

[0121] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile 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 the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0122] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the intelligent electronic medical record management method based on big data proposed in the above embodiment is implemented.

[0123] The storage medium proposed in this embodiment and the data storage method proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be found in the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0124] Example 2

[0125] Reference Figure 3 and Figure 4 , which is the second embodiment of the present invention, and this embodiment provides an intelligent electronic medical record management method based on big data, including: In order to verify the beneficial effects of the solution of the present invention, the following simulation experiment was carried out;

[0126] The simulation experiment selected 500 patients, and processed the data using the existing image processing system and the system based on the present invention. The medical image data came from CT scans and MRI scans.

[0127] In the image processing stage, Gaussian filters are used to process the image data to eliminate the interference of random noise and ensure the preservation of image details; feature extraction is performed through convolutional neural networks to obtain the edge, texture and shape features of the image; after feature extraction, the features in the image are enhanced based on multi-scale Gaussian pyramids;

[0128] In the lesion detection stage, the bounding box of the detected candidate area is adjusted through the confidence threshold and regression loss function; then, the annotated data is summarized and connected to the medical record database, using the patient ID as the primary key and the annotation ID as the foreign key, and managing the lesion annotation information and patient medical record data through a one-to-many relationship;

[0129] Finally, the simulation experiment generates a diagnosis report for each case based on the report template generated by the system, and uses natural language processing (NLP) technology to convert the analysis results into text, automatically update the medical records and generate a trend chart of the disease change. Before updating the medical records, doctors review the reports of some sensitive cases (such as cancer and genetic diseases) to ensure the privacy of patients.

[0130] The simulation experiment data are shown in Table 3;

[0131] Table 3 Average indicators of the method of the present invention and the traditional method

[0132]

[0133] As can be seen from Table 3, the average denoising accuracy of the system based on the invention is 95%, which is much higher than the 82% of the traditional system, an increase of 13%. This shows that through Gaussian pyramid processing, the system can capture the global structure and pay attention to local details. In terms of lesion detection, the accuracy of the scheme of the invention is 90%, which is significantly higher than the 75% of the traditional method. This shows that calculating the Euclidean distance between the medical feature vector and the medical feature map can greatly improve the accuracy of lesion detection, especially in processing complex structures.

[0134] Finally, according to Figure 3 Through this curve, we can observe that the loss value of the candidate region after adjustment is significantly reduced, which proves the effectiveness of optimizing the bounding box through the regression loss function, and the corresponding high-confidence candidate region has a better performance in the optimization process. Figure 4 You can see that the loss value is significantly reduced;

[0135] In summary, simulation experiments have verified the effectiveness of the present invention in optimizing bounding boxes and improving lesion detection accuracy. Especially when processing medical images with complex structures, the system can greatly improve the detection accuracy and has good application prospects.

[0136] Those skilled in the art will appreciate that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the application can adopt the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program codes. The scheme in the embodiments of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0137] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0138] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0139] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0140] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0141] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. An intelligent electronic medical record management method based on big data, characterized in that: include: Acquire medical image data of the patient, use a Gaussian filter to eliminate image noise in the medical image data, and perform feature extraction on the medical image after noise elimination through a convolutional neural network; Detecting the medical image after feature extraction to obtain the lesion area, generating annotation data according to the lesion area, and associating the annotation data with the medical record database; According to the association between the annotated data and the medical record database, the annotated data and the patient information in the medical record database are summarized and filled in the report template. The report template is converted into report text through the natural language processing model for publication, and the patient's medical record is updated.

2. The intelligent electronic medical record management method based on big data as claimed in claim 1, characterized in that: Using a Gaussian filter to eliminate image noise in the medical image data includes: Suppose the pixel coordinates of the medical image are (x, y), and the pixel value of the medical image data at the pixel coordinates (x, y) is I(x, y). The Gaussian filter is used to perform weighted averaging on each pixel in the medical image and its surrounding neighboring pixels, that is, the output value of each pixel is the weighted sum of itself and the surrounding pixels.

3. The intelligent electronic medical record management method based on big data as claimed in claim 2, characterized in that: The convolutional neural network is used to extract features from medical images after noise removal, including: The medical images after noise elimination are divided into a training set and a validation set, wherein the training set contains the medical image set I T (n); The medical image set I T (n) input into the convolutional neural network, convolve each medical image in the medical image set through the convolution layer in the convolutional neural network, generate a corresponding medical image feature map by the convolution kernel, and flatten the medical image feature map to obtain a feature vector F of the medical image. T (n); Where n represents the nth training image.

4. The intelligent electronic medical record management method based on big data as claimed in claim 3, characterized in that: Also includes: The medical image feature map extracted by the convolution kernel in the convolution layer is represented as F(I); Calculate the extracted medical image feature map F(I) and the feature vector F of the medical image T (n), and according to the cosine similarity, the medical image with the highest cosine similarity is obtained as the medical image set I T (n) category field, repeat the cosine similarity calculation process until the medical image set I T (n) All medical images are classified; A Gaussian pyramid is established to enhance the features of each classified medical image, and multi-resolution processing is performed on each layer of the Gaussian pyramid to generate medical images of different scales.

5. The intelligent electronic medical record management method based on big data according to claim 1 or 4, characterized in that: Detecting the medical image after feature extraction to obtain the lesion area, and generating annotation data according to the lesion area, including: Based on medical images of different scales, multiple candidate regions are generated, a confidence score and a confidence threshold are set for each candidate region, a bounding box B(x1, y1, x2, y2) is generated for each candidate region, and the bounding box is adjusted by a regression loss function; When the confidence score exceeds the confidence threshold, the candidate region is a lesion region, and the position of the lesion region is marked to generate annotation data; Among them, x1, y1 are the coordinates of the upper left corner, and x2, y2 are the coordinates of the lower right corner.

6. The intelligent electronic medical record management method based on big data as claimed in claim 5, characterized in that: Associating the annotated data with the medical record database includes: The annotated data are aggregated as a sub-table, and each annotated data has a unique annotated ID, the medical record database is used as the main table, and an association relationship between the main table and the sub-table is established; The association relationship between the main table and the sub-table includes: The patient ID in the medical record database is used as the primary key, and the summarized annotation data, that is, the annotation ID in the child table, is used as the foreign key; The primary key of the main table is connected to the foreign key of the sub-table to obtain a one-to-many relationship between the primary keys of the main table and the sub-table.

7. The intelligent electronic medical record management method based on big data according to claim 6, characterized in that: According to the association between the annotated data and the medical record database, the annotated data and the patient information in the medical record database are summarized and filled in the report template, and the report template is converted into report text through the natural language processing model for publication, and the patient medical record is updated, including: Before filling in the report template, sensitive information about the patient's condition is obtained based on the patient's medical records in the medical record database and the patient's current lesions in the annotated data; Upload the sensitive information to the doctor's review list, and the doctor will decide whether to fill in the report template with the sensitive information; When the report is released, the newly generated report will be updated to the medical record database, and the report will be compared with the historical medical records in the medical record database to generate a trend chart of the disease change for the doctor's reference.

8. An intelligent electronic medical record management system based on big data, based on the intelligent electronic medical record management method based on big data according to any one of claims 1 to 7, characterized in that: include: The medical image preprocessing and feature extraction module is configured to obtain medical image data of the patient, eliminate image noise in the medical image data using a Gaussian filter, and perform feature extraction on the medical image after the noise is eliminated using a convolutional neural network; The lesion detection and annotation generation module is configured to detect the medical image after feature extraction, obtain the lesion area, generate annotation data according to the lesion area, and associate the annotation data with the medical record database; The report generation and medical record update module is configured to summarize the patient information in the annotated data and the medical record database according to the association between the annotated data and the medical record database, fill in the report template, convert the report template into report text through the natural language processing model for publication, and update the patient's medical record.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

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

Images