Medical record big data-based eye disease medication characteristic intelligent processing method and system

Abstract

The invention discloses an eye disease medication characteristic intelligent processing method and system based on medical record big data, and belongs to the technical field of medical record data processing. Eye images of a patient before and after medication are obtained through a wearable eye scanner, and a focus feature group containing a pixel gray average value and a shape area value is constructed through grid division and focus feature marking and extraction; calculating focus state attribute values based on the mathematical model, marking disease states and analyzing trends before and after medication; the correlation trend between the medical records is mined through trend tagging and digital matrix comparison; according to the method, quantitative analysis of the eye disease medication effect and intelligent association of medical record data are achieved, the objectivity of medication evaluation and the making efficiency of a personalized treatment scheme are improved, and the method is suitable for clinical auxiliary diagnosis of the ophthalmology department and self-health management of patients.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical record data processing, in particular to an eye disease medication feature intelligent processing method and system based on medical record big data. BACKGROUND

[0002] The treatment effect of eye diseases (such as glaucoma, conjunctivitis, macular degeneration, etc.) is highly dependent on the rationality of the medication scheme, and the evaluation of the medication effect is the key basis for adjusting the scheme. In the prior art, the following deficiencies exist in the evaluation of the medication effect:

[0003] Strong subjectivity, lack of quantitative standard: Traditional evaluation relies on doctors' naked-eye observation of eye images or patients' subjective description, which is difficult to accurately quantify the lesion changes and is easily affected by experience differences, resulting in poor consistency of evaluation results. For example, for the "improvement" of the lesion of a patient with keratitis after medication, different doctors may draw different conclusions due to different judgment standards.

[0004] Data is scattered, and it is difficult to mine the correlation trend: patient medical record data is stored in the form of text or isolated images, lacking structured processing, making it difficult to mine the medication response rules of similar diseases from massive medical records. For example, it is difficult to quickly find historical medical records with the same lesion change trend as a glaucoma patient after medication, resulting in delayed personalized medication recommendations.

[0005] Insufficient real-time, not timely adjustment: The lesion changes before and after medication need to be recorded through regular re-examination, which cannot be tracked in real time, and the best adjustment opportunity may be missed. For example, if a patient uses hormone eye drops and the lesion spreads, if not monitored in time, it may lead to the aggravation of the disease.

[0006] Single feature dimension: Existing technologies focus on a single feature of the lesion (such as area), ignoring details such as gray scale changes, resulting in one-sided evaluation. For example, the area of some lesions does not change but the gray scale deepens, which may indicate active inflammation, but traditional methods are prone to miss. SUMMARY

[0007] The purpose of the present application is to provide an eye disease medication feature intelligent processing method and system based on medical record big data to solve the problems raised in the background art.

[0008] In order to solve the above technical problems, the present application provides the following technical solutions:

[0009] An eye disease medication feature intelligent processing system based on medical record big data, the system comprises:

[0010] An image acquisition and medical record generation module for acquiring eye images before and after medication of a patient, dividing a background grid, and generating an eye disease medical record;

[0011] a feature marking and extraction module, configured to mark features of the background grid with the lesion points before and after the medication, generate a medical record feature group and extract a lesion feature group;

[0012] a disease state evaluation and trend analysis module, configured to evaluate disease state attribute values of the lesion points based on the lesion feature group, mark the disease state and analyze the disease trend before and after the medication;

[0013] a trend correlation determination module, configured to perform trend labeling processing on the background grid, construct a digital one-dimensional matrix and determine the trend correlation between the eye disease medical records.

[0014] Further, the image acquisition and medical record generation module comprises:

[0015] a wearable image acquisition unit, configured to acquire eye images of the patient before and after the medication through a wearable eye scanner, and the time interval between the before and after medication scans is a fixed period;

[0016] a grid division unit, configured to divide the eye images into background grids according to a fixed pixel size;

[0017] a medical record generation unit, configured to integrate the before-medication eye images and the after-medication eye images to generate an eye disease medical record containing corresponding image information.

[0018] Further, the feature marking and extraction module comprises:

[0019] a lesion feature set marking unit, configured to mark features of the background grid with the lesion points before and after the medication, generate a before-medication lesion feature set and an after-medication lesion feature set, and form a medical record feature group;

[0020] a pixel and area extraction unit, configured to acquire the average pixel gray value of the background grid with the lesion points and the shape area value of the lesion points based on the medical record feature group, and form a lesion feature group.

[0021] Further, the disease state evaluation and trend analysis module comprises:

[0022] a state attribute value calculation unit, configured to calculate disease state attribute values of the lesion points based on the lesion feature group;

[0023] a disease state marking unit, configured to compare the disease state attribute values with preset standard values, and mark the lesion points as a repair state or an active state;

[0024] a trend analysis unit, configured to compare the disease state attribute values before and after the medication, and analyze the disease trend of the background grid with the lesion points.

[0025] Further, the trend correlation determination module comprises:

[0026] a trend labeling unit configured to label a background grid based on a disease trend and set a label value;

[0027] a matrix construction unit configured to construct a digital one-dimensional matrix of an eye disease medical record based on the trend label value, and set a matrix element as the trend label value of the background grid;

[0028] a correlation determination unit configured to compare the digital one-dimensional matrices of different medical records, count the proportion of equal elements at the same position, and determine that the two medical records are correlated in trend and output the result if the proportion meets a preset trend correlation index value.

[0029] The method for intelligent processing of eye disease medication characteristics based on medical record big data comprises the following steps:

[0030] Step S1: obtaining eye images of a patient before and after medication by using a wearable eye scanner, dividing the eye images into background grids, and generating an eye disease medical record comprising the eye images before and after medication;

[0031] Step S2: marking the features of the background grids with lesions before and after medication, generating a lesion feature set before medication and a lesion feature set after medication to form a medical record feature group, and extracting the average pixel gray value and shape area value of the lesion points based on the medical record feature group to form the lesion feature group;

[0032] Step S3: evaluating the disease state attribute value of the lesion points based on the lesion feature group, marking the disease state of the lesion points by comparing with a preset standard value, and analyzing the disease trend of the background grids with lesions before and after medication;

[0033] Step S4: performing trend labeling on the background grids according to the disease trend, constructing a digital one-dimensional matrix corresponding to the eye disease medical record, comparing the digital one-dimensional matrices of different medical records, and determining and outputting the eye disease medical records with trend correlation.

[0034] Further, the specific implementation process of step S1 comprises:

[0035] obtaining the eye images of the patient by using the wearable eye scanner, and dividing the eye images of the patient into background grids according to fixed pixel size, thereby obtaining G background grids;

[0036] The patient uses the wearable eye scanner to scan the eye images before and after medication at each time of medication, and generates an eye disease medical record, which comprises the eye images before and after medication, wherein the time interval between the use of the wearable eye scanner to scan the eye images before and after medication is a fixed period.

[0037] Further, the specific implementation process of the step S2 includes:

[0038] Let any gth background grid be denoted as B g At each time of using the wearable eye scanner to perform eye image scanning, if the background grid B g has a lesion point, then the background grid B g with the lesion point is marked with feature distinction before and after medication to generate a lesion feature set LC1 = {B g | g ∈ [1, G]} before medication and a lesion feature set LC2 = {B g | g ∈ [1, G]} after medication respectively, and form a medical record feature group, denoted as MC i : [LC1, LC2], wherein i is the index number of the eye disease medical record, and MC i is the i th eye disease medical record.

[0039] Based on the medical record feature group, the pixel gray average value of the background grid with the lesion point and the shape area value of the lesion point in the background grid are obtained respectively to form a lesion feature group, denoted as [avg, S], wherein represents the background grid belonging to the lesion feature set before medication or the lesion feature set after medication, x = 1 or 2, avg represents the pixel gray average value, and S represents the shape area value.

[0040] Further, the specific implementation process of the step S3 includes:

[0041] Based on the lesion feature group, the disease state attribute value of the lesion point is evaluated The disease state attribute value is compared with a preset disease state standard value, if the disease state attribute value is greater than or equal to the disease state standard value, then the disease state mark of the lesion point in the lesion feature group [avg, S] is feedback as the repair state, if the disease state attribute value is less than the disease state standard value, then the disease state mark of the lesion point in the lesion feature group [avg, S] is feedback as the active state.

[0042] Based on the medical record feature group, the disease trend of the background grid with the lesion point before and after medication is analyzed, if then it indicates that the disease trend of the background grid B g with the lesion point is approaching the active state, if then it indicates that the disease trend of the background grid B g with the lesion point is approaching the repair state, if then it indicates that the disease trend of the background grid B g with the lesion point is stable.

[0043] Further, the specific implementation process of the step S4 includes:

[0044] Based on the disease trend under the pre-drug and post-drug conditions, the background grid B g An additional trend label is added, the trend label includes approaching to active state, approaching to repair state and stable, the trend label value is set, if the trend label is approaching to active state, the background grid B g The corresponding trend label value is -1, if the trend label is approaching to repair state, the background grid B g The corresponding trend label value is 1, if the trend label is stable, the background grid B g The corresponding trend label value is 0;

[0045] Based on the trend label value of the background grid B g , a digital one-dimensional matrix of the eye disease medical record is constructed, and the gth column matrix element of the digital one-dimensional matrix is the trend label value of the background grid B g , the corresponding generated digital one-dimensional matrix of the eye disease medical record MC i is recorded as R(MC i ), and the jth eye disease medical record MC j is obtained. j ;

[0046] The trend label values of the same matrix element positions between the digital one-dimensional matrix R(MC i ) and the digital one-dimensional matrix R(MC j ) are compared, the proportion of the number of equal matrix elements is counted, if the proportion meets the set trend correlation index value, it is determined that there is a trend correlation between the eye disease medical record MC i and the eye disease medical record MC j , and the eye disease medical record with the trend correlation is output.

[0047] Compared with the prior art, the beneficial effects achieved by the present application are: in the eye disease medication feature intelligent processing method and system based on medical record big data provided by the present application, the eye images of the patient before and after medication are obtained by the wearable eye scanner, the lesion feature group containing the average pixel gray value and the shape area value is constructed through grid division, lesion feature marking and extraction, the lesion state attribute value is calculated based on the mathematical model, the disease state is marked and the trend before and after medication is analyzed, the correlation trend between medical records is mined through trend labeling and digital matrix comparison, the present application realizes the quantitative analysis of the eye disease medication effect and the intelligent correlation of the medical record data, improves the objectivity of the medication evaluation and the efficiency of the personalized treatment scheme, and is suitable for ophthalmic clinical auxiliary diagnosis and patient self-health management. BRIEF DESCRIPTION OF DRAWINGS

[0048] The accompanying drawings are included to provide a further understanding of the application and are incorporated in and constitute a part of the specification, illustrate embodiments of the application and are used to explain the application, but are not intended to limit the application.

[0049] Figure 1 is a step schematic diagram of the eye disease drug feature intelligent processing method based on medical record big data of the application. DETAILED DESCRIPTION

[0050] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the application.

[0051] The application is based on the technical logic of "quantitative feature-state evaluation-trend correlation", and the core principles include:

[0052] The eye image is divided into a fixed grid, the abstract "lesion change" is converted into a calculable digital feature through the pixel gray average value (reflecting the lesion density or inflammation degree) and the shape area value (reflecting the lesion size), and the drug effect evaluation is converted from "qualitative description" to "quantitative value", reducing human error and solving subjective problems. For example, the AS value of a certain conjunctivitis patient after taking medicine increases from 30 to 60 (preset standard value 50), which can be clearly judged as "repair state". At the same time, based on the fixed period scanning (such as 24 hours / once) of the wearable device, the lesion change is recorded in real time, avoiding the delay caused by the interval between reexaminations. For example, the AS value of a glaucoma patient is found to decrease after taking medicine for 12 hours, which can be intervened in advance.

[0053] Fusion of gray scale and area features, quantification of disease state: the higher the AS value, the more obvious the lesion repair trend (high gray scale and small area), and vice versa, to avoid misjudgment in a single dimension. For example, the gray scale of a certain lesion increases from 80 to 120, and the AS value increases, indicating that the inflammation is reduced. The traditional method of looking at the area only is prone to missed judgment.

[0054] The lesion trend of each medical record is converted into a digital one-dimensional matrix, and the similarity of the matrix elements is compared to realize the rapid mining of similar drug response rules in a large number of medical records, provide a reference for doctors, and improve the correlation efficiency based on the pattern recognition principle. For example, the medical record matrix of a newly diagnosed keratitis patient has a matching degree of 85% with the medical record of 50 cases of using "Levofloxacin Eye Drops", which can be preferentially recommended.

[0055] In the first embodiment: provide an intelligent processing system for eye disease drug characteristics based on medical record big data, the system comprises:

[0056] An image acquisition and medical record generation module is configured to acquire eye images before and after drug use of a patient, divide a background grid, and generate an eye disease medical record.

[0057] The image acquisition and medical record generation module comprises:

[0058] A wearable image acquisition unit is configured to acquire eye images before and after drug use of a patient through a wearable eye scanner, and the time interval between the before and after drug use scans is a fixed period.

[0059] A grid division unit is configured to divide the eye images into a background grid according to a fixed pixel size.

[0060] A medical record generation unit is configured to integrate the eye images before and after drug use, and generate an eye disease medical record containing corresponding image information.

[0061] A feature marking and extraction module is configured to mark the before and after drug use characteristics of the background grid with a lesion point, generate a medical record feature group, and extract a lesion feature group.

[0062] The feature marking and extraction module comprises:

[0063] A lesion feature set marking unit is configured to mark the before and after drug use characteristics of the background grid with a lesion point, generate a before-drug-use lesion feature set and an after-drug-use lesion feature set, and form a medical record feature group.

[0064] A pixel and area extraction unit is configured to acquire the average pixel gray value of the background grid with a lesion point and the shape area value of the lesion point based on the medical record feature group, and form a lesion feature group.

[0065] A disease state evaluation and trend analysis module is configured to evaluate the disease state attribute value of the lesion point based on the lesion feature group, mark the disease state, and analyze the disease trend before and after drug use.

[0066] The disease state evaluation and trend analysis module comprises:

[0067] A state attribute value calculation unit is configured to calculate the disease state attribute value of the lesion point based on the lesion feature group.

[0068] A disease state marking unit is configured to compare the disease state attribute value with a preset standard value, and mark the lesion point as a repair state or an active state.

[0069] A trend analysis unit is configured to compare the disease state attribute values before and after drug use, and analyze the disease trend of the background grid with a lesion point.

[0070] a trend correlation determination module for trend labeling of the background grid, constructing a digital one-dimensional matrix, and determining the trend correlation between the eye disease medical records;

[0071] The trend correlation determination module comprises:

[0072] a trend labeling unit for adding a trend label to the background grid based on the disease trend and setting a label value;

[0073] a matrix construction unit for constructing a digital one-dimensional matrix of the eye disease medical record based on the trend label value, and the matrix element is the trend label value of the background grid;

[0074] a correlation determination unit for comparing the digital one-dimensional matrices of different medical records, counting the proportion of equal elements at the same position, and if the proportion meets a preset trend correlation index value, determining that the two medical records have trend correlation and outputting the result.

[0075] Please refer to Figure 1 In this embodiment two: an intelligent processing method for eye disease medication characteristics based on medical record big data is provided, which is applicable to the above-mentioned embodiment one, and the method comprises the following steps:

[0076] Step S1: obtaining the eye images of the patient before and after medication through the wearable eye scanner, dividing the eye images into background grids, and generating an eye disease medical record containing the eye images before and after medication;

[0077] For example, the eye images of the patient are obtained through the wearable eye scanner, and the eye images of the patient are divided into background grids according to the fixed pixel size, and G background grids are obtained.

[0078] The patient uses the wearable eye scanner to scan the eye images before and after medication at each medication time, and generates an eye disease medical record, which includes the eye images before and after medication, wherein the time interval between the use of the wearable eye scanner to scan the eye images before and after medication is a fixed period.

[0079] Step S2: marking the features of the background grid with a lesion point before and after medication, generating a lesion feature set before medication and a lesion feature set after medication to constitute a medical record feature group, and extracting the pixel gray average value and shape area value of the lesion point based on the medical record feature group to form the lesion feature group.

[0080] For example, any gthbackground grid is denoted as B g When the wearable eye scanner is used to scan the eye images each time, if there is a lesion point in the background grid B g , the background grid B gThe characteristic area before and after the medication is distinguished to generate the lesion characteristic set LC1={B g |g∈[1,G]} and the lesion characteristic set LC2={B g |g∈[1,G]} respectively, and constitute the medical record characteristic group, denoted as MC i :[LC1,LC2], wherein i is the index number of the eye disease medical record, MC i is the i-th eye disease medical record.

[0081] Based on the medical record characteristic group, the pixel gray average value of the background grid with the lesion point and the shape area value of the lesion point in the background grid are obtained respectively to constitute the lesion characteristic group, denoted as [avg,S], wherein, represents the background grid belonging to the lesion characteristic set before the medication or the lesion characteristic set after the medication, x=1 or 2, avg represents the pixel gray average value, and S represents the shape area value.

[0082] Step S3: Based on the lesion characteristic group, the disease state attribute value of the lesion point is evaluated, the disease state of the lesion point is marked by comparing with the preset standard value, and the disease trend of the background grid with the lesion point before and after the medication is analyzed.

[0083] For example, based on the lesion characteristic group, the disease state attribute value of the lesion point is evaluated The disease state attribute value is compared with the preset disease state standard value, if the disease state attribute value is greater than or equal to the disease state standard value, the disease state marked in the lesion characteristic group [avg,S] is marked as the repair state, if the disease state attribute value is less than the disease state standard value, the disease state marked in the lesion characteristic group [avg,S] is marked as the active state.

[0084] Based on the medical record characteristic group, the disease trend of the background grid with the lesion point before and after the medication is analyzed, if , it indicates that the disease trend of the background grid B g with the lesion point is approaching the active state, if , it indicates that the disease trend of the background grid B g with the lesion point is approaching the repair state, if , it indicates that the disease trend of the background grid B g with the lesion point is stable.

[0085] Step S4: The trend labeling processing is performed on the background grid according to the disease trend, the digital one-dimensional matrix corresponding to the eye disease medical record is constructed, the eye disease medical records with the trend correlation are determined and output by comparing the digital one-dimensional matrices of different medical records.

[0086] For example, based on the disease trend before and after the medication, the background grid B g The additional trend label includes approaching to active state, approaching to repair state and stable, set the trend label value, if the trend label is approaching to active state, then let the background grid B g The corresponding trend label value is-1, if the trend label is approaching to repair state, then let the background grid B g The corresponding trend label value is 1, if the trend label is stable, then let the background grid B g The corresponding trend label value is 0.

[0087] Based on the trend label value of the background grid B g , a digital one-dimensional matrix of the eye disease medical record is constructed, and the gth column matrix element of the digital one-dimensional matrix is the trend label value of the background grid B g , the digital one-dimensional matrix corresponding to the generated eye disease medical record MC i is recorded as R(MC i ), and the jth eye disease medical record MC j is obtained. j );

[0088] Compare whether the trend label values at the same matrix element positions of the digital one-dimensional matrix R(MC i ) and the digital one-dimensional matrix R(MC j ) are equal, and count the proportion of the number of equal matrix elements, if the proportion meets the set trend correlation index value, it is determined that there is a trend correlation between the eye disease medical record MC i and the eye disease medical record MC j , and the eye disease medical record with trend correlation is output.

[0089] It should be noted that in this paper, relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the term "includes", "contains" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment.

[0090] Finally, it should be noted that the above only describes the preferred embodiments of the present application and is not intended to limit the present application, and although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent replacements to some technical features. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. An intelligent processing method for eye disease medication characteristics based on medical record big data, characterized by: The method comprises the following steps: Step S1: Obtain eye images of the patient before and after medication using a wearable eye scanner, perform background grid division on the eye images, and generate an eye disease medical record including the eye images before and after medication; Step S2: Mark the background grid with lesions before and after medication to distinguish the features, generate a lesion feature set before medication and a lesion feature set after medication to form a medical record feature group, and extract the pixel grayscale average value and shape area value of the lesion points based on the medical record feature group to form a lesion feature group; Step S3: evaluating the symptom status attribute value of the lesion point based on the lesion feature group, marking the symptom status of the lesion point by comparing it with a preset standard value, and analyzing the symptom trend of the background grid where the lesion point exists before and after medication; Step S4: Trend labeling is performed on the background grid according to the symptom trend, and a digital one-dimensional matrix corresponding to the eye disease medical records is constructed. By comparing the digital one-dimensional matrices of different medical records, the eye disease medical records with trend association are determined and output.

2. The intelligent processing method for eye disease medication characteristics based on medical record big data according to claim 1 is characterized in that: The specific implementation process of step S1 includes: Acquire an eye image of the patient using a wearable eye scanner, and divide the eye image of the patient into background grids according to a fixed pixel size, obtaining a total of G background grids; Each time a patient takes medication, a wearable eye scanner is used to scan eye images before and after taking the medication, and an eye disease medical record is generated. The eye disease medical record includes the eye image before and after taking the medication. Among them, the time interval before and after the eye image scan using the wearable eye scanner before and after taking the medication is a fixed period.

3. The intelligent processing method for eye disease medication characteristics based on medical record big data according to claim 2 is characterized in that: The specific implementation process of step S2 includes: Let any g-th background grid be B g , each time the wearable eye scanner is used to scan the eye image, if the background grid B g If there is a lesion point in the background grid B, g Perform feature differentiation and labeling before and after medication to generate the feature set of lesions before medication LC1 = {B g |g∈[1,G]} and the feature set of lesions after medication LC2={B g |g∈[1, G]}, and constitute a medical record feature group, recorded as MC i : [LC1, LC2], where i is the index number of the eye disease medical record, MC i is the i-th eye disease medical record; Based on the medical record feature group, the average grayscale value of the background grid where the lesion point exists and the shape area value of the lesion point in the background grid are obtained respectively to form the lesion feature group, which is recorded as [avg, S], where Represents the background grid belonging to the lesion feature set before or after medication, x = 1 or 2, avg represents the average pixel grayscale value, and S represents the shape area value.

4. The intelligent processing method for eye disease medication characteristics based on medical record big data according to claim 3 is characterized in that: The specific implementation process of step S3 includes: Based on the lesion feature group, evaluate the disease state attribute value of the lesion point Compare the symptom state attribute value with the preset symptom state standard value. If the symptom state attribute value is greater than or equal to the symptom state standard value, the lesion point is placed in the lesion feature group. The disease state fed back in [avg, S] is marked as the repair state. If the disease state attribute value is less than the disease state standard value, the lesion point is marked in the lesion feature group. The disease state fed back in [avg, S] is marked as active state; Based on the medical record feature group, the symptom trend of the background grid with lesions before and after medication is analyzed. It means that there is a background grid B with lesions. g The symptom trend is approaching the active state. It means that there is a background grid B with lesions. g The disease trend is approaching the repair state. It means that there is a background grid B with lesions. g The trend of the disease is stable.

5. The intelligent processing method for eye disease medication characteristics based on medical record big data according to claim 4 is characterized in that: The specific implementation process of step S4 includes: Based on the symptom trend before and after medication, the background grid B g Add trend labels, including trend labels approaching to active state, approaching to repair state and stable, set trend label value, if the trend label is approaching to active state, then set background grid B g The corresponding trend label value is -1. If the trend label is approaching the repair state, then the background grid B g The corresponding trend label value is 1. If the trend label is stable, then the background grid B g The corresponding trend label value is 0; Based on background grid B g The trend label value of the eye disease medical record is constructed by a digital one-dimensional matrix, and the matrix element of the g-th column of the digital one-dimensional matrix is ​​the background grid B g The trend label value of eye disease medical records MC i The corresponding generated digital one-dimensional matrix is ​​recorded as R(MC i ), get the jth eye disease medical record MC j The corresponding generated digital one-dimensional matrix is ​​recorded as R(MC j ); Compare the digital one-dimensional matrix R(MC i ) and the digital one-dimensional matrix R(MC j ) are equal, and the proportion of the number of equal matrix elements is counted. If the proportion meets the set trend correlation index value, the eye disease medical record MC is determined. i and eye disease medical history MC j There is a trend correlation between them, and the medical records of eye diseases with trend correlation are output.

6. An intelligent processing system for eye disease medication characteristics based on medical record big data, which executes the intelligent processing method for eye disease medication characteristics according to any one of claims 1 to 5, characterized in that: The system comprises: Image acquisition and medical record generation module, used to obtain eye images of patients before and after medication, divide the background grid and generate eye disease medical records; The feature marking and extraction module is used to distinguish the features of the background grid with lesions before and after medication, generate medical record feature groups and extract lesion feature groups; Symptom status assessment and trend analysis module, which is used to assess the symptom status attribute value of the lesion point based on the lesion feature group, mark the symptom status and analyze the symptom trend before and after medication; The trend association determination module is used to perform trend labeling on the background grid, construct a digital one-dimensional matrix and determine the trend association between eye disease medical records.

7. The intelligent processing system for eye disease medication characteristics based on medical record big data according to claim 6 is characterized in that: The image acquisition and medical record generation module includes: A wearable image acquisition unit is used to obtain eye images of the patient before and after medication through a wearable eye scanner, and the time interval between scans before and after medication is a fixed period; A grid division unit, used for dividing the background grid of the eye image according to a fixed pixel size; The medical record generating unit is used to integrate the eye images before and after medication to generate an eye disease medical record containing corresponding image information.

8. The intelligent processing system for eye disease medication characteristics based on medical record big data according to claim 6 is characterized in that: The feature marking and extraction module includes: The lesion feature set marking unit is used to distinguish the features of the background grid with lesions before and after medication, generate the lesion feature set before medication and the lesion feature set after medication, and form a medical record feature group; The pixel and area extraction unit is used to obtain the pixel grayscale average value of the background grid where the lesion point exists and the shape and area value of the lesion point based on the medical record feature group to form a lesion feature group.

9. The intelligent processing system for eye disease medication characteristics based on medical record big data according to claim 6 is characterized in that: The symptom status assessment and trend analysis module includes: A state attribute value calculation unit calculates the symptom state attribute value of the lesion point based on the lesion feature group; A symptom status marking unit is used to compare the symptom status attribute value with a preset standard value and mark the lesion point as a repair state or an active state; The trend analysis unit is used to compare the symptom status attribute values ​​before and after medication, and analyze the symptom trend of the background grid where the lesion points exist.

10. The intelligent processing system for eye disease medication characteristics based on medical record big data according to claim 6 is characterized in that: The trend association determination module includes: A trend labeling unit is used to add trend labels to the background grid based on the symptom trend and set the label value; A matrix construction unit, used for constructing a digital one-dimensional matrix of eye disease medical records based on trend label values, wherein the matrix elements are trend label values ​​of the background grid; The association determination unit is used to compare the digital one-dimensional matrices of different medical records, count the proportion of equal elements in the same position, and if the proportion meets the preset trend association index value, it is determined that there is a trend association between the two medical records and output the result.

Images