Eye medical data acquisition system and method based on artificial intelligence

Abstract

The invention discloses an eye medical data collection system and method based on artificial intelligence, and relates to the technical field of eye medical data collection, and the method comprises the steps: S10, generating an eye lesion feature data set of a to-be-detected eye of a target patient; s20, acquiring a first eye lesion feature data set and a second eye lesion feature data set based on a discrete condition of a target transaction deviation angle in the target lesion feature data set; s30, analyzing and determining an auxiliary eye lesion feature data set of the to-be-detected eyes of the target patient; and S40, generating eye medical data of the to-be-tested eyes of the target patient. The abnormal lesion feature data can be visually and clearly identified through the data set, the analysis efficiency of the data set is improved to a certain extent, and the generated eye medical data is more representative and less in data size compared with eye medical data directly acquired through an eye detection instrument.

Description

Technical Field

[0001] This invention relates to the field of ocular medical data acquisition technology, specifically an ocular medical data acquisition system and method based on artificial intelligence. Background Technology

[0002] Eye medical data, through multi-dimensional quantitative indicators and dynamic tracking capabilities, has become a core driving force for precision ophthalmology, disease prevention and control, and technological innovation.

[0003] Currently, eye testing instruments are used to collect medical data on the eyes of target patients. However, the exposure time of these instruments is relatively short, making it difficult to effectively collect intermittent abnormal medical data. This hinders the accurate assessment of the patient's eye condition. Furthermore, the large volume of continuous medical data collected by these instruments requires a significant amount of time to interpret. Summary of the Invention

[0004] The purpose of this invention is to provide an artificial intelligence-based eye medical data acquisition system and method to solve the problems raised in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for acquiring ocular medical data based on artificial intelligence, the method comprising: S10: Using a camera installed inside the eye position monitoring glasses, real-time eye movement data of the target patient under different gaze behavior types is captured, and a dataset of ocular lesion features of the target patient's eye to be tested is generated. S20: Search for the target lesion feature dataset in the ocular lesion feature dataset of the target patient's eye to be tested, and obtain the first ocular lesion feature dataset and the second ocular lesion feature dataset based on the discreteness of the target abnormal deviation angle in the target lesion feature dataset. S30: Analyze and identify the auxiliary ocular lesion feature dataset of the target patient's eye; S40: Generate ocular medical data for the target patient's eye to be tested, and transmit the generated ocular medical data to the medical terminal.

[0006] Furthermore, S10 includes: S101: After the target patient wears eye position monitoring glasses, they look at the observation interface. A three-dimensional spatial coordinate system is constructed with the center of the eye socket of the target patient's eye as the origin. The positive direction of the X-axis of the three-dimensional spatial coordinate system points to the root of the nose of the target patient. The camera that acquires the eye image of the target eye is called the target camera. The eye movement data of the target patient's eye under different gaze behavior types are captured in real time through the target camera. The eye movement data includes the position coordinates of the pupil center of the target patient's eye, the gap distance between the upper and lower eyelid edges of the target patient's eye, and the position coordinates of the gaze point of the target patient's eye. The gaze behavior types include fixation and saccade. S102: At time t, when the target patient's gaze behavior type is gaze, the position coordinates of the pupil center of the target patient's eye at time t are collected; At time t, when the target patient's gaze behavior type changes, eye movement data before and after the change in the target patient's gaze behavior type are collected; S103: Compare the coordinates A of the pupil center of the target patient's eye at time t with the origin C. If coordinates A and C do not coincide, the distance fACt and the deviation angle hACt of the pupil at time t are taken as the first ocular lesion feature of the target patient's eye at time t. If coordinates A and C coincide, fAct=0 and hAct=0. The method for calculating the deviation angle hACt is as follows: Draw a ray parallel to the observation interface through the origin C, with the ray pointing towards the root of the nose of the target patient. Randomly select a point D on the ray and denote the coordinates of the selected point D as (x1,0,0). Let the real-time position coordinates A of the center of the pupil of the target patient's eye be (x3,y3,z3), and the position coordinates C of the origin be (0,0,0). Calculate the angle between vector CA´ and vector CD using the formula for calculating the angle between three-dimensional spatial vectors to obtain the deviation angle hAct of the pupil of the target patient's eye at time t, where x1>0; S104: Compare the coordinates B of the eye position of the target patient before the change in the target patient's gaze behavior type with the coordinates G of the eye position of the target patient after the change in the target patient's gaze behavior type. If coordinates B and G do not coincide, the deviation angle hBGt of the eye position of the target patient at time t is taken as the second ocular lesion feature of the eye at time t. If coordinates B and G coincide, then hBGt=0. The method for calculating the deviation angle hBGt is as follows: Let the position coordinates B be (x4, y4, z4), the position coordinates G be (x5, y5, z5), and the position coordinates of the origin C be (0, 0, 0). The angle between vectors CB and CG is calculated using the formula for calculating the angle between vectors in three-dimensional space. This gives the deviation angle hBGt of the target patient's line of sight at time t. If the deviation angle hBGt is close to the root of the target patient's nose, a negative sign is added before the value hBGt to make it negative. If the deviation angle hBGt is far from the root of the target patient's nose, no processing is done on the value hBGt. S105: Generate an ocular lesion feature dataset M for the target patient's eye based on the first and second ocular lesion features, M={St,St+d,…,St+(n-1)*d}, St=(fAct,hAct,hBGt), where d represents the acquisition interval of eye movement data, i=1,2,…,n, indicating that the obtained ocular lesion features are numbered according to the chronological order, n represents the total number of numbers, and St+(n-1)*d represents the ocular lesion features obtained by the target patient's eye at time t+(n-1)*d.

[0007] Furthermore, S20 includes: S201: Search for the target lesion feature data in the ocular lesion feature dataset M. The target lesion feature data format is St+(i-1)*d=(0,0,0), where St+(i-1)*d represents the ocular lesion feature obtained by the target patient's eye at time t+(i-1)*d. S202: Remove the target lesion feature data from the ocular lesion feature dataset M to obtain the target ocular lesion feature dataset M1. In the target ocular lesion feature dataset M1, search for the target abnormal deviation angle. The specific search method is as follows: Determine whether 135°<hAc[t+(i-1)*d]<215° holds true. If it does, then hAc[t+(i-1)*d] is considered to be the target deviation angle. If it does not, then hAc[t+(i-1)*d] is not considered to be the target deviation angle. Determine whether hBG[t+(i-1)*d]>0° holds true. If it does, then hBG[t+(i-1)*d] is considered to be the target deviation angle. If it does not, then hBG[t+(i-1)*d] is not considered to be the target deviation angle. S203: Determine the total number of target deviation angles Pt+(i-1)*d and the total number of St+(i-1)*d stored in the target eye lesion feature dataset M1, and calculate the sum Nt+(i-1)*d between Pt+(i-1)*d / (2*Qt+(i-1)*d) and the value 1; If 1≤Nt+(i-1)*d≤1+z or 2-z≤Nt+(i-1)*d≤2, then sum all hAc[t+(i-1)*d] from i=1 to i=n to obtain HAc, and store the HAc / Qt+(i-1)*d value in set N1 to obtain the first ocular lesion feature dataset; If 1+z<Nt+(i-1)*d<2-z, then the target deviation angle is stored in set N2 to obtain the second ocular lesion feature dataset.

[0008] Furthermore, S30 includes: The range of distance values ​​[T1, T2] between the upper and lower eyelid margins of the target patient's eye was obtained using a clinical experience database. If Kt+(i-1)*d < T1, then the auxiliary ocular lesion characteristic coefficient of the target patient's eye at time t+(i-1)*d is calculated according to Ut+(i-1)*d=(T1-Kt+(i-1)*d) / T1. Here, Kt+(i-1)*d represents the average gap distance between the upper and lower eyelid edges of the target patient's eye at time t+(i-1)*d. Kt+(i-1)*d=(1 / n)*Σkt+(i-1)*d, where the superscript of Σ is n and the subscript is i=1. kt+(i-1)*d represents the gap distance between the upper and lower eyelid edges of the target patient's eye at time t+(i-1)*d. If T1≤Kt+(i-1)*d≤T2, then the auxiliary ocular lesion characteristic coefficient Ut+(i-1)*d of the target patient's eye at time t+(i-1)*d is considered to be 0. The calculated real-time auxiliary ocular lesion feature coefficients of the target patient's eye to be tested are stored in set V to obtain the auxiliary ocular lesion feature dataset of the target patient's eye to be tested, V={Ut,Ut+d,…,Ut+(i-1)*d}.

[0009] Furthermore, S40 includes: The system generates ocular medical data for the target patient's eye, including a first ocular lesion feature dataset, a second ocular lesion feature dataset, and an auxiliary ocular lesion feature dataset. The generated ocular medical data is then transmitted to a medical terminal.

[0010] An artificial intelligence-based ophthalmic medical data acquisition system includes an ophthalmic lesion feature dataset generation module, an ophthalmic lesion data classification module, an auxiliary ophthalmic lesion feature dataset acquisition module, and an ophthalmic medical data generation module. The ocular lesion feature dataset generation module is used to capture eye movement data of the target patient under different gaze behavior types in real time and generate an ocular lesion feature dataset of the target patient's eye to be tested. The ocular lesion data classification module is used to search for the target lesion feature dataset in the ocular lesion feature dataset of the target patient's eye to be tested, and to obtain the first ocular lesion feature dataset and the second ocular lesion feature dataset based on the discreteness of the target abnormal deviation angle in the target lesion feature dataset. The auxiliary ocular lesion feature dataset acquisition module is used to analyze and determine the auxiliary ocular lesion feature dataset of the target patient's eye to be tested; The ophthalmic medical data generation module is used to generate ophthalmic medical data for the target patient's eye to be tested, and transmit the generated ophthalmic medical data to the medical terminal.

[0011] Furthermore, the ocular lesion feature dataset generation module includes an eye movement data capture unit, a first ocular lesion feature determination unit, a second ocular lesion feature determination unit, and an ocular lesion feature dataset generation unit; The eye-tracking data capture unit captures eye-tracking data of the target patient's eyes under different types of fixation behavior in real time using a target camera; The first ocular lesion feature determination unit determines the first ocular lesion feature of the target patient's eye based on the overlap between the real-time position coordinates of the pupil center of the target patient's eye and the coordinate origin. The second ocular lesion feature determination unit determines the second ocular lesion feature of the target patient's eye based on the overlap between the coordinates of the gaze point of the target patient's eye before the change in the target patient's gaze behavior type and the coordinates of the gaze point of the target patient's eye after the change in the target patient's gaze behavior type. The ocular lesion feature dataset generation unit generates an ocular lesion feature dataset of the target patient's eye to be tested based on the first ocular lesion feature and the second ocular lesion feature.

[0012] Furthermore, the ocular lesion data classification module includes a target ocular lesion feature dataset acquisition unit, a target abnormal deviation angle finding unit, and an ocular lesion feature dataset classification unit; The target ocular lesion feature dataset acquisition unit removes the target lesion feature data from the ocular lesion feature dataset to obtain the target ocular lesion feature dataset; The target deviation angle finding unit searches for the target deviation angle in the target ocular lesion feature dataset; The ocular lesion feature dataset classification unit obtains the first ocular lesion feature dataset and the second ocular lesion feature dataset based on the discreteness of the target deviation angle in the target lesion feature dataset.

[0013] Furthermore, the auxiliary ocular lesion feature dataset acquisition module includes a judgment unit, an auxiliary ocular lesion feature coefficient calculation unit, and an auxiliary ocular lesion feature dataset acquisition unit; The judgment unit determines whether the real-time average gap distance between the upper and lower eyelid edges of the target patient's eye is within the range of the gap distance between the upper and lower eyelid edges of the target patient's eye obtained through the clinical experience database, and transmits the judgment result to the auxiliary ocular lesion characteristic coefficient calculation unit. The auxiliary ocular lesion feature coefficient calculation unit calculates the real-time auxiliary ocular lesion feature coefficient of the target patient's eye based on the judgment result transmitted by the judgment unit. The auxiliary ocular lesion feature dataset acquisition unit stores the calculated real-time auxiliary ocular lesion feature coefficients of the target patient's eye to be tested into a set, thereby obtaining the auxiliary ocular lesion feature dataset of the target patient's eye to be tested.

[0014] Furthermore, the ocular medical data generation module includes an ocular medical data generation unit and a transmission unit; The ophthalmic medical data generation unit generates ophthalmic medical data for the target patient's eye based on the first ophthalmic lesion feature dataset, the second ophthalmic lesion feature dataset, and the auxiliary ophthalmic lesion feature dataset. The transmission unit transmits the generated ocular medical data to the medical terminal.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention uses eye position monitoring glasses to continuously capture eye movement data of the target patient's eye under different fixation behavior types in real time. Based on the captured eye movement data, the real-time movement distance and real-time deviation angle of the pupil of the target patient's eye, as well as the real-time deviation angle of the eye's line of sight, are calculated to obtain a dataset of ocular lesion features of the target patient's eye. The dataset allows for intuitive and clear identification of abnormal lesion features, thereby improving the efficiency of dataset analysis to a certain extent.

[0016] 2. Based on the discreteness of the target abnormal deviation angle in the target lesion feature dataset, this invention analyzes and classifies the abnormal deviation angle in the target lesion feature dataset, thereby enabling continuous screening of the lesion feature data of the target patient's eye to be tested. Moreover, the lesion feature data retained after screening is more representative, further improving the system's accuracy in collecting ophthalmic medical data.

[0017] 3. This invention transmits the generated ocular medical data to a medical terminal, enabling medical personnel to quickly understand the lesions of the target patient's eye. Furthermore, the ocular medical data generated by this invention differs from ocular medical data directly collected by ocular testing instruments, and is more representative and contains less data. Attached Figure Description

[0018] Figure 1 This is a schematic diagram illustrating the workflow of an artificial intelligence-based ocular medical data acquisition method according to the present invention. Detailed Implementation

[0019] Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] like Figure 1 As shown, this invention provides a technical solution for an artificial intelligence-based ocular medical data acquisition system and method, and an artificial intelligence-based ocular medical data acquisition method, characterized in that: the method includes: S10: Using a camera installed inside the eye position monitoring glasses, real-time eye movement data of the target patient under different gaze behavior types is captured, and a dataset of ocular lesion features of the target patient's eye to be tested is generated. S10 includes: S101: After wearing eye position monitoring glasses, the target patient looks at the observation interface. A three-dimensional spatial coordinate system is constructed with the center of the target patient's eye socket as the origin. The positive X-axis of the three-dimensional spatial coordinate system points towards the root of the target patient's nose, and the positive Y-axis is perpendicular to the observation interface. The camera that acquires the eye image of the target eye is designated as the target camera. The target camera captures eye movement data of the target patient's eye under different gaze behavior types in real time. The eye movement data includes the position coordinates of the pupil center of the target patient's eye, the gap distance between the edges of the upper and lower eyelids of the target patient's eye, and the position coordinates of the gaze point of the target patient's eye when looking at the observation interface. The location of the eye being tested, and the position of the gaze point, are obtained by analyzing the eye images acquired by the target camera. The coordinates of the center of the pupil of the target patient's eye are calculated using a built-in algorithm to obtain the position of the current gaze point of the target patient's eye on the observation interface. This method belongs to the prior art. The types of gaze behavior include fixation and saccade. Fixation is the process of keeping the foveal vision in one position to obtain sufficient visual information, while saccade is the eye movement behavior of turning from one gaze to another. S102: At time t, when the target patient's gaze behavior type is gaze, the position coordinates of the pupil center of the target patient's eye at time t are collected; At time t, when the target patient's fixation behavior type changes, eye movement data before and after the change in the target patient's fixation behavior type are collected. The change in fixation behavior type refers to the change from fixation to saccade or from saccade to fixation. S103: Compare the coordinates A of the center of the pupil of the target patient's eye at time t with the origin C. If coordinates A and C do not coincide, the movement distance fACt and the deviation angle hACt of the pupil of the target patient's eye at time t are taken as the first ocular lesion feature of the target patient's eye at time t. The movement distance is calculated using the distance formula between two points in space, and the deviation angle is calculated using the angle formula between three-dimensional spatial vectors. Both are existing technologies. If coordinates A and C coincide, then fAct=0 and hAct=0. The method for calculating the deviation angle hACt is as follows: Draw a ray parallel to the observation interface through the origin C, with the ray pointing towards the root of the nose of the target patient. Randomly select a point D on the ray and denote the coordinates of the selected point D as (x1,0,0). Let the real-time position coordinates A of the center of the pupil of the target patient's eye be (x3,y3,z3), and the position coordinates C of the origin be (0,0,0). Calculate the angle between vector CA´ and vector CD using the formula for calculating the angle between three-dimensional spatial vectors to obtain the deviation angle hAct of the pupil of the target patient's eye at time t, where x1>0; S104: Compare the coordinates B of the eye position of the target patient before the change in the target patient's gaze behavior type with the coordinates G of the eye position of the target patient after the change in the target patient's gaze behavior type. If coordinates B and G do not coincide, the deviation angle hBGt of the eye position of the target patient at time t is taken as the second ocular lesion feature of the eye at time t. If coordinates B and G coincide, then hBGt=0. The method for calculating the deviation angle hBGt is as follows: Let the position coordinates B be (x4, y4, z4), the position coordinates G be (x5, y5, z5), and the position coordinates of the origin C be (0, 0, 0). The angle between vectors CB and CG is calculated using the formula for calculating the angle between vectors in three-dimensional space. This gives the deviation angle hBGt of the target patient's line of sight at time t. If the deviation angle hBGt is close to the root of the target patient's nose, a negative sign is added before the value hBGt to make it negative. If the deviation angle hBGt is far from the root of the target patient's nose, no processing is done on the value hBGt. S105: Generate an ocular lesion feature dataset M for the target patient's eye based on the first and second ocular lesion features, M={St,St+d,…,St+(n-1)*d}, St=(fAct,hAct,hBGt), where d represents the acquisition interval of eye movement data, i=1,2,…,n, indicating that the obtained ocular lesion features are numbered according to the chronological order, n represents the total number of numbers, and St+(n-1)*d represents the ocular lesion features obtained by the target patient's eye at time t+(n-1)*d; S20: Search for the target lesion feature dataset in the ocular lesion feature dataset of the target patient's eye to be tested, and obtain the first ocular lesion feature dataset and the second ocular lesion feature dataset based on the discreteness of the target abnormal deviation angle in the target lesion feature dataset. S20 includes: S201: Search for the target lesion feature data in the ocular lesion feature dataset M. The target lesion feature data format is St+(i-1)*d=(0,0,0), where St+(i-1)*d represents the ocular lesion feature obtained by the target patient's eye at time t+(i-1)*d. S202: Remove the target lesion feature data from the ocular lesion feature dataset M to obtain the target ocular lesion feature dataset M1. In the target ocular lesion feature dataset M1, search for the target abnormal deviation angle. The specific search method is as follows: Determine whether 135°<hAc[t+(i-1)*d]<215° holds true. If it does, then hAc[t+(i-1)*d] is considered to be the target deviation angle. If it does not, then hAc[t+(i-1)*d] is not considered to be the target deviation angle. Determine whether hBG[t+(i-1)*d]>0° holds true. If it does, then hBG[t+(i-1)*d] is considered to be the target deviation angle. If it does not, then hBG[t+(i-1)*d] is not considered to be the target deviation angle. S203: Determine the total number of target deviation angles Pt+(i-1)*d and the total number of St+(i-1)*d stored in the target eye lesion feature dataset M1, and calculate the sum Nt+(i-1)*d between Pt+(i-1)*d / (2*Qt+(i-1)*d) and the value 1; If 1≤Nt+(i-1)*d≤1+z or 2-z≤Nt+(i-1)*d≤2, then sum all hAc[t+(i-1)*d] from i=1 to i=n to obtain HAc, and store the HAc / Qt+(i-1)*d value in set N1 to obtain the first ocular lesion feature dataset; If 1+z<Nt+(i-1)*d<2-z, then the target deviation angle is stored in set N2 to obtain the second ocular lesion feature dataset; S30: Analyze and identify the auxiliary ocular lesion feature dataset of the target patient's eye; S30 includes: The range of distance values ​​[T1,T2] between the upper and lower eyelid margins of the target patient was obtained by collecting and organizing a large amount of clinical experience from the clinical experience database. If Kt+(i-1)*d < T1, then the auxiliary ocular lesion characteristic coefficient of the target patient's eye at time t+(i-1)*d is calculated according to Ut+(i-1)*d=(T1-Kt+(i-1)*d) / T1. Here, Kt+(i-1)*d represents the average gap distance between the upper and lower eyelid edges of the target patient's eye at time t+(i-1)*d. Kt+(i-1)*d=(1 / n)*Σkt+(i-1)*d, where the superscript of Σ is n and the subscript is i=1. kt+(i-1)*d represents the gap distance between the upper and lower eyelid edges of the target patient's eye at time t+(i-1)*d. If T1≤Kt+(i-1)*d≤T2, then the auxiliary ocular lesion characteristic coefficient Ut+(i-1)*d of the target patient's eye at time t+(i-1)*d is considered to be 0. The calculated real-time auxiliary ocular lesion feature coefficients of the target patient's eye to be tested are stored in set V to obtain the auxiliary ocular lesion feature dataset of the target patient's eye to be tested, V={Ut,Ut+d,…,Ut+(i-1)*d}; S40: Generate ocular medical data for the target patient's eye to be tested, and transmit the generated ocular medical data to the medical terminal; S40 includes: The system generates ocular medical data for the target patient's eye, including a first ocular lesion feature dataset, a second ocular lesion feature dataset, and an auxiliary ocular lesion feature dataset. The generated ocular medical data is then transmitted to a medical terminal, which refers to the computer of the medical staff.

[0021] An artificial intelligence-based ophthalmic medical data acquisition system includes an ophthalmic lesion feature dataset generation module, an ophthalmic lesion data classification module, an auxiliary ophthalmic lesion feature dataset acquisition module, and an ophthalmic medical data generation module. The ocular lesion feature dataset generation module is used to capture eye movement data of target patients under different gaze behavior types in real time and generate an ocular lesion feature dataset of the target patient's eye to be tested. The ocular lesion feature dataset generation module includes an eye movement data capture unit, a first ocular lesion feature determination unit, a second ocular lesion feature determination unit, and an ocular lesion feature dataset generation unit; The eye-tracking data capture unit captures eye-tracking data of the target patient's eyes in real time under different types of fixation behavior using a target camera; The first ocular lesion feature determination unit determines the first ocular lesion features of the target patient's eye based on the overlap between the real-time position coordinates of the pupil center of the target patient's eye and the coordinate origin. The second ocular lesion feature determination unit determines the second ocular lesion features of the target patient's eye based on the overlap between the coordinates of the gaze point of the target patient's eye before the change in the target patient's gaze behavior type and the coordinates of the gaze point of the target patient's eye after the change in the target patient's gaze behavior type. The ocular lesion feature dataset generation unit generates an ocular lesion feature dataset for the target patient's eye based on the first ocular lesion feature and the second ocular lesion feature; The ocular lesion data classification module is used to search for the target lesion feature dataset in the ocular lesion feature dataset of the target patient's eye to be tested, and to obtain the first ocular lesion feature dataset and the second ocular lesion feature dataset based on the discreteness of the target abnormal deviation angle in the target lesion feature dataset. The ocular lesion data classification module includes a target ocular lesion feature dataset acquisition unit, a target anomaly deviation angle detection unit, and an ocular lesion feature dataset classification unit; The target ocular lesion feature dataset acquisition unit removes the target ocular lesion feature data from the ocular lesion feature dataset to obtain the target ocular lesion feature dataset; The target deviation angle finding unit searches for the target deviation angle in the target ocular lesion feature dataset; The ocular lesion feature dataset classification unit obtains the first ocular lesion feature dataset and the second ocular lesion feature dataset based on the discreteness of the target aberration deviation angle in the target lesion feature dataset; The auxiliary ocular lesion feature dataset acquisition module is used to analyze and determine the auxiliary ocular lesion feature dataset of the target patient's eye to be tested; The auxiliary ocular lesion feature dataset acquisition module includes a judgment unit, an auxiliary ocular lesion feature coefficient calculation unit, and an auxiliary ocular lesion feature dataset acquisition unit; The system includes a judgment unit to determine whether the real-time average gap distance between the upper and lower eyelid edges of the target patient's eye is within the range obtained from the clinical experience database, and transmits the judgment result to the auxiliary ocular lesion characteristic coefficient calculation unit. The auxiliary ocular lesion characteristic coefficient calculation unit calculates the real-time auxiliary ocular lesion characteristic coefficient of the target patient's eye based on the judgment result transmitted by the judgment unit; The auxiliary ocular lesion feature dataset acquisition unit stores the calculated real-time auxiliary ocular lesion feature coefficients of the target patient's eye to be tested into a set, thereby obtaining the auxiliary ocular lesion feature dataset of the target patient's eye to be tested. The ophthalmic medical data generation module is used to generate ophthalmic medical data for the target patient's eye to be tested, and transmit the generated ophthalmic medical data to the medical terminal; The ophthalmic medical data generation module includes an ophthalmic medical data generation unit and a transmission unit; The ophthalmic medical data generation unit generates ophthalmic medical data for the target patient's eye based on the first ophthalmic lesion feature dataset, the second ophthalmic lesion feature dataset, and the auxiliary ophthalmic lesion feature dataset. The transmission unit transmits the generated ocular medical data to the medical terminal.

[0022] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention 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 substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for acquiring ocular medical data based on artificial intelligence, characterized in that: The method includes: S10: Using a camera installed inside the eye position monitoring glasses, real-time eye movement data of the target patient under different gaze behavior types is captured, and a dataset of ocular lesion features of the target patient's eye to be tested is generated. S20: Search for the target lesion feature dataset in the ocular lesion feature dataset of the target patient's eye to be tested, and obtain the first ocular lesion feature dataset and the second ocular lesion feature dataset based on the discreteness of the target abnormal deviation angle in the target lesion feature dataset. S30: Analyze and identify the auxiliary ocular lesion feature dataset of the target patient's eye; S40: Generate ocular medical data for the target patient's eye to be tested, and transmit the generated ocular medical data to the medical terminal.

2. The method for acquiring ocular medical data based on artificial intelligence according to claim 1, characterized in that: S10 includes: S101: After the target patient wears eye position monitoring glasses, they look at the observation interface. A three-dimensional spatial coordinate system is constructed with the center of the eye socket of the target patient's eye as the origin. The positive direction of the X-axis of the three-dimensional spatial coordinate system points to the root of the nose of the target patient. The camera that acquires the eye image of the target eye is called the target camera. The eye movement data of the target patient's eye under different gaze behavior types are captured in real time through the target camera. The eye movement data includes the position coordinates of the pupil center of the target patient's eye, the gap distance between the upper and lower eyelid edges of the target patient's eye, and the position coordinates of the gaze point of the target patient's eye. The gaze behavior types include fixation and saccade. S102: At time t, when the target patient's gaze behavior type is gaze, the position coordinates of the pupil center of the target patient's eye at time t are collected; At time t, when the target patient's gaze behavior type changes, eye movement data before and after the change in the target patient's gaze behavior type are collected; S103: Compare the coordinates A of the center of the pupil of the target patient's eye at time t with the origin C. If coordinates A and C do not coincide, then measure the distance f that the pupil of the target patient's eye has moved at time t. ACt and the deviation angle h ACt As the first ocular lesion feature of the target patient's eye at time t, if coordinates A and C coincide, then f Act =0, h Act =0; Angle of deviation h ACt The calculation method is as follows: Draw a ray parallel to the observation interface through the origin C, with the ray pointing towards the root of the nose of the target patient. Randomly select a point D on the ray, and denote the coordinates of the selected point D as (x1,0,0). Let the real-time position coordinates A of the center of the pupil of the target patient's eye be (x3,y3,z3), and the position coordinates C of the origin be (0,0,0). Calculate the angle between vectors CA´ and CD using the formula for calculating the angle between three-dimensional spatial vectors, and obtain the deviation angle h of the pupil of the target patient's eye at time t. Act ; S104: Compare the coordinates B of the target patient's eye's gaze point position before the change in the target patient's gaze behavior type with the coordinates G of the target patient's eye's gaze point position after the change in the target patient's gaze behavior type. If coordinates B and G do not coincide, then calculate the deviation angle h of the target patient's eye's gaze point position at time t. BGt As the second ocular lesion feature of the target patient's eye at time t, if coordinates B and G coincide, then h BGt =0; Angle of deviation h BGt The calculation method is as follows: Let the position coordinates B be (x4, y4, z4), the position coordinates G be (x5, y5, z5), and the position coordinates of the origin C be (0, 0, 0). The angle between vectors CB and CG is calculated using the formula for calculating the angle between vectors in three-dimensional space, thus obtaining the deviation angle h of the target patient's line of sight at time t. BGt If the deviation angle h BGt If the target patient's nasal root is close, then the value h BGt Add a negative sign to make it negative; if the deviation angle h is abnormal. BGt If the target patient's nasal root is far away, the numerical value h will not be affected. BGt Process it; S105: Generate an ocular lesion feature dataset M for the target patient's eye based on the first and second ocular lesion features, M={S t ,S t+d ,…,S t+(n-1)*d }, S t =(f Act ,h Act h BGt ), where d represents the acquisition interval of eye-tracking data, i=1,2,…,n, represents the numbering of the obtained ocular lesion features according to the chronological order, n represents the total number of numbers, S t+(n-1)*d This represents the ocular lesion characteristics of the target patient's eye at time t+(n-1)*d.

3. The method for acquiring ocular medical data based on artificial intelligence according to claim 2, characterized in that: S20 includes: S201: Search for target lesion feature data in the ocular lesion feature dataset M. The target lesion feature data format is S. t+(i-1)*d =(0,0,0), S t+(i-1)*d This represents the ocular lesion characteristics obtained in the target patient's eye at time t+(i-1)*d; S202: Remove the target lesion feature data from the ocular lesion feature dataset M to obtain the target ocular lesion feature dataset M1. In the target ocular lesion feature dataset M1, search for the target abnormal deviation angle. The specific search method is as follows: Determine if 135° < h Ac[t+(i-1)*d] Does <215° hold true? If so, then we consider h to be true. Ac[t+(i-1)*d] Let h be the target deviation angle. If this condition is not met, then h is considered to be... Ac[t+(i-1)*d] It is not the target deviation angle; Determine h BG[t+(i-1)*d] If the value > 0° is true, then h is considered true. BG[t+(i-1)*d] Let h be the target deviation angle. If this condition is not met, then h is considered to be... BG[t+(i-1)*d] It is not the target deviation angle; S203: The total number P of the target deviation angles being searched t+(i-1)*d and the S stored in the target ocular lesion feature dataset M1 t+(i-1)*d Total number Q t+(i-1)*d To determine, for P t+(i-1)*d / (2*Q t+(i-1)*d The sum N between 1 and 1 t+(i-1)*d Perform calculations; If 1≤N t+(i-1)*d ≤1+z or 2-z≤N t+(i-1)*d If ≤2, then for all h from i=1 to i=n Ac[t+(i-1)*d] Summation is performed to obtain H Ac and H Ac / Q t+(i-1)*d The values ​​are stored in set N1 to obtain the first ocular lesion feature dataset; If 1+z<N t+(i-1)*d If <2-z, then the target deviation angle is stored in set N2 to obtain the second ocular lesion feature dataset.

4. The method for acquiring ocular medical data based on artificial intelligence according to claim 3, characterized in that: S30 includes: The range of distance values ​​[T1, T2] between the upper and lower eyelid margins of the target patient's eye was obtained using a clinical experience database. If K t+(i-1)*d <T1, then according to U t+(i-1)*d =(T1-K t+(i-1)*d The auxiliary ocular lesion characteristic coefficients of the target patient's eye at time t+(i-1)*d are calculated using K / T1, where K t+(i-1)*d K represents the average gap distance between the upper and lower eyelid margins of the target patient's eye at time t+(i-1)*d. t+(i-1)*d =(1 / n)*Σk t+(i-1)*d The superscript of Σ is n, and the subscript is i=1, k t+(i-1)*d This represents the distance between the edges of the upper and lower eyelids of the target patient at time t+(i-1)*d. If T1≤K t+(i-1)*d If ≤T2, then the auxiliary ocular lesion characteristic coefficient U of the target patient's eye at time t+(i-1)*d is considered to be... t+(i-1)*d =0; The calculated real-time auxiliary ocular lesion feature coefficients of the target patient's eye are stored in set V, resulting in the auxiliary ocular lesion feature dataset of the target patient's eye, V={U t U t+d ,…,U t+(i-1)*d } 5. The method for acquiring ocular medical data based on artificial intelligence according to claim 4, characterized in that: S40 includes: The system generates ocular medical data for the target patient's eye, including a first ocular lesion feature dataset, a second ocular lesion feature dataset, and an auxiliary ocular lesion feature dataset. The generated ocular medical data is then transmitted to a medical terminal.

6. An AI-based ocular medical data acquisition system applied to the AI-based ocular medical data acquisition method according to any one of claims 1-5, characterized in that: The system includes an ocular lesion feature dataset generation module, an ocular lesion data classification module, an auxiliary ocular lesion feature dataset acquisition module, and an ocular medical data generation module; The ocular lesion feature dataset generation module is used to capture eye movement data of the target patient under different gaze behavior types in real time and generate an ocular lesion feature dataset of the target patient's eye to be tested. The ocular lesion data classification module is used to search for the target lesion feature dataset in the ocular lesion feature dataset of the target patient's eye to be tested, and to obtain the first ocular lesion feature dataset and the second ocular lesion feature dataset based on the discreteness of the target abnormal deviation angle in the target lesion feature dataset. The auxiliary ocular lesion feature dataset acquisition module is used to analyze and determine the auxiliary ocular lesion feature dataset of the target patient's eye to be tested; The ophthalmic medical data generation module is used to generate ophthalmic medical data for the target patient's eye to be tested, and transmit the generated ophthalmic medical data to the medical terminal.

7. The artificial intelligence-based ocular medical data acquisition system according to claim 6, characterized in that: The ocular lesion feature dataset generation module includes an eye movement data capture unit, a first ocular lesion feature determination unit, a second ocular lesion feature determination unit, and an ocular lesion feature dataset generation unit. The eye-tracking data capture unit captures eye-tracking data of the target patient's eyes under different types of fixation behavior in real time using a target camera; The first ocular lesion feature determination unit determines the first ocular lesion feature of the target patient's eye based on the overlap between the real-time position coordinates of the pupil center of the target patient's eye and the coordinate origin. The second ocular lesion feature determination unit determines the second ocular lesion feature of the target patient's eye based on the overlap between the coordinates of the gaze point of the target patient's eye before the change in the target patient's gaze behavior type and the coordinates of the gaze point of the target patient's eye after the change in the target patient's gaze behavior type. The ocular lesion feature dataset generation unit generates an ocular lesion feature dataset of the target patient's eye to be tested based on the first ocular lesion feature and the second ocular lesion feature.

8. The artificial intelligence-based ocular medical data acquisition system according to claim 7, characterized in that: The ocular lesion data classification module includes a target ocular lesion feature dataset acquisition unit, a target abnormal deviation angle search unit, and an ocular lesion feature dataset classification unit; The target ocular lesion feature dataset acquisition unit removes the target lesion feature data from the ocular lesion feature dataset to obtain the target ocular lesion feature dataset; The target deviation angle finding unit searches for the target deviation angle in the target ocular lesion feature dataset; The ocular lesion feature dataset classification unit obtains the first ocular lesion feature dataset and the second ocular lesion feature dataset based on the discreteness of the target deviation angle in the target lesion feature dataset.

9. The artificial intelligence-based ocular medical data acquisition system according to claim 8, characterized in that: The auxiliary ocular lesion feature dataset acquisition module includes a judgment unit, an auxiliary ocular lesion feature coefficient calculation unit, and an auxiliary ocular lesion feature dataset acquisition unit; The judgment unit determines whether the real-time average gap distance between the upper and lower eyelid edges of the target patient's eye is within the range of the gap distance between the upper and lower eyelid edges of the target patient's eye obtained through the clinical experience database, and transmits the judgment result to the auxiliary ocular lesion characteristic coefficient calculation unit. The auxiliary ocular lesion feature coefficient calculation unit calculates the real-time auxiliary ocular lesion feature coefficient of the target patient's eye based on the judgment result transmitted by the judgment unit. The auxiliary ocular lesion feature dataset acquisition unit stores the calculated real-time auxiliary ocular lesion feature coefficients of the target patient's eye to be tested into a set, thereby obtaining the auxiliary ocular lesion feature dataset of the target patient's eye to be tested.

10. The artificial intelligence-based ocular medical data acquisition system according to claim 9, characterized in that: The ophthalmic medical data generation module includes an ophthalmic medical data generation unit and a transmission unit; The ophthalmic medical data generation unit generates ophthalmic medical data for the target patient's eye based on the first ophthalmic lesion feature dataset, the second ophthalmic lesion feature dataset, and the auxiliary ophthalmic lesion feature dataset. The transmission unit transmits the generated ocular medical data to the medical terminal.

Images