A three-dimensional model construction method based on ophthalmic detection

Abstract

The application relates to the field of eye detection, and discloses a three-dimensional model construction method based on ophthalmic detection. First, human eye state information is collected, and the human eye state information is pretreated to obtain initial eye state information; data analysis is performed on the initial eye state information to establish initial parameters of an eye three-dimensional model; parameter optimization is performed to obtain screening parameters of the eye three-dimensional model; the eye three-dimensional model is established, eye key points in the eye three-dimensional model are collected, and eye screening point data are bound for the eye key points; the eye key points are normalized, screening operation is performed through big data comparison, and final eye state information is obtained; geometric structure reconstruction is performed to obtain a reconstructed eye three-dimensional model; the application solves the problems that the prior art eye detection is inaccurate, is easy to affect a diagnosis result, and the existing eye model has a large difference in size, position and shape from actual patient eyes.

Description

TECHNICAL FIELD

[0001] The application relates to the field of eye detection, in particular to a three-dimensional model construction method based on ophthalmic detection. BACKGROUND

[0002] With the development of science and technology, the medical level is gradually improved, and for ophthalmology, the medical technology is gradually more developed, but since the eye is a relatively fragile body organ of the human body, great care is needed in the process of ophthalmic detection and eye treatment.

[0003] The current ophthalmic detection technology basically adopts the laser mode to perform eye monitoring and scanning, but in the existing eye monitoring technology, the detection effect is not good, since the eye is not a plane, the image scanned by the laser is very poor due to the coincidence, which easily affects the detection result and the diagnosis result, so a three-dimensional model construction method based on ophthalmic detection is very necessary. SUMMARY

[0004] The application aims to provide a three-dimensional model construction method based on ophthalmic detection, to solve the problems of inaccurate eye detection, easy influence on diagnosis result, and large difference between the existing eye model and the actual patient's eye size, position and shape.

[0005] In order to achieve the above-mentioned purpose, the application provides the following method:

[0006] The three-dimensional model construction method based on ophthalmic detection provided by the application is:

[0007] S1: collecting human eye state information by an eye detection device, and preprocessing the human eye state information to obtain preliminary eye state information;

[0008] S2: performing data analysis on the preliminary eye state information according to historical ophthalmic detection data, and determining initial parameters of an eye three-dimensional model;

[0009] S3: performing parameter optimization on the initial parameters of the eye three-dimensional model to obtain screening parameters of the eye three-dimensional model;

[0010] S4: establishing an eye three-dimensional model according to the screening parameters of the eye three-dimensional model, collecting eye key points in the eye three-dimensional model, and binding eye screening point data to the eye key points;

[0011] S5: performing normalization processing on the eye key points, performing screening operation through big data comparison, and obtaining final eye state information;

[0012] S6: Using the final eye state information after filtering, perform geometric correction processing to reconstruct the geometric structure and obtain a reconstructed three-dimensional eye model.

[0013] Preferably, the collection of human eye status information through eye detection equipment includes: detecting basic human eye information, as well as measuring human eye refractive power, visual field, ophthalmoscope, slit lamp, and intraocular pressure using ophthalmic ultrasound diagnostic instrument, OTC, ophthalmic AI screening software, and ophthalmic surgical microscope.

[0014] Preferably, the step of preprocessing the human eye state information to obtain preliminary eye state information includes: filtering invalid data from the collected human eye state information and calculating the average data value of the human eye state information; if the human eye state information is higher or lower than 70% of the average data value of the human eye state information, then the human eye state information is filtered out to obtain valid human eye state information; converting the valid human eye state information into a unified format and storing the valid human eye state information in a unified database to obtain preliminary eye state information.

[0015] Preferably, the step of analyzing the preliminary eye condition information based on historical ophthalmological examination data to establish the initial parameters of the three-dimensional eye model includes: summarizing and analyzing the basic information of the human eye and the historical ophthalmological examination data, dividing the eye into three parts from the inside out: the core body, the outer eye body, and the inner eye body; dividing the preliminary eye condition information into the three parts of the eye body and calculating the average initial parameters of each part of the eye body by combining the data with the historical ophthalmological examination data; fitting the average parameters of the three parts of the eye body to the data, and combining multiple sets of data for the average initial parameters at the junction of the three parts of the eye body to obtain the initial parameter range of the junction of the eye body; and encapsulating the average initial parameters of each part of the three parts of the eye body and the initial parameter range of the junction of the eye body to obtain the initial parameters of the three-dimensional eye model.

[0016] Preferably, the initial parameters of the three-dimensional eye model are optimized to obtain the formula for selecting the three-dimensional eye model parameters:

[0017] P = (1-a)S(t-1) + M

[0018] Where P is the selection parameter for the 3D eye model, S is the initial parameter for the 3D eye model, a is the mixing parameter, a∈(0,1), t is the t-th cycle in the model training cycle, and M is the parameter update amount during model training.

[0019] Preferably, after optimizing the initial parameters of the three-dimensional eye model to obtain the three-dimensional eye model screening parameters, the method further includes verifying the accuracy of the three-dimensional eye model screening parameters: the server obtains the current time point and the previous verification time point; the server determines whether the verification time value has been reached based on the current time point and the previous verification time point; if so, the server sends a verification start command to the target subsystem; the server determines whether the three-dimensional eye model screening parameters have been received within a preset time; if so, the server performs a verification comparison between the three-dimensional eye model screening parameters and the historical ophthalmological examination data stored in the server, wherein the verification comparison is a comparison of the three-dimensional eye model screening parameters and the data set with the same attributes as the historical ophthalmological examination data; the server performs corresponding verification operations based on the comparison result of the verification comparison, wherein the comparison result is the result of whether the verification data set is abnormal.

[0020] Preferably, the step of the server verifying and comparing the three-dimensional eye model screening parameters with the historical ophthalmological examination data stored in the server includes: the server performing a set search on the historical ophthalmological examination data stored in the server based on the three-dimensional eye model screening parameters, wherein the set search is a search of data sets of the same type as the three-dimensional eye model screening parameters; the server extracting a first comparison data set based on the search results of the set search; the server verifying and comparing the three-dimensional eye model screening parameters with the first comparison data set; and the server verifying and comparing the three-dimensional eye model screening parameters with the first comparison data set. The verification and comparison process includes: the server calculating a first verification value for the eye 3D model screening parameters, the first verification value being the ratio of the eye 3D model screening parameters to the first comparison data set; the server calculating a second verification value for the first comparison data set, the second verification value being either 0 or 1; the server determining whether there is a value of 0 among the first verification value and the second verification value; if not, the server determining whether the first verification value and the second verification value are consistent; if the server determines that the first verification value and the second verification value are consistent, the server determining that the eye 3D model screening parameters are not abnormal.

[0021] Preferably, the steps of establishing a three-dimensional eye model based on the eye three-dimensional model selection parameters, collecting key eye points in the eye three-dimensional model, and binding eye selection point data to the key eye points include: establishing a three-dimensional eye model based on the eye three-dimensional model selection parameters; collecting key eye points in the eye three-dimensional model, wherein the key eye points are the center points of the core body, outer eye body, and inner eye body, the center intersection points of each part, and the center point of the core area of ​​eye detection; binding the eye detection three-dimensional model selection parameters to the key eye points, and connecting all the key eye points adjacently to form a three-dimensional eye structure network.

[0022] Preferably, the steps of normalizing the key points of the eye and performing filtering operations through big data comparison to obtain the final eye state information include: normalizing the three-dimensional structure network of the eye; converting the three-dimensional structure network of the eye into data; formatting each three-dimensional structure network of the eye into a unique data string to obtain a three-dimensional structure string of the eye; performing filtering operations on the three-dimensional structure string of the eye through big data, comparing multiple three-dimensional structure strings of the eye; filtering based on the distinguishability of the adjacent points of each key point of the eye in the three-dimensional structure string of the eye, obtaining the three-dimensional structure string of the eye with the lowest distinguishability, and obtaining the final eye state information.

[0023] Preferably, the step of performing geometric correction processing on the final eye state information after filtering and reconstructing the geometric structure to obtain a reconstructed three-dimensional eye model includes: filtering overlapping eye key points based on the repetition rate of the eye key points in the final eye state information; geometrically reconstructing the three-dimensional eye model based on the eye key points with the most overlaps at the same location based on the overlapping eye key points; and establishing the reconstructed three-dimensional eye model based on the geometrically reconstructed eye key points.

[0024] The beneficial effects of this invention are as follows: By collecting human eye state information, this invention performs preprocessing, data analysis, and parameter screening steps to correct the parameters of the three-dimensional eye model, making the data of the three-dimensional eye model more accurate and preventing errors in eye detection results and data. After establishing the three-dimensional eye model, key eye points are screened and normalized, allowing the three-dimensional eye model to be adjusted specifically for different patients. This enables the establishment of three-dimensional eye models for patients of different ages and physical conditions. Finally, geometric correction is performed to reconstruct the three-dimensional eye model, making the establishment of the three-dimensional eye model faster each time. It also allows for fine-tuning using historical three-dimensional eye models, making the reconstructed three-dimensional eye model more widely applicable and more convenient to use. Attached Figure Description

[0025] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0026] Figure 1 This is a flowchart illustrating a method for constructing a three-dimensional model based on ophthalmic examination, provided in an embodiment of the present invention.

[0027] Figure 2 This is a flowchart of the model construction steps for the three-dimensional eye model provided in this embodiment of the invention. Detailed Implementation

[0028] To enable those skilled in the art to better understand the present invention, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. 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.

[0029] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this invention are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product, or end that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or ends.

[0030] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0031] Current ophthalmic testing technologies primarily employ lasers for eye monitoring and scanning. However, existing eye monitoring technologies are not very effective. Because the eye is not a flat surface, the images generated by laser scanning often overlap, resulting in poor quality and potentially affecting test and diagnostic results. Therefore, a three-dimensional model construction method based on ophthalmic testing is essential.

[0032] The present invention aims to provide a method for constructing a three-dimensional model based on ophthalmic examination, in order to solve the problems of inaccurate eye examination in existing technologies, which can easily affect diagnostic results, and the significant differences between existing eye models and the actual size, position and shape of patients' eyes.

[0033] This invention provides a method for constructing a three-dimensional model based on ophthalmic examination, as described in the following embodiments: Figure 1 , Figure 2 As shown, it includes the following steps:

[0034] S1: Collect human eye condition information through eye detection equipment, and preprocess the human eye condition information to obtain preliminary eye condition information.

[0035] In this embodiment of the invention, collecting human eye status information through eye detection equipment includes: detecting basic human eye information, as well as measuring refractive power, visual field, ophthalmoscope, slit lamp, and intraocular pressure using ophthalmic ultrasound diagnostic instrument, OTC, ophthalmic AI screening software, and ophthalmic surgical microscope; filtering invalid data from the collected human eye status information and calculating the average data value of the human eye status information; if the human eye status information is higher / lower than 70% of the average data value of the human eye status information, the human eye status information is filtered out to obtain valid human eye status information; converting the valid human eye status information into a unified format and storing the valid human eye status information in a unified database to obtain preliminary eye status information.

[0036] S2: Based on historical ophthalmological examination data, perform data analysis on preliminary eye condition information to establish initial parameters for the three-dimensional eye model.

[0037] In this embodiment of the invention, basic information about the human eye and historical ophthalmological examination data are summarized and analyzed. The eye is divided into three parts from the inside out: the core body, the outer eye body, and the inner eye body. Preliminary eye state information is divided into data based on the three parts of the eye body, and the average initial parameters of each part of the eye body are calculated by combining them with historical ophthalmological examination data. The average parameters of the three parts of the eye body are fitted with data, and multiple sets of data are combined to obtain the initial parameter range of the eye body junction. The average initial parameters of each part of the eye body and the initial parameter range of the eye body junction are encapsulated to obtain the initial parameters of the three-dimensional eye model.

[0038] S3: Optimize the initial parameters of the 3D eye model to obtain the selection parameters for the 3D eye model.

[0039] In this embodiment of the invention, the initial parameters of the three-dimensional eye model are optimized to obtain the formula for selecting the three-dimensional eye model parameters:

[0040] P = (1-a)S(t-1) + M

[0041] Where P represents the eye 3D model selection parameters, S represents the initial parameters of the eye 3D model, a represents the hybrid parameters, a∈(0,1), t represents the t-th cycle in the model training period, and M represents the parameter update amount during model training. After optimizing the initial parameters of the eye 3D model to obtain the eye 3D model selection parameters, the accuracy of the eye 3D model selection parameters is also verified: the server obtains the current time point and the previous verification time point; the server determines whether the verification time value has been reached based on the current time point and the previous verification time point; if so, the server sends a verification start command to the target subsystem; the server determines whether the eye 3D model selection parameters have been received within a preset time; if so, the server performs a verification comparison between the eye 3D model selection parameters and the historical ophthalmological examination data stored in the server. The verification comparison is a comparison of the data sets with the same attributes of the eye 3D model selection parameters and the historical ophthalmological examination data; the server performs corresponding verification operations based on the comparison results, and the comparison result is whether the verification data set is accurate. An abnormal result; the server performs a set search on historical ophthalmological examination data stored on the server based on the eye 3D model screening parameters. The set search is a search of data sets of the same type as the eye 3D model screening parameters; the server extracts a first comparison data set based on the search results of the set search; the server performs a verification comparison between the eye 3D model screening parameters and the first comparison data set; the verification comparison between the server and the first comparison data set includes: the server calculates a first verification value for the eye 3D model screening parameters, which is the ratio of the eye 3D model screening parameters to the first comparison data set; the server calculates a second verification value for the first comparison data set, which can be 0 or 1; the server determines whether there is a value of 0 in the first verification value and the second verification value; if not, the server determines whether the first verification value and the second verification value are consistent; if the server determines that the first verification value and the second verification value are consistent, the server determines that the eye 3D model screening parameters are not abnormal.

[0042] S4: Establish a 3D eye model based on the eye 3D model selection parameters, collect key eye points in the eye 3D model, and bind eye selection point data to the key eye points.

[0043] In this embodiment of the invention, a three-dimensional eye model is established based on the eye three-dimensional model screening parameters; key points of the eye in the eye three-dimensional model are collected, including the center points of the core body, outer eye body, inner eye body, the center intersection points of each part, and the center point of the core area of ​​eye detection; the eye detection eye three-dimensional model screening parameters are bound to the key points of the eye, and all key points of the eye are connected adjacently to form a three-dimensional eye structure network.

[0044] S5: Normalize key points around the eyes, perform filtering calculations through big data comparison, and obtain the final eye condition information.

[0045] In this embodiment of the invention, the three-dimensional structure network of the eye is normalized, the three-dimensional structure network of the eye is converted into data, and each three-dimensional structure network of the eye is formatted into a unique data string to obtain a three-dimensional structure string of the eye; the three-dimensional structure string of the eye is filtered and compared through big data; the string of the eye is filtered based on the difference between the adjacent points of each key point of the eye, and the three-dimensional structure string of the eye with the lowest difference is obtained, thus obtaining the final eye state information.

[0046] S6: Use the final eye state information after filtering to perform geometric correction processing, reconstruct the geometric structure, and obtain a reconstructed three-dimensional eye model.

[0047] In this embodiment of the invention, the final eye state information is filtered for overlapping key points based on the repetition rate of key points; based on the overlapping key points, the eye three-dimensional model is geometrically reconstructed according to the key point position with the most overlaps at the same location; and the reconstructed eye three-dimensional model is established based on the geometrically reconstructed key points.

[0048] The beneficial effects of this invention are as follows: By collecting human eye state information, this invention performs preprocessing, data analysis, and parameter screening steps to correct the parameters of the three-dimensional eye model, making the data of the three-dimensional eye model more accurate and preventing errors in eye detection results and data. After establishing the three-dimensional eye model, key eye points are screened and normalized, allowing the three-dimensional eye model to be adjusted specifically for different patients. This enables the establishment of three-dimensional eye models for patients of different ages and physical conditions. Finally, geometric correction is performed to reconstruct the three-dimensional eye model, making the establishment of the three-dimensional eye model faster each time. It also allows for fine-tuning using historical three-dimensional eye models, making the reconstructed three-dimensional eye model more widely applicable and more convenient to use.

[0049] The above descriptions are merely embodiments of the present invention. Commonly known technical solutions or characteristics are not described in detail here. It should be noted that those skilled in the art can make various modifications and improvements without departing from the present invention, and these should also be considered within the scope of protection of the present invention. These modifications and improvements will not affect the effectiveness of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A three-dimensional model construction method based on ophthalmic examination, characterized by, The method comprises: S1: collecting human eye state information by an eye detection device, and preprocessing the human eye state information to obtain preliminary eye state information; S2: performing data analysis on the preliminary eye state information according to historical ophthalmic detection data to determine initial parameters of an eye three-dimensional model; S3: performing parameter optimization on the initial parameters of the eye three-dimensional model to obtain screening parameters of the eye three-dimensional model; S4: establishing an eye three-dimensional model according to the screening parameters of the eye three-dimensional model, collecting eye key points in the eye three-dimensional model, and binding eye screening point data to the eye key points; S5: performing normalization processing on the eye key points, performing screening operation through big data comparison, and obtaining final eye state information; S6: performing geometric correction processing on the final eye state information after screening operation, performing geometric structure reconstruction, and obtaining a reconstructed eye three-dimensional model; The formula for performing parameter optimization on the initial parameters of the eye three-dimensional model to obtain the screening parameters of the eye three-dimensional model is: P=(1-a)S(t-1)+M Wherein, P is the screening parameter of the eye three-dimensional model, S is the initial parameter of the eye three-dimensional model, a is a mixing parameter, a∈(0,1), t is the tth cycle in the model training period, and M is the parameter update amount of model training; The step of establishing an eye three-dimensional model according to the screening parameters of the eye three-dimensional model, collecting eye key points in the eye three-dimensional model, and binding eye screening point data to the eye key points comprises: Establishing an eye three-dimensional model according to the screening parameters of the eye three-dimensional model; Collecting eye key points in the eye three-dimensional model, wherein the eye key points are the center points of core bodies, outer eye bodies, inner eye bodies, the center intersection points of each part, and the center points of core areas of eye detection; Binding the eye three-dimensional model screening parameters of eye detection to the eye key points, and connecting all the eye key points adjacently to form an eye three-dimensional structure network; The step of performing normalization processing on the eye key points, performing screening operation through big data comparison, and obtaining final eye state information comprises: Performing normalization processing on the eye three-dimensional structure network, converting the eye three-dimensional structure network into data, converting each eye three-dimensional structure network into a unique data string in a format to obtain an eye three-dimensional structure string; Performing screening operation on the eye three-dimensional structure string through big data, and comparing a plurality of eye three-dimensional structure strings; Performing screening on the eye three-dimensional structure string according to the difference degree of the position of each eye key point adjacent to the eye three-dimensional structure string to obtain the eye three-dimensional structure string with the lowest difference degree, and obtaining final eye state information; The step of performing geometric correction processing on the final eye state information after screening operation, performing geometric structure reconstruction, and obtaining a reconstructed eye three-dimensional model comprises: Performing screening and overlapping of the eye key points according to the repetition rate of the eye key points. According to the superimposed eye key points, the eye three-dimensional model is geometrically reconstructed according to the position with the most superimpositions; The eye three-dimensional model is reconstructed according to the geometrically reconstructed eye key points.

2. The three-dimensional model construction method based on ophthalmic detection according to claim 1, characterized in that: The human eye state information collected by the eye detection device includes the basic information of the human eye detected by the ophthalmic ultrasonic diagnostic instrument, OTC, ophthalmic AI screening software, ophthalmic surgical microscope, diopter, visual field, ophthalmoscope, slit lamp, and tonometer.

3. The method of claim 2, wherein the method further comprises: The step of preprocessing the human eye state information to obtain preliminary eye state information includes: Invalid data filtering is performed on the collected human eye state information to calculate the average data value of the human eye state information; If the human eye state information is higher / lower than 70% of the average data value of the human eye state information, the human eye state information is excluded to obtain valid human eye state information; The valid human eye state information is converted into a unified format, and the valid human eye state information is stored in a unified database to obtain preliminary eye state information.

4. The method of claim 3, wherein the method further comprises: The step of analyzing the preliminary eye state information based on historical ophthalmic detection data to determine the initial parameters of the eye three-dimensional model includes: The human eye basic information and historical ophthalmic detection data are analyzed by dividing the eye into three parts: core body, outer eye body, and inner eye body; The preliminary eye state information is data-split according to the three parts of the eye, and the average initial parameters of each part of the eye are calculated by combining the historical ophthalmic detection data; The average parameters of the three parts of the eye are data-fitted, and the average initial parameters of the junctions of the three parts of the eye are combined to obtain the initial parameter interval of the eye junctions; The average initial parameters of each part of the eye and the initial parameter interval of the eye junctions are data-encapsulated to obtain the initial parameters of the eye three-dimensional model.

5. The three-dimensional model construction method based on ophthalmic detection according to claim 1, characterized in that: After the initial parameters of the eye three-dimensional model are optimized to obtain the screening parameters of the eye three-dimensional model, the accuracy of the screening parameters of the eye three-dimensional model is verified: The server obtains the current time point and the last verification time point; The server determines whether the verification time value is reached according to the current time point and the last verification time point; If yes, the server sends a verification start instruction to the target subsystem; The server determines whether the screening parameters of the eye three-dimensional model are received within a preset time; If yes, the server verifies and compares the screening parameters of the eye three-dimensional model with the historical ophthalmic detection data stored in the server, and the verification and comparison are the comparison of data sets with the same attributes between the screening parameters of the eye three-dimensional model and the historical ophthalmic detection data. The server performs a corresponding checking operation according to a comparison result of the comparison, and the comparison result is a result of whether the checking data set is abnormal.

6. The method of claim 5, wherein the three-dimensional model is constructed based on ophthalmic detection. The step of performing the checking comparison between the eye three-dimensional model screening parameter and the historical ophthalmic detection data stored in the server includes: The server performs a set retrieval according to the eye three-dimensional model screening parameter in the historical ophthalmic detection data stored in the server, and the set retrieval is a retrieval of a data set of the same type as the eye three-dimensional model screening parameter; The server extracts a first comparison data set according to a retrieval result of the set retrieval; The server performs the checking comparison between the eye three-dimensional model screening parameter and the first comparison data set; The server performs the checking comparison between the eye three-dimensional model screening parameter and the first comparison data set, including: The server calculates a first checking value of the eye three-dimensional model screening parameter, and the first checking value is a ratio of the eye three-dimensional model screening parameter to the first comparison data set; The server calculates a second checking value of the first comparison data set, and the second checking value is 0 or 1; The server determines whether there is a value of 0 in the first checking value and the second checking value; If not, the server determines whether the first checking value and the second checking value are consistent; If the server determines that the first checking value and the second checking value are consistent, the server determines that the eye three-dimensional model screening parameter is not abnormal.

Images