A myopia early warning method for teenagers and related device

Abstract

The application provides a myopia early warning method for teenagers and related equipment, and is applied to the technical field of data processing. The application obtains historical eye images of a plurality of users and eye habits of the corresponding users; the historical eye images of different users are compared with each other respectively to generate similarity results; the historical eye images of different users are divided based on the similarity results to generate corresponding user groups; real-time eye images of the users are obtained based on the eye habits; the real-time eye images are processed based on a preset classification model to generate target eye images; the historical eye images and the target eye images are processed based on an eye image calculation model to generate eye image change rates of the users; if the eye image change rate is greater than a preset threshold, the eye image change rate is processed based on a risk factor evaluation model to generate early warning information, thereby reducing the high incidence of myopia in teenagers.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a myopia early warning method for teenagers and related equipment. BACKGROUND

[0002] At present, when detecting whether a user has myopia, a professional detection method is usually used, such as a visual acuity detection table. A tester asks the user to identify the symbols on the detection table, and the visual acuity detection result is obtained according to the identification result. When myopia occurs, in order to see objects more clearly, the user often squints to reduce light incidence and increase the depth of field, so as to obtain a clearer field of view. At the same time, if the user with myopia has a long-term eye use condition, “dry eye syndrome” will also occur, which needs to be treated by frequent blinking to moisten the eyeball. Therefore, identifying the squinting and frequent blinking phenomenon is of great significance for early diagnosis of myopia.

[0003] At present, the gold standard for evaluating myopia is an optometry report, which needs to be carried out in a designated place (such as a hospital or an optical store) with an optometry machine, and cannot be carried out in real time in daily life for myopia risk assessment, and the application range is limited. It should be noted that mild myopia has less impact on visual acuity, and users are easy to get used to the blur of distant scenes, and because the near vision field is very clear, it will not greatly affect daily learning and life, so users themselves are also difficult to find visual abnormalities in the early stage of myopia.

[0004] Therefore, how to timely discover the risk in the early stage of myopia and then carry out effective intervention measures to slow down or even reverse the myopia trend is a problem that needs to be solved under the current severe myopia situation.

[0005] It should be noted that the information disclosed in the above background section is only used to strengthen the understanding of the background of the present disclosure, and therefore can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0006] The purpose of the present application is to provide a myopia early warning method for teenagers and related equipment and system, which at least to some extent overcomes the problems existing in the prior art, classifies different users according to the eye images of the users, screens the corresponding user groups, and obtains the eye habits of different users in the group, obtains the real-time eye images of the corresponding users according to the eye habits, so as to know the harmfulness of different eye habits to the eyes of the users, and early warns the users according to this, so as to reduce the high incidence of myopia in teenagers.

[0007] Other characteristics and advantages of the present application will become apparent from the following detailed description, or will be learned partly through practice of the present application.

[0008] According to an aspect of the present application, a myopia early warning method and related device for teenagers are provided, including: obtaining historical eye images of a plurality of users and eye habits of the corresponding users; respectively comparing the historical eye images of different users with each other to generate similarity results; dividing the historical eye images of different users based on the similarity results to generate corresponding user groups, wherein the user groups include identification information representing the eye image states of different users in the same group; obtaining real-time eye images of the users based on the eye habits; processing the real-time eye images based on a preset classification model to generate target eye images, wherein the target eye images include the influence degrees of a plurality of physiological state factors on the real-time eye images; processing the historical eye images and the target eye images based on an eye image calculation model to generate an eye image change rate of the user, wherein the eye image calculation model includes a calculation formula for calculating the eye image change rate, and the formula is: P = a * Tn / (S1-S2); P represents the eye image change rate, a is a preset normalization factor, Tn is the duration of a second time period, and (S1-S2) is the change value of the eye image in a first time period and a second time period; if the eye image change rate is greater than a preset threshold, processing the eye image change rate based on a risk factor evaluation model to generate early warning information.

[0009] In an embodiment of the present application, the processing of the real-time eye images based on the preset classification model to generate the target eye images includes: obtaining life environment data of the user; and setting a preset weight value for the physiological state factors of the user based on the life environment data.

[0010] In an embodiment of the present application, the processing of the eye image change rate based on the risk factor evaluation model to generate the early warning information includes: processing the eye image change rate based on the risk factor evaluation model to generate risk attributes of each physiological state factor; obtaining the influence degrees of the physiological state factors and the eye image change rate based on the risk attributes; processing the preset weight values of the physiological state factors based on a risk classification rule to generate risk abnormal indicators, wherein the risk abnormal indicators are sorted in descending order according to the sizes of the weight values; and generating the early warning information based on the risk abnormal indicators.

[0011] In an embodiment of the present application, the processing of the real-time eye image based on the preset classification model to generate a target eye image comprises: performing grayscale processing on the real-time eye image to generate an initial color image; obtaining initial features based on the initial color image; performing normalization processing on the initial color image based on a residual neural network model to generate a target color image of a preset size; performing global average pooling processing on the target color image of the preset size based on the residual neural network model to generate target features; and generating a target eye image based on the target features.

[0012] In an embodiment of the present application, before the normalization processing on the initial color image based on the residual neural network model to generate a target color image of a preset size, the method further comprises: performing slice processing on the target color image to sequentially generate a plurality of continuous slice images; processing the slice images to generate a multi-channel image; and inputting the multi-channel image into a residual neural network model for processing to generate the initial color image.

[0013] In an embodiment of the present application, the processing of the real-time eye image based on the preset classification model to generate a target eye image further comprises: performing classification processing on the slice images to generate target slice images of different categories; processing the target slice images based on a preset classification threshold to generate inflammatory cells in an abnormal state; processing the inflammatory cells in the abnormal state based on a preset rule to generate an inflammatory cell proportion value; and generating a target eye image based on the inflammatory cell proportion value.

[0014] In an embodiment of the present application, the processing of the eye image change rate based on the risk factor evaluation model to generate early warning information comprises: obtaining a real-time parameter fluctuation probability of the physiological state factor; if the fluctuation probability is greater than a preset fluctuation threshold, determining whether the physiological state factor corresponding to the fluctuation probability is consistent based on the historical physiological state factor; if yes, increasing the risk value of the physiological state factor; if no, obtaining the number of times that the physiological state factor fluctuation probability is greater than the preset fluctuation threshold within a preset time, and if the number of times is greater than a preset number of times, increasing the risk value of the physiological state factor; and generating early warning information based on the physiological state factor.

[0015] In another aspect of the present application, a myopia early warning device for teenagers is characterized in that it comprises: an acquisition module configured to acquire historical eye images of a plurality of users and eye habits of the corresponding users; acquire real-time eye images of the users based on the eye habits; a processing module configured to compare the historical eye images of different users with each other respectively to generate similarity results; divide the historical eye images of different users based on the similarity results to generate corresponding user groups, wherein the user groups comprise identification information for representing the eye image states of different users in the same group; process the real-time eye images based on a preset classification model to generate target eye images, wherein the target eye images comprise the influence degrees of a plurality of physiological state factors on the real-time eye images; process the historical eye images and the target eye images based on an eye image calculation model to generate an eye image change rate of the user, wherein the eye image calculation model comprises a calculation formula for calculating the eye image change rate, and the formula is: P=a*Tn / (S1-S2); P represents the eye image change rate, a is a preset normalization factor, Tn is the time length of a second time period, and (S1-S2) is the change value of the eye images in a first time period and a second time period; if the eye image change rate is greater than a preset threshold, process the eye image change rate based on a risk factor evaluation model to generate early warning information.

[0016] According to still another aspect of the present application, an electronic device is characterized in that it comprises: a first processor; and a memory configured to store executable instructions of the first processor; wherein the first processor is configured to execute the executable instructions to implement the above-mentioned myopia early warning method for teenagers.

[0017] According to still another aspect of the present application, a computer readable storage medium is provided, which stores a computer program, and the computer program is executed by a second processor to implement the above-mentioned myopia early warning method for teenagers.

[0018] According to still another aspect of the present application, a computer program product is provided, which comprises a computer program, and the computer program is executed by a third processor to implement the above-mentioned myopia early warning method for teenagers.

[0019] The method and related device for early warning of myopia of teenagers provided by the application, a server acquires historical eye images of a plurality of users and eye habits of the corresponding users; the historical eye images of different users are compared with each other respectively to generate similarity results; the historical eye images of different users are divided based on the similarity results to generate corresponding user groups, wherein the user groups include identification information for representing the eye image states of different users in the same group; a real-time eye image of the user is acquired based on the eye habits; the real-time eye image is processed based on a preset classification model to generate a target eye image, wherein the target eye image includes the influence degrees of a plurality of physiological state factors on the real-time eye image; the historical eye images and the target eye image are processed based on an eye image calculation model to generate an eye image change rate of the user, wherein the eye image calculation model includes a calculation formula for calculating the eye image change rate, and the formula is P=a*Tn / (S1-S2); P represents the eye image change rate, a is a preset normalization factor, Tn is the time length of a second time period, and (S1-S2) is the change value of the eye image in a first time period and a second time period; if the eye image change rate is greater than a preset threshold, the eye image change rate is processed based on a risk factor evaluation model to generate early warning information. Different users are classified according to the eye images of the users, corresponding user groups are screened out, and the eye habits of different users in the groups are acquired, the real-time eye image of the corresponding user is acquired according to the eye habits, the harmfulness of different eye habits to the eyes of the user is learned, and the user is warned according to this, so as to reduce the high incidence of myopia of teenagers.

[0020] It should be understood that the foregoing general description and the following detailed description are only exemplary and explanatory, and cannot limit the present disclosure. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 A flowchart of a method for early warning of myopia of teenagers provided by an embodiment of the application is shown;

[0022] Figure 2 A structural schematic diagram of an early warning device for myopia of teenagers provided by an embodiment of the application is shown;

[0023] Figure 3 A structural schematic diagram of an electronic device provided by an embodiment of the application is shown;

[0024] Figure 4 A schematic diagram of a storage medium provided by an embodiment of the application is shown. DETAILED DESCRIPTION

[0025] The preferred embodiments of the present application will be described below in conjunction with the drawings, which should be understood as illustrative in nature and are not used to limit the present application.

[0026] The early warning method for myopia of teenagers according to the exemplary embodiments of the present application will be described below in conjunction with Figure 1 It should be noted that the following application scenarios are only shown for the purpose of facilitating the understanding of the spirit and principles of the present application, and the embodiments of the present application are not limited in this respect. On the contrary, the embodiments of the present application can be applied to any applicable scenario.

[0027] In one embodiment, the present application further provides an early warning method for myopia of teenagers and related equipment. Figure 1 The flowchart of an early warning method for myopia of teenagers according to the embodiments of the present application is schematically shown. As Figure 1 shown, the method is applied to a server and includes:

[0028] S101, obtaining historical eye images of a plurality of users and eye habits of the corresponding users.

[0029] In one embodiment, the server will obtain eye images of a plurality of users within a specific time period, and the number of the eye images is at least one, and if there are multiple eye images, a suitable eye image can be selected according to actual needs. In addition, the server will also obtain the eye habits of the corresponding users, wherein the eye habits include but are not limited to the daily life habits of the users, such as the posture of reading books, the light of the scene, the eye use time of the users, whether the users often play games, the game time, the game frequency, etc.

[0030] S102, respectively comparing and processing the historical eye images of different users with each other to generate similarity results.

[0031] In one embodiment, the server will randomly or purposefully select at least one group of eye images from the historical eye images of different users as a reference group, and the remaining eye images as a control group, and by comparing the eye images in the control group with the eye images in the reference group, the similarity between different eye images is obtained.

[0032] S103, dividing the historical eye images of different users based on the similarity results to generate corresponding user groups.

[0033] In an embodiment, the server divides the historical eye images of different users according to the similarity between different eye images, and generates different user groups, wherein the user groups include a healthy group, a cataract group, an eye dryness group, an eye disease group, etc., and different users are divided into corresponding groups according to different user groups. In addition, the user groups include identification information for representing the state of the eye images of different users in the same group.

[0034] S104, acquiring a real-time eye image of the user based on the eye use habit.

[0035] In an embodiment, the server acquires the real-time eye image of the user according to the user's living habits, such as reading posture, daily activity light, eye use time, whether playing games, whether taking a break for the eyes (eye exercises), etc., so as to understand the real-time eye image of each user under different eye use habits, and further understand the correlation between the abnormal eye image of the user and the corresponding eye use habit.

[0036] S105, processing the real-time eye image based on a preset classification model to generate a target eye image.

[0037] In an embodiment, the server acquires the living environment data of the user, sets a preset weight value for the physiological state factors of the user based on the living environment data. The server acquires the correlation degree between each physiological risk factor and the eye abnormality based on the living environment data, and sorts based on the correlation degree, so as to filter out the physiological risk factor that has the greatest impact on the eye abnormality, and sets a preset weight value for the physiological state factors of the user based on the sorting result. For example, if the user lives in a place that is too dark, or the light in the home or work environment is lower than the normal light intensity, working or living in such an environment will affect the eye health of the user to some extent. In addition, if the user is habitually in a too humid or dry environment for a long time, it will also affect the eye health of the user to some extent. The target eye image includes the influence degree of a plurality of physiological state factors on the real-time eye image. Through the above factors, the weight value of the physiological risk factor that may affect the eye of the user is adjusted, in addition, the home environment or living environment of the user can also be improved according to the above situation, so as to reduce the further deterioration of the eye of the user.

[0038] In another implementation, the real-time eye image is subjected to grayscale processing to generate an initial color image, initial features are obtained based on the initial color image, the initial color image is subjected to normalization processing based on the residual neural network model to generate a target color image of a preset size, the target color image of the preset size is subjected to global average pooling processing based on the residual neural network model to generate target features, and a target eye image is generated based on the target features. The preset classification model is an image classification model, which includes a convolution layer, residual blocks, and a fully connected layer. Specifically, the convolution layer includes an input layer, the input layer includes a single-channel 2048x2048 grayscale image data as the input of the network, contains a 6x6 convolution kernel, the input channel number is 1, the output channel number is 64, and the stride is 2. This convolution layer is used to extract low-level features from the input image, where the low-level features refer to edge features, texture features, corner features, shape features, etc.

[0039] In addition, batch normalization is added after the convolution layer to accelerate convergence and improve the stability of the model. The output of the convolution layer is subjected to ReLU activation function processing to increase the non-linear expression capability of the network. In addition, a max pooling layer is arranged after the convolution layer, which uses a 3x3 pooling kernel and a stride of 2 to reduce the size of the feature map. The preset classification model also includes multiple residual blocks, each of which includes two convolution layers and a skip connection for learning residual. After the last residual module, a global average pooling layer is connected, which converts the feature map into a fixed-size feature vector for classification tasks. The preset classification model finally has a fully connected layer that maps the feature vector to the probability of 4 output categories to adapt to the classification task of the target eye image.

[0040] In S106, the eye image computing model is used to process the historical eye image and the target eye image to generate an eye image change rate of the user.

[0041] In one implementation, the eye image computing model includes a calculation formula for calculating the eye image change rate, which is P=a*Tn / (S1-S2). P represents the eye image change rate, a is a preset normalization factor, Tn is the duration of the second time period, and (S1-S2) is the change value of the eye image in the first time period and the second time period. For example, if the eye image change rate of the user in the previous two time periods is greater than 1%, it means that the current eye image of the user may deteriorate, including but not limited to deepening the eye power, inflammation of the eye, eyeball deformation, etc. Similarly, if the user completes the treatment of the eyeball according to the doctor's guidance, it indicates that the current eye image of the user has improved compared to before. In addition, the size of the eye image change rate is not limited in this solution.

[0042] S107, if the eye image change rate is greater than a preset threshold, processing the eye image change rate based on a risk factor evaluation model to generate early warning information.

[0043] In an embodiment, the eye image change rate is processed based on the risk factor evaluation model to generate a risk attribute of each physiological state factor, the influence degree of the physiological state factor on the eye image change rate is obtained based on the risk attribute, a preset weight value of the physiological state factor is processed based on a risk classification rule to generate a risk abnormality index, wherein the risk abnormality index is sorted in descending order according to the size of the weight value, and the early warning information is generated based on the risk abnormality index. The risk factor evaluation model learns from a large number of user eye images, filters out the basic parameters related to user risk assessment and prediction warning, and collects information according to the filtered information. Specifically, dynamic data and static data can be collected from data sources such as clinical data center, electronic medical record system, laboratory test system, network information, surgical nursing information system, and post-hospital follow-up system.

[0044] In another embodiment, the real-time parameter floating probability of the physiological state factor is obtained, if the floating probability is greater than a preset floating threshold, it is judged whether the physiological state factor corresponding to the floating probability is consistent based on the historical physiological state factor, if yes, the risk value of the physiological state factor is increased, if no, the number of times that the physiological state factor floating probability is greater than the preset floating threshold within a preset time is obtained, if the number of times is greater than a preset number of times, the risk value of the physiological state factor is increased, and the early warning information is generated based on the physiological state factor.

[0045] In the present application, the server obtains historical eye images of a plurality of users and eye habits of the corresponding users, respectively compares and processes the historical eye images of different users to generate a similarity result, divides and processes the historical eye images of different users based on the similarity result to generate corresponding user groups, wherein the user groups include identification information for representing the eye image states of different users in the same group, obtains real-time eye images of the users based on the eye habits, processes the real-time eye images based on a preset classification model to generate target eye images, wherein the target eye images include the influence degree of a plurality of physiological state factors on the real-time eye images, processes the historical eye images and the target eye images based on an eye image calculation model to generate an eye image change rate of the user, and processes the eye image change rate based on a risk factor evaluation model to generate early warning information if the eye image change rate is greater than a preset threshold. According to the eye images of the users, different users are classified, corresponding user groups are selected, and eye habits of different users in the group are obtained. The real-time eye images of the corresponding users are obtained according to the eye habits, so as to know the harmfulness of different eye habits to the eyes of the users, and the users are warned according to this, so as to reduce the high incidence of myopia in teenagers.

[0046] Optionally, in another embodiment based on the above method of the application, before the residual neural network model is used to normalize the initial color image to generate a target color image of a preset size, the method further comprises:

[0047] slicing the target color image to generate a plurality of continuous slice images in sequence;

[0048] processing the slice images to generate a multi-channel image;

[0049] inputting the multi-channel image into a residual neural network model for processing to generate the initial color image.

[0050] In one embodiment, the initial color image is processed based on a preset color division rule to generate a target color image containing different classification numbers, the target color image is divided based on a preset classification number to generate a plurality of initial slice images, the initial slice images are preprocessed based on a target segmentation model to generate preprocessed slice images, and the preprocessed slice images are processed based on a preset bilinear interpolation to generate a plurality of slice images of a preset size. The color image is labeled and divided according to the color depth of the target image, all slices are preprocessed, all preprocessed slices are input into an optimal segmentation model to obtain a segmentation result, and the segmentation result is bilinearly interpolated and upsampled to restore to the original size.

[0051] Optionally, in another embodiment based on the above method of the application, the real-time eye image is processed based on a preset classification model to generate a target eye image, and the method further comprises:

[0052] classifying the slice images to generate target slice images of different categories;

[0053] processing the target slice images based on a preset classification threshold to generate inflammatory cells in an abnormal state;

[0054] processing the inflammatory cells in an abnormal state based on a preset rule to generate an inflammatory cell proportion value;

[0055] generating a target eye image based on the inflammatory cell proportion value.

[0056] In one embodiment, the training samples of different age groups and different user groups are compared and learned with the slice images of the current user to strengthen the learning and reinforcement of the preset classification model on the slice images of the eye images, thereby improving the cross-domain generalization of the preset classification model. The same clinical phenotype may have different symptoms, and understanding the pathogenesis of different types of eye diseases can provide more basis for formulating precise medical strategies.

[0057] By applying the above technical solution, the server obtains historical eye images of a plurality of users and eye habits of the corresponding users, respectively compares the historical eye images of different users with each other, generates a similarity result, divides the historical eye images of different users based on the similarity result, and generates a corresponding user group. The user group includes identification information representing the eye image states of different users in the same group. The real-time eye image of the user is obtained based on the eye habits.

[0058] The server obtains life environment data of the user, sets a preset weight value for the physiological state factors of the user based on the life environment data, performs grayscale processing on the real-time eye image to generate an initial color image, obtains an initial feature based on the initial color image, performs slice processing on the target color image to generate a plurality of continuous slice images in sequence, processes the slice images to generate a multi-channel image, inputs the multi-channel image into a residual neural network model for processing to generate an initial color image, performs classification processing on the slice image to generate target slice images of different categories, processes the target slice images based on a preset classification threshold to generate inflammatory cells in an abnormal state, processes the inflammatory cells in the abnormal state based on a preset rule to generate an inflammatory cell proportion value, generates a target eye image based on the inflammatory cell proportion value, performs normalization processing on the initial color image based on the residual neural network model to generate a target color image of a preset size, performs global average pooling processing on the target color image of the preset size based on the residual neural network model to generate a target feature, and generates a target eye image based on the target feature. The target eye image includes the influence degree of a plurality of physiological state factors on the real-time eye image.

[0059] Further, the server also processes the historical eye image and the target eye image based on an eye image calculation model to generate an eye image change rate of the user, wherein the eye image calculation model comprises a calculation formula for calculating the eye image change rate, and the formula is P=a*Tn / (S1-S2), P represents the eye image change rate, a is a preset normalization factor, Tn is the time length of the second time period, (S1-S2) is the change value of the eye image in the first time period and the second time period, if the eye image change rate is greater than a preset threshold, the eye image change rate is processed based on the risk factor evaluation model to generate the risk attribute of each physiological state factor, the influence degree of the physiological state factor and the eye image change rate is obtained based on the risk attribute, and the preset weight value of the physiological state factor is processed based on the risk classification rule to generate a risk abnormal index, wherein the risk abnormal index is sorted in descending order according to the size of the weight value, the warning information is generated based on the risk abnormal index, the real-time parameter floating probability of the physiological state factor is obtained, if the floating probability is greater than a preset floating threshold, whether the physiological state factor corresponding to the floating probability is consistent is determined based on the historical physiological state factor, if yes, the risk value of the physiological state factor is increased, if no, the number of times that the physiological state factor floating probability is greater than the preset floating threshold is obtained within a preset time, if the number of times is greater than a preset number of times, the risk value of the physiological state factor is increased, and the warning information is generated based on the physiological state factor. According to the eye images of the users, different users are classified, corresponding user groups are screened out, and the eye habits of different users in the group are obtained, the real-time eye images of the corresponding users are obtained according to the eye habits, so that the harmfulness of different eye habits to the eyes of the users is known. By comparing and learning the training samples of different age groups and different user groups with the slice images of the current user, the learning and reinforcement of the preset classification model on the slice images of the eye images are strengthened, and the cross-domain generalization of the preset classification model is improved. The same clinical phenotype may have different symptoms, and understanding the pathogenesis of different types of eye diseases can provide more basis for formulating precise medical strategies, and the user is warned according to this, so as to reduce the high incidence of myopia in adolescents.

[0060] In an embodiment, as shown in Figure 2 The application also provides a warning device for adolescent myopia, comprising:

[0061] The acquisition module 201 is configured to acquire historical eye images of a plurality of users and eye habits of corresponding users, and acquire real-time eye images of the users based on the eye habits.

[0062] The processing module 202 is configured to respectively compare the historical eye images of different users with each other to generate a similarity result, divide the historical eye images of different users based on the similarity result to generate a corresponding user group, wherein the user group includes identification information for representing the eye image states of different users in the same group, process the real-time eye image based on a preset classification model to generate a target eye image, wherein the target eye image includes the influence degree of a plurality of physiological state factors on the real-time eye image, process the historical eye image and the target eye image based on an eye image calculation model to generate an eye image change rate of the user, wherein the eye image calculation model includes a calculation formula for calculating the eye image change rate, and the formula is P = a * Tn / (S1-S2); P represents the eye image change rate, a is a preset normalization factor, Tn is the time length of a second time period, and (S1-S2) is the change value of the eye image in the first time period and the second time period; and if the eye image change rate is greater than a preset threshold, process the eye image change rate based on a risk factor evaluation model to generate early warning information.

[0063] In the present application, the server obtains the historical eye images of a plurality of users and the eye habits of the corresponding users, respectively compares the historical eye images of different users with each other to generate a similarity result, divides the historical eye images of different users based on the similarity result to generate a corresponding user group, wherein the user group includes identification information for representing the eye image states of different users in the same group, obtains the real-time eye image of a user based on the eye habits, processes the real-time eye image based on a preset classification model to generate a target eye image, wherein the target eye image includes the influence degree of a plurality of physiological state factors on the real-time eye image, processes the historical eye image and the target eye image based on an eye image calculation model to generate an eye image change rate of the user, and if the eye image change rate is greater than a preset threshold, processes the eye image change rate based on a risk factor evaluation model to generate early warning information. According to the eye images of the users, different users are classified, the corresponding user group is screened out, the eye habits of different users in the group are obtained, the real-time eye image of the corresponding user is obtained according to the eye habits, so as to know the harmfulness of different eye habits to the eyes of the user, and the user is warned according to this, so as to reduce the high incidence of myopia in teenagers.

[0064] In another embodiment of the present application, the processing module 202 is configured to process the real-time eye image based on a preset classification model to generate a target eye image, including: obtaining the life environment data of the user; and setting a preset weight value for the physiological state factors of the user based on the life environment data.

[0065] In another implementation of the present application, the processing module 202 is configured to process the eye image change rate based on the risk factor evaluation model to generate early warning information, including: processing the eye image change rate based on the risk factor evaluation model to generate risk attributes of each physiological state factor; obtaining the influence degree of the physiological state factor and the eye image change rate based on the risk attributes; processing the preset weight value of the physiological state factor based on the risk classification rule to generate a risk abnormality index, wherein the risk abnormality index is sorted in descending order according to the size of the weight value; generating early warning information based on the risk abnormality index.

[0066] In another implementation of the present application, the processing module 202 is configured to process the real-time eye image based on the preset classification model to generate a target eye image, including: performing grayscale processing on the real-time eye image to generate an initial color image; obtaining initial features based on the initial color image; performing normalization processing on the initial color image based on a residual neural network model to generate a target color image of a preset size; performing global average pooling processing on the target color image of the preset size based on the residual neural network model to generate target features; and generating a target eye image based on the target features.

[0067] In another implementation of the present application, before the processing module 202 is configured to perform normalization processing on the initial color image based on the residual neural network model to generate a target color image of a preset size, it further includes: performing slice processing on the target color image to sequentially generate a plurality of continuous slice images; processing the slice images to generate a multi-channel image; and inputting the multi-channel image into a residual neural network model for processing to generate the initial color image.

[0068] In another implementation of the present application, the processing module 202 is configured to process the real-time eye image based on the preset classification model to generate a target eye image, and further includes: performing classification processing on the slice image to generate target slice images of different categories; processing the target slice image based on a preset classification threshold to generate inflammation cells in an abnormal state; processing the inflammation cells in the abnormal state based on a preset rule to generate an inflammation cell proportion value; and generating a target eye image based on the inflammation cell proportion value.

[0069] In another implementation of the present application, the processing module 202 is configured to process the eye image change rate based on the risk factor evaluation model to generate early warning information, including: obtaining a real-time parameter floating probability of the physiological state factor; if the floating probability is greater than a preset floating threshold, determining whether the physiological state factor corresponding to the floating probability is consistent based on the historical physiological state factor; if yes, increasing the risk value of the physiological state factor; if no, obtaining the number of times that the physiological state factor floating probability is greater than the preset floating threshold within a preset time, and if the number of times is greater than a preset number of times, increasing the risk value of the physiological state factor; and generating early warning information based on the physiological state factor.

[0070] By applying the above technical solutions, the server obtains historical eye images of a plurality of users and eye habits of the corresponding users, respectively compares and processes the historical eye images of different users with each other to generate similarity results, performs division processing on the historical eye images of different users based on the similarity results, and generates corresponding user groups, wherein the user groups include identification information for representing the eye image states of different users in the same group, and the real-time eye image of the user is obtained based on the eye habits.

[0071] The server obtains life environment data of the user, sets a preset weight value for the physiological state factor of the user based on the life environment data, performs grayscale processing on the real-time eye image to generate an initial color image, obtains an initial feature based on the initial color image, performs slice processing on the target color image to sequentially generate a plurality of continuous slice images, processes the slice images to generate a multi-channel image, inputs the multi-channel image into a residual neural network model for processing to generate an initial color image, performs classification processing on the slice images to generate target slice images of different categories, processes the target slice images based on a preset classification threshold to generate inflammation cells in an abnormal state, processes the inflammation cells in the abnormal state based on a preset rule to generate an inflammation cell proportion value, generates a target eye image based on the inflammation cell proportion value, performs normalization processing on the initial color image based on the residual neural network model to generate a target color image of a preset size, performs global average pooling processing on the target color image of the preset size based on the residual neural network model to generate a target feature, and generates the target eye image based on the target feature, wherein the target eye image includes the influence degree of a plurality of physiological state factors on the real-time eye image.

[0072] Further, the server also processes the historical eye image and the target eye image based on an eye image calculation model to generate an eye image change rate of the user, wherein the eye image calculation model comprises a calculation formula for calculating the eye image change rate, and the formula is P=a*Tn / (S1-S2), P represents the eye image change rate, a is a preset normalization factor, Tn is the time length of the second time period, (S1-S2) is the change value of the eye image in the first time period and the second time period, if the eye image change rate is greater than a preset threshold, the eye image change rate is processed based on the risk factor evaluation model to generate the risk attribute of each physiological state factor, the influence degree of the physiological state factor and the eye image change rate is obtained based on the risk attribute, and the preset weight value of the physiological state factor is processed based on the risk classification rule to generate a risk abnormal index, wherein the risk abnormal index is sorted in descending order according to the size of the weight value, the warning information is generated based on the risk abnormal index, the real-time parameter floating probability of the physiological state factor is obtained, if the floating probability is greater than a preset floating threshold, whether the physiological state factor corresponding to the floating probability is consistent is determined based on the historical physiological state factor, if yes, the risk value of the physiological state factor is increased, if no, the number of times that the physiological state factor floating probability is greater than the preset floating threshold within a preset time is obtained, if the number of times is greater than a preset number of times, the risk value of the physiological state factor is increased, and the warning information is generated based on the physiological state factor. Different users are classified according to the eye image of the user, the corresponding user group is screened out, and the eye habits of different users in the group are obtained, the real-time eye image of the corresponding user is obtained according to the eye habits, so that the harmfulness of different eye habits to the user's eyes is known. By comparing and learning the training samples of different age groups and different user groups with the slice image of the current user, the learning and reinforcement of the preset classification model on the slice image of the eye image are strengthened, and the cross-domain generalization of the preset classification model is improved. The same clinical phenotype may have different symptoms, and understanding the pathogenesis of different types of eye diseases can provide more basis for formulating precise medical strategies, and the user is warned according to this, so as to reduce the high incidence of myopia in adolescents.

[0073] The embodiment of the present application provides an electronic device, such as Figure 3 As shown in the figure, the electronic device 3 comprises a first processor 300, a memory 301, a bus 302 and a communication interface 303, the first processor 300, the communication interface 303 and the memory 301 are connected through the bus 302; the memory 301 stores a computer program capable of running on the first processor 300, and the first processor 300 executes the computer program to perform the early warning method for myopia of adolescents provided by any one of the preceding embodiments of the present application.

[0074] The memory 301 can include a random access memory (RAM) and can also include a non-volatile memory, such as at least one disk memory. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 303 (which can be wired or wireless), and the Internet, a wide area network, a local network, a metropolitan area network, etc. can be used.

[0075] The bus 302 can be an ISA bus, a PCI bus, an EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. The memory 301 is used to store a program, and the first processor 300 executes the program after receiving an execution instruction. The early warning method for myopia of teenagers disclosed in any of the embodiments of the present application can be applied to the first processor 300 or implemented by the first processor 300.

[0076] The first processor 300 can be an integrated circuit chip with signal processing capability. In the implementation process, each step of the above method can be completed by integrated logic circuits of hardware in the first processor 300 or instructions in the form of software. The first processor 300 described above can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a ready-to-program gate array (FPGA), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. Each method, step and logic block disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in combination with the embodiments of the present application can be embodied as a hardware code processor for execution or a combination of hardware and software modules in the code processor for execution. The software module can be located in a random memory, a flash memory, a read-only memory, a programmable read-only memory, an electrically erasable programmable memory, a register, etc. The storage medium in the art. The storage medium is located in the memory 301, and the first processor 300 reads the information in the memory 301 and combines the hardware to complete the steps of the above method.

[0077] The electronic device provided by the above embodiments of the present application has the same beneficial effects as the early warning method for myopia of teenagers provided by the embodiments of the present application and the methods adopted, run or implemented by the application programs stored therein.

[0078] The computer readable storage medium provided by the embodiments of the present application, such as a computer readable storage medium,Figure 4 As shown, the computer readable storage medium stores 401 a computer program, which, when read and run by the second processor 402, implements the early warning method for myopia of teenagers as described above.

[0079] The technical solutions of the embodiments of the present application essentially or the parts that make contributions to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product, which is stored in a storage medium and includes instructions for making an electronic device (which can be an air conditioner, a refrigeration device, a personal computer, a server, or a network device) or a processor execute all or part of the steps of the method described in the embodiments of the present application. The storage medium described above includes a U disk, a mobile hard disk, a ROM, a RAM, a magnetic disk or an optical disk, and various media that can store program codes.

[0080] The computer readable storage medium provided by the above embodiments of the present application and the early warning method for myopia of teenagers provided by the embodiments of the present application are based on the same inventive concept, and have the same beneficial effects as the method adopted, run or implemented by the application program stored therein.

[0081] The embodiments of the present application provide a computer program product, including a computer program, which, when executed by a third processor, implements the method as described above.

[0082] The computer program product provided by the above embodiments of the present application and the early warning method for myopia of teenagers provided by the embodiments of the present application are based on the same inventive concept, and have the same beneficial effects as the method adopted, run or implemented by the application program stored therein.

[0083] It should be noted that in the present application, 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 terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device 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 device. Without more limitations, the element defined by the statement "including a" does not exclude the presence of other identical elements in the process, method, article or device including the element.

[0084] The various embodiments in the present application are described in a related manner, and the same or similar parts between various embodiments can be referred to each other. Each embodiment focuses on the difference from other embodiments. In particular, the early warning method for evaluating myopia of teenagers, the electronic device, the electronic equipment, and the readable storage medium embodiment are basically similar to the early warning method for evaluating myopia of teenagers described above, so the description is relatively simple, and the relevant parts can be referred to the description of the early warning method for evaluating myopia of teenagers described above.

[0085] Although the present application is disclosed as above, the present application is not limited thereto. Any person skilled in the art can make various changes and modifications without departing from the spirit and scope of the present application, and the protection scope of the present application should be subject to the scope defined by the claims.

Claims

1. A method of early warning of myopia in adolescents, characterized in that, The method comprises the following steps: acquiring historical eye images of a plurality of users and eye habits of corresponding users; respectively comparing the historical eye images of different users with each other to generate similarity results; dividing the historical eye images of different users based on the similarity results to generate corresponding user groups, wherein the user groups include identification information representing the eye image states of different users in the same group; acquiring real-time eye images of the users under the eye habits based on the eye habits; processing the real-time eye images based on a preset classification model to generate target eye images, wherein the target eye images include the influence degrees of a plurality of physiological state factors on the real-time eye images; The historical eye image and the target eye image are processed based on an eye image calculation model to generate an eye image change rate of the user, wherein the eye image calculation model comprises a calculation formula for calculating the eye image change rate, and the formula is: P=a Tn / (S1-S2); P represents the eye image change rate, a is a preset normalization factor, Tn is a time length of a second time period, (S1-S2) is a change value of the eye image in the first time period and the second time period, the first time period is a time interval for acquiring the historical eye image, and the second time period is a time interval for acquiring the real-time eye image. if the eye image change rate is greater than a preset threshold, processing the eye image change rate based on a risk factor evaluation model to generate warning information, including acquiring a real-time parameter fluctuation probability of the physiological state factors; if the fluctuation probability is greater than a preset fluctuation threshold, determining whether the physiological state factors corresponding to the fluctuation probability are consistent based on historical physiological state factors; if yes, increasing the risk value of the physiological state factors; if no, acquiring the number of times that the fluctuation probability of the physiological state factors is greater than the preset fluctuation threshold within a preset time, and if the number of times is greater than a preset number of times, increasing the risk value of the physiological state factors; and generating warning information based on the physiological state factors.

2. The method of claim 1, wherein, The method further comprises the following steps: acquiring life environment data of the users; setting a preset weight value for the physiological state factors of the users based on the life environment data.

3. The method of claim 2, wherein, The method further comprises the following steps: processing the eye image change rate based on a risk factor evaluation model to generate risk attributes of each physiological state factor; acquiring the influence degrees of the physiological state factors and the eye image change rate based on the risk attributes; processing the preset weight values of the physiological state factors based on risk classification rules to generate risk abnormality indexes, wherein the risk abnormality indexes are sorted in descending order according to the weight values; generating warning information based on the risk abnormality indexes.

4. The method of claim 1, wherein, The method further comprises the following steps: performing grayscale processing on the real-time eye images to generate initial color images; acquiring initial features based on the initial color images; performing normalization processing on the initial color images based on a residual neural network model to generate target color images with a preset size; performing global average pooling processing on the target color images with the preset size based on the residual neural network model to generate target features; generating target eye images based on the target features.

5. The method of claim 4, wherein, The method further comprises the following steps before performing normalization processing on the initial color images based on the residual neural network model to generate target color images with a preset size: performing slice processing on the target color images to sequentially generate a plurality of continuous slice images; The slice images are processed to generate multi-channel images; The multi-channel images are input into a residual neural network model for processing to generate the initial color image.

6. The method of claim 5, wherein, The processing of the real-time eye image based on the preset classification model to generate the target eye image further includes: The slice images are classified to generate target slice images of different categories; The target slice images are processed based on a preset classification threshold to generate inflammatory cells in an abnormal state; The inflammatory cells in the abnormal state are processed based on a preset rule to generate an inflammatory cell proportion value; The target eye image is generated based on the inflammatory cell proportion value.

7. A myopia early warning device for teenagers, characterized in that, The device includes: An acquisition module configured to acquire historical eye images of a plurality of users and eye habits of corresponding users, and acquire real-time eye images of the users under the eye habits based on the eye habits; The processing module is configured to respectively compare and process historical eye images of different users to generate a similarity result; divide the historical eye images of different users based on the similarity result to generate corresponding user groups, wherein the user groups include identification information representing the state of eye images of different users in the same group; process the real-time eye image based on a preset classification model to generate a target eye image, wherein the target eye image includes the influence degree of several physiological state factors on the real-time eye image; process the historical eye image and the target eye image based on an eye image calculation model to generate an eye image change rate of the user, wherein the eye image calculation model includes a calculation formula for calculating the eye image change rate, and the formula is: P=a Tn / (S1-S2); P represents the eye image change rate, a is a preset normalization factor, Tn is the duration of the second time period, (S1-S2) is the change value of the eye image in the first time period and the second time period, the first time period is the time interval for obtaining the historical eye image, and the second time period is the time interval for obtaining the real-time eye image; if the eye image change rate is greater than a preset threshold, process the eye image change rate based on a risk factor evaluation model to generate warning information, including obtaining a real-time parameter floating probability of the physiological state factor; if the floating probability is greater than a preset floating threshold, determine whether the physiological state factor corresponding to the floating probability is consistent based on the historical physiological state factor; if yes, increase the risk value of the physiological state factor; if no, obtain the number of times that the floating probability of the physiological state factor is greater than the preset floating threshold within a preset time, and if the number of times is greater than a preset number of times, increase the risk value of the physiological state factor; and generate warning information based on the physiological state factor.

8. An electronic device, comprising: The device includes: A first processor; and a memory configured to store executable instructions of the first processor; The first processor is configured to execute the executable instructions to implement the myopia early warning method of any one of claims 1-6.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the second processor to implement the myopia early warning method of any one of claims 1-6.

Images