AI intelligent correction method for preventing and controlling juvenile pseudomyopia based on big data

Abstract

The invention relates to the field of data processing, in particular to an AI intelligent correction method for prevention and control of juvenile pseudomyopia based on big data, the method is executed by a server, and the method comprises the following steps: obtaining vision information and initial body information of a target user, the vision information comprising various indexes for evaluating vision, the body information comprises motion data and eye data; the vision information and the initial body information are processed based on a preset scheme recommendation model, a correction scheme is obtained, and the correction scheme comprises an exercise scheme and an eye use scheme; in each unit period, collecting real-time body information and illness state information of the target user; on the basis of the illness state information, the real-time body information is compared with a correction scheme, whether first prompt information is generated or not is determined on the basis of a comparison result, and the first prompt information represents that eye using information and / or motion information of the target user are / is not matched with the correction scheme. According to the method, the correction scheme can be more accurately formulated for the pseudo-myopia patient.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to an AI intelligent correction method for the prevention and control of pseudomyopia in adolescents based on big data. Background Art

[0002] Nowadays, with the gradual popularization of electronic devices, more and more people are suffering from myopia problems due to excessive eye use and poor eye habits; moreover, the myopic population has gradually spread from young people to adolescents, and more and more parents have begun to pay attention to their children's eye problems.

[0003] Myopia is usually divided into false myopia and true myopia. False myopia is usually caused by excessive eye use or unhealthy eye habits. The vision of patients with false myopia can be restored to a certain extent by correcting their eye habits reasonably.

[0004] Therefore, how to more accurately formulate correction plans for patients with pseudomyopia and monitor the implementation effect of the correction plans is an urgent problem that needs to be solved. Summary of the invention

[0005] In order to solve the above problems, the present application provides an AI intelligent correction method for the prevention and control of pseudomyopia in adolescents based on big data.

[0006] In the first aspect, the present application provides an AI intelligent correction method for the prevention and control of pseudomyopia in adolescents based on big data, which adopts the following technical solutions: An AI intelligent correction method for adolescent pseudomyopia prevention and control based on big data, executed by a server, comprising: obtaining vision information and initial physical information of a target user, wherein the vision information comprises various indicators for evaluating vision, and the initial physical information is physical information for which a correction scheme has not been implemented, and the physical information comprises movement data and eye use data; Processing the vision information and the initial body information based on a preset solution recommendation model to obtain a correction solution, wherein the correction solution includes an exercise solution and an eye use solution; In each unit period, the real-time physical and medical information of the target user is collected; Based on the medical condition information, the real-time body information is compared with the correction plan, and based on the comparison result, it is determined whether to generate a first prompt information, wherein the first prompt information indicates that the target user's eye information and / or motion information does not match the correction plan.

[0007] By adopting the above technical solution, it is possible to collect the vision information and physical information of the target user, use the preset solution recommendation model, and train based on the historical data of multiple correction users, so as to scientifically evaluate the positive contribution of different types of exercise and eye habits to the correction effect, and ensure the accuracy and effectiveness of the correction solution; and then generate personalized correction solutions according to the specific situation of the target user, including exercise solutions and eye solutions, to improve the correction effect; and further, in each unit cycle, the physical information and medical information of the target user are collected in real time by monitoring equipment, and compared with the correction solution, so as to timely discover and remind the user whether the correction solution is implemented, thereby improving the implementation effect of the correction solution; at the same time, taking into account the possible medical conditions of the user, the exercise duration and other requirements in the correction solution are flexibly adjusted through the medical information input by the user, thereby increasing the flexibility and convenience of the correction solution; by generating the first prompt information, the user is promptly reminded to pay attention to the eye posture, control the eye duration, etc., to help the user develop good eye habits and exercise habits.

[0008] In a possible implementation, the motion data includes the motion type and the motion duration, the eye use data includes the eye use posture and the eye use duration, and the real-time physical information and medical condition information of the target user collected in each unit period includes: In each unit period, motion data collected by a first monitoring device is acquired, where the first monitoring device is used to monitor the motion state of the target user, and the motion data includes a motion type and a motion duration; In each unit period, a monitoring image collected by a second monitoring device is obtained, and eye use data, eye use posture and eye use duration are determined based on the monitoring image, where the second monitoring device is used to monitor the image of the target user's eye use status; In each unit period, the medical condition information input by the target user is obtained.

[0009] In a possible implementation, comparing the real-time body information with the correction plan based on the condition information, and determining whether to generate the first prompt information based on the comparison result, includes: If the disease condition information indicates that the target user has a disease of a preset type, comparing the eye use data with the eye use plan to determine whether to generate a first prompt information; If the medical condition information indicates that the target user does not have a preset type of disease, the eye use data is compared with the eye use plan, and the exercise data is compared with the exercise plan to determine whether to generate the first prompt information.

[0010] In one possible implementation, the method further includes: Obtaining fault information sent by any monitoring device, and determining that any monitoring device is a faulty device; Determine a device number of a standby device corresponding to the faulty device, and send a wake-up instruction to the standby device based on the device number; A second prompt message is sent to the normal device based on the device number, where the second prompt message is used to prompt the target user to switch to the backup device, and the normal device is a monitoring device that is not faulty.

[0011] In one possible implementation, the method further includes: At the end of each unit cycle, a feedback record is generated based on the physical information of the target user collected by each monitoring device and / or the faulty device; the feedback record is sent to a designated device, which is a coach-end device or a parent-end device; After receiving the reception information of the feedback record fed back by the designated device of the target user, a questionnaire is sent to the designated device every preset time period.

[0012] In a possible implementation, the training step of the solution recommendation model includes: Constructing a training set, wherein the training set includes a plurality of training samples, each of the training samples includes a historical correction scheme, a historical vision index, and a correction score, wherein the correction score represents the change of the vision index after adopting the historical correction scheme; Constructing a positive sample group and a negative sample group, wherein the positive sample group includes two different positive samples, and the negative sample group includes two different negative samples; Training is performed based on the initial feature extraction model of each positive sample group and each negative sample group to obtain a feature extraction model, and feature extraction is performed on each positive sample and each negative sample based on the feature extraction model to obtain a positive correction feature and a negative correction feature corresponding to each positive sample; The initial network model is trained based on the positive correction features and the negative correction features to obtain a solution recommendation model.

[0013] In a possible implementation, the constructing of the positive sample group and the negative sample group includes any of the following: Two positive samples with matching historical vision indicators are constructed into a positive sample group, and two negative samples with matching historical indicators are constructed into a negative sample group; Two positive samples with matching correction scores are constructed into a positive sample group, and two negative samples with matching correction scores are constructed into a negative sample group.

[0014] In one possible implementation, the method further includes: recording the receiving time each time supplementary information input by the user is received, wherein the supplementary information includes sleep information and nutritional intake information; Based on the receiving time, generating a statistical table distributed along a timeline for the supplementary information received each time; A mapping relationship is established between the supplementary information received each time and the real-time physical information and condition information of the target user collected.

[0015] In the second aspect, the present application provides an AI intelligent correction device for the prevention and control of pseudomyopia in adolescents based on big data, which adopts the following technical solutions: An AI intelligent correction device for adolescent pseudomyopia prevention and control based on big data, comprising: A first information acquisition module is used to acquire vision information and initial physical information of the target user, wherein the vision information includes various indicators for evaluating vision, and the initial physical information is physical information without implementing a correction plan, and the physical information includes movement data and eye use data; A correction scheme determination module, used for processing the vision information and the initial body information based on a preset scheme recommendation model to obtain a correction scheme, wherein the correction scheme includes an exercise scheme and an eye use scheme; The second information acquisition module is used to collect the real-time physical information and disease information of the target user in each unit period; A comparison module is used to compare the real-time body information with the correction plan based on the disease information, and determine whether to generate a first prompt information based on the comparison result, wherein the first prompt information indicates that the target user's eye information and / or movement information does not match the correction plan.

[0016] By adopting the above-mentioned technical solution, the device can collect the vision information and physical information of the target user, use the preset solution recommendation model, and train based on the historical data of multiple correction users. It can scientifically evaluate the positive contribution of different types of exercise and eye habits to the correction effect, and ensure the accuracy and effectiveness of the correction solution; and then it can generate personalized correction solutions according to the specific situation of the target user, including exercise solutions and eye solutions, to improve the correction effect; and further, in each unit cycle, the physical information and medical condition information of the target user are collected in real time through the monitoring device, and compared with the correction plan, so as to timely discover and remind the user whether the correction plan is implemented, thereby improving the implementation effect of the correction plan; at the same time, taking into account the possible disease conditions of the user, the exercise duration and other requirements in the correction plan are flexibly adjusted through the medical condition information input by the user, thereby increasing the flexibility and convenience of the correction plan; by generating the first prompt information, the user is promptly reminded to pay attention to the eye posture, control the eye duration, etc., to help the user develop good eye habits and exercise habits.

[0017] In a possible implementation, the motion data includes motion type and motion duration, the eye use data includes eye use posture and eye use duration, and the second information acquisition module is specifically used to: In each unit period, motion data collected by a first monitoring device is acquired, where the first monitoring device is used to monitor the motion state of the target user, and the motion data includes a motion type and a motion duration; In each unit period, a monitoring image collected by a second monitoring device is obtained, and eye use data, eye use posture and eye use duration are determined based on the monitoring image, where the second monitoring device is used to monitor the image of the target user's eye use status; In each unit period, the medical condition information input by the target user is obtained.

[0018] In a possible implementation, the comparison module is specifically used to: If the disease condition information indicates that the target user has a disease of a preset type, comparing the eye use data with the eye use plan to determine whether to generate a first prompt information; If the medical condition information indicates that the target user does not have a preset type of disease, the eye use data is compared with the eye use plan, and the exercise data is compared with the exercise plan to determine whether to generate the first prompt information.

[0019] In a possible implementation, the device further includes a fault processing module, and the fault processing module is specifically configured to: Obtaining fault information sent by any monitoring device, and determining that any monitoring device is a faulty device; Determine a device number of a standby device corresponding to the faulty device, and send a wake-up instruction to the standby device based on the device number; A second prompt message is sent to the normal device based on the device number, where the second prompt message is used to prompt the target user to switch to the backup device, and the normal device is a monitoring device that is not faulty.

[0020] In a possible implementation, the method further includes a feedback collection module, and the feedback collection module is specifically used to: At the end of each unit cycle, a feedback record is generated based on the physical information of the target user collected by each monitoring device and / or the faulty device; the feedback record is sent to a designated device, which is a coach-end device or a parent-end device; After receiving the reception information of the feedback record fed back by the designated device of the target user, a questionnaire is sent to the designated device every preset time period.

[0021] In a possible implementation, the device further includes a model training module, and the model training module is specifically used to: Constructing a training set, wherein the training set includes a plurality of training samples, each of the training samples includes a historical correction scheme, a historical vision index, and a correction score, wherein the correction score represents the change of the vision index after adopting the historical correction scheme; Constructing a positive sample group and a negative sample group, wherein the positive sample group includes two different positive samples, and the negative sample group includes two different negative samples; Training is performed based on the initial feature extraction model of each positive sample group and each negative sample group to obtain a feature extraction model, and feature extraction is performed on each positive sample and each negative sample based on the feature extraction model to obtain a positive correction feature and a negative correction feature corresponding to each positive sample; The initial network model is trained based on the positive correction features and the negative correction features to obtain a solution recommendation model.

[0022] In a possible implementation, the model training module is further used to: Two positive samples with matching historical vision indicators are constructed into a positive sample group, and two negative samples with matching historical indicators are constructed into a negative sample group; Two positive samples with matching correction scores are constructed into a positive sample group, and two negative samples with matching correction scores are constructed into a negative sample group.

[0023] In a possible implementation, the device further includes a statistical module, which is specifically configured to: recording the receiving time each time supplementary information input by the user is received, wherein the supplementary information includes sleep information and nutritional intake information; Based on the receiving time, generating a statistical table distributed along a timeline for the supplementary information received each time; A mapping relationship is established between the supplementary information received each time and the real-time physical information and condition information of the target user collected.

[0024] In the third aspect, the present application provides an AI intelligent correction system for adolescent pseudomyopia prevention and control based on big data, which adopts the following technical solutions: An AI intelligent correction system for adolescent pseudomyopia prevention and control based on big data, the system includes: The system includes a monitoring device, a server, a prompting device and: A monitoring device is used to collect vision information and initial physical information of the target user, and send the collected information to a server; the vision information includes various indicators for evaluating vision, the initial physical information is physical information without implementing a correction plan, and the physical information includes movement data and eye use data; The server is used to receive the vision information and initial physical information of the target user, and process the vision information and the initial physical information based on the preset solution recommendation model to obtain a correction solution, wherein the correction solution includes an exercise solution and an eye use solution; The server is further configured to collect real-time physical information and disease information of the target user in each unit period; compare the real-time physical information with the correction plan based on the disease information, and determine whether to generate first prompt information to be sent to the prompt device based on the comparison result, wherein the first prompt information indicates that the eye information and / or movement information of the target user does not match the correction plan; The prompting device performs a prompting in response to the first prompting information sent by the server.

[0025] Furthermore, the server is also used to execute any one of the AI ​​intelligent correction methods for the prevention and control of pseudomyopia in adolescents based on big data in the first aspect mentioned above.

[0026] In a fourth aspect, the present application provides a computer-readable storage medium, which adopts the following technical solution: A computer-readable storage medium, comprising: a computer program storing the above-mentioned AI intelligent correction method for the prevention and control of pseudomyopia in adolescents based on big data, which can be loaded by a processor and execute.

[0027] To sum up, by collecting the correction method of the present application, it is possible to collect the vision information and physical information of the target user, use the preset solution recommendation model, and train based on the historical data of multiple correction users, so as to scientifically evaluate the positive contribution of different types of exercise and eye habits to the correction effect, and ensure the accuracy and effectiveness of the correction solution; and then generate personalized correction solutions according to the specific situation of the target user, including exercise solutions and eye solutions, to improve the correction effect; and further, in each unit cycle, the physical information and medical condition information of the target user are collected in real time through the monitoring equipment, and compared with the correction plan, so as to timely discover and remind the user whether the correction plan is implemented, thereby improving the implementation effect of the correction plan; at the same time, taking into account the possible disease conditions of the user, the exercise duration and other requirements in the correction plan are flexibly adjusted through the medical condition information input by the user, thereby increasing the flexibility and convenience of the correction plan; by generating the first prompt information, the user is promptly reminded to pay attention to the eye posture, control the eye duration, etc., to help the user develop good eye habits and exercise habits. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 is a schematic diagram of an application scenario of the correction method in an embodiment of the present application; Figure 2 is a flowchart of the steps of the correction method in the embodiment of the present application; Figure 3 It is a schematic diagram of the DBN model training process in an embodiment of the present application; Figure 4 It is a structural diagram of the DBN model in the embodiment of the present application; Figure 5 This is a schematic diagram of the process steps of data collection in the embodiment of the present application; Figure 6 is a schematic diagram of the steps of fault monitoring of a monitoring device in an embodiment of the present application; Figure 7 Schematic diagram of the steps of the feedback questionnaire in the embodiment of the present application; Figure 8 It is a schematic diagram of the structure of the correction device in the embodiment of the present application; Fig. 9 It is a schematic diagram of the structure of the server in the embodiment of the present application. DETAILED DESCRIPTION

[0029] The following is combined with Figure 1 -Attached Fig. 9 This application is described in further detail.

[0030] After reading this specification, those skilled in the art may make non-creative modifications to this embodiment as needed, but such modifications are protected by patent law as long as they are within the scope of the claims of this application.

[0031] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0032] In addition, the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article, unless otherwise specified, generally means that the associated objects before and after are in an "or" relationship.

[0033] The present application embodiment provides an AI intelligent correction method for the prevention and control of pseudomyopia in adolescents based on big data, which can be used for example Figure 1 In the application scenario shown, the server can be configured with a solution recommendation model, and can generate a correction solution based on the acquired physical data and initial physical information of the target user. Furthermore, each monitoring device can obtain the real-time physical information and disease information of the target user in each unit period, and judge the real-time physical information and disease information of the target user in combination with the correction solution to determine whether to generate the first prompt information. If the first prompt information is generated, it is sent to the prompt device to remind the target user that the eye information and / or movement information does not match the correction solution.

[0034] The present application embodiment provides an AI intelligent correction method for adolescent pseudomyopia prevention and control based on big data, which is executed by a server, referring to Figure 2 The method specifically comprises steps 10 to 40, wherein: Step 10: Obtain the target user's vision information and initial physical information. The vision information includes various indicators for evaluating vision. The initial physical information is physical information without any correction plan. The physical information includes movement data and eye use data.

[0035] For the embodiments of the present application, the vision information and physical information of the target user are all information without correction. Among them, the indicators of vision information may specifically include but are not limited to the reading of the eye chart, diopter, and myopia reading, etc., and may also include other indicators that can characterize the degree of myopia. The motion data includes the user's motion habits, and the motion habits include the user's motion type, motion frequency, and motion duration. The eye information includes the eye environment, the average daily eye duration, and the eye posture, wherein the eye posture can be represented by the distance and angle between the eye and the visual object.

[0036] The target user's strength information and initial physical information need to be provided by the user, or collected by monitoring equipment. Among them, the monitoring equipment includes smart wearable devices for recording sports data, such as smart watches / bracelets; the monitoring equipment can also include image acquisition equipment, which can collect eye images of the target user when using his eyes, and the computer equipment obtains the eye images and determines the eye use time based on the time of the collected images; the computer equipment can also determine the target user's eye habits based on the collected eye images.

[0037] Step 20: Process the vision information and initial body information based on the preset solution recommendation model to obtain a correction solution, which includes an exercise solution and an eye use solution.

[0038] For the embodiment of the present application, the preset scheme recommendation model is trained based on the historical data of multiple correction users, wherein the correction users are users who can be diagnosed with pseudomyopia and whose vision information reaches the preset indicators after the correction scheme.

[0039] Different types of exercise have different positive contributions to the correction effect, and users with different vision information have different eye postures in the correction plan. The model trained with historical data of multiple correction users can learn the positive contribution of each type of exercise to pseudomyopia correction, and can also learn the eye habits corresponding to different vision information. The target user's vision information and initial physical information are input into the solution recommendation model, and the output of the solution recommendation model is the correction plan for the target user. The correction plan includes an exercise plan and an eye plan. The exercise plan includes the type of exercise and the average daily exercise duration, and the eye plan includes the average daily eye threshold and eye posture.

[0040] Step 30: In each unit period, real-time physical information and disease condition information of the target user is collected.

[0041] For the embodiment of the present application, after the correction scheme for the target user is determined, the correction scheme usually depends on the target user to execute it autonomously. In order to improve the execution effect of the correction scheme, the embodiment of the present application collects the execution information of the correction scheme by the target user by setting a monitoring device. Physical information includes movement data and eye data.

[0042] Specifically, the target user's medical condition information needs to be input by the target user. The unit period can be set by the user, for example, 6 hours or 24 hours. The specific duration of the unit period is not specifically limited in the embodiment of the present application. In the embodiment of the present application, the unit period is taken as one hour for example to illustrate the feasibility of the logic of the present application scheme.

[0043] Step 40: Based on the disease condition information, the real-time body information is compared with the correction plan, and based on the comparison result, it is determined whether to generate a first prompt information, the first prompt information indicating that the target user's eye information and / or motion information does not match the correction plan.

[0044] For the embodiment of the present application, within a unit period, if there is a situation where the target user does not meet the standard of the correction plan, such as the exercise time of the day is less than the exercise time in the correction plan, and / or the eye posture is not standard, and / or the eye time exceeds the eye time threshold in the correction plan, then the target user can be prompted through the corresponding prompt device. However, considering the actual situation, the target user may have a disease that makes it impossible to complete the correction plan for the day, so the target user may not be reminded. Therefore, the judgment is made based on the disease information input by the target user. When the disease information input by the target user indicates that the target user has a preset type of disease, the judgment standard within the current unit period can be lowered, that is, the exercise time in the correction plan for the day is reduced, but the eye plan is not modified. For example, the type of exercise recommended by the correction plan is running, and the exercise time is 1 hour. In the current unit period, the exercise time can be set to a temporary value, such as 0; the temporary value can also be modified by the user.

[0045] As long as the user has eye use behaviors within the unit period, such as reading, using electronic devices, etc., it is necessary to count the target user's eye use time, and use the glasses image to determine the target user's eye use data and compare it with the eye use plan, and then determine whether to generate the first prompt information.

[0046] In an embodiment of the present application, by collecting the vision information and physical information of the target user, using a preset solution recommendation model, and training based on the historical data of multiple correction users, it is possible to scientifically evaluate the positive contribution of different types of exercise and eye habits to the correction effect, thereby ensuring the accuracy and effectiveness of the correction solution; and then, it is possible to generate a personalized correction solution, including an exercise solution and an eye solution, for the specific situation of the target user to improve the correction effect; and further, within each unit cycle, the physical information and medical condition information of the target user are collected in real time by monitoring equipment, and compared with the correction solution, so that it is possible to timely discover and remind the user whether the correction solution is being implemented, thereby improving the implementation effect of the correction solution; at the same time, taking into account the possible medical conditions of the user, the exercise duration and other requirements in the correction solution are flexibly adjusted through the medical condition information input by the user, thereby increasing the flexibility and convenience of the correction solution; by generating the first prompt information, the user is promptly reminded to pay attention to the eye posture, control the eye duration, etc., to help the user develop good eye habits and exercise habits.

[0047] In the embodiments of this application, refer to Figure 3 The step of training a solution recommendation model may specifically include steps 01 to 04, wherein: Step 01, constructing a training set, the training set includes multiple training samples, each training sample includes a historical correction scheme, a historical vision index and a correction score, wherein the correction score represents the change of the vision index after adopting the historical correction scheme; Step 02: Construct a positive sample group and a negative sample group, wherein the positive sample group includes two different positive samples, and the negative sample group includes two different negative samples.

[0048] Specifically, a correction user may correspond to multiple groups of historical data, and each group of correction data corresponds to a correction cycle. The historical data of each correction user includes historical correction schemes, historical vision indicators, and correction scores. Among them, the correction score is determined based on the changes in each indicator in the current vision information compared with the previous vision information. For example, a weight is set for each vision indicator, and each vision indicator is multiplied by the corresponding weight compared with the previous time, and then the sum is obtained to obtain the correction score. The correction score represents the change in vision indicators after adopting the historical correction scheme. For example, the larger the correction score, the greater the change in vision indicators, that is, the better the correction effect; the smaller the correction score, the smaller the change in vision indicators, that is, the worse the correction effect. Specifically, if the correction score is positive, it represents that the correction effect is positive; if the correction score is negative, it represents that the correction effect is reverse. Among them, positive samples are samples with positive correction scores, and negative samples are samples with negative correction scores.

[0049] Step 03: Train the initial feature extraction model based on each positive sample group and each negative sample group to obtain a feature extraction model, and extract features from each positive sample and each negative sample based on the feature extraction model to obtain positive correction features and negative correction features corresponding to each positive sample; Step 04: Train the initial network model based on the positive correction features and the negative correction features to obtain a solution recommendation model.

[0050] Specifically, in an embodiment of the present application, a DBN model is used as the initial network model for training. DBN is a neural network structure composed of multiple RBMs stacked with a classification layer or regression layer superimposed on the top. This structure enables DBN to have the advantages of both unsupervised learning and supervised learning, so that it can effectively process complex data sets. DBN uses a greedy algorithm to train RBM layer by layer. This layer-by-layer training method makes DBN more stable and efficient during unsupervised training. Each layer of RBM is optimized to retain the information of the previous layer and extract higher-level features. In this way, DBN can gradually abstract the inherent structure and laws of the data, thereby generating a low-dimensional feature representation with more information. For multi-dimensional unlabeled data samples containing redundant and noisy information, the DBN model can remove these redundant information by learning layer by layer, and then extract key features.

[0051] In one embodiment, referring to Figure 4The initial network model includes two cascaded RBM (Restricted Boltzmann Machine) networks, namely RBM1 and RBM2, and an encoding layer AE (AutoEncoder) network.

[0052] The RBM1 layer is the input layer, which is used to receive the input positive sample group and negative sample group and perform preliminary feature extraction. The output result of the RBM1 layer is the input data of the RBM2 layer; the RBM2 layer is used to further extract features so that the deep features of the sample data can be gradually learned. Since the function of the DBN model is unsupervised feature extraction, an AE encoding layer is superimposed after the RBM2 network layer to achieve unsupervised feature extraction, thereby adjusting the weights of the entire network. The AE network includes an encoder and a decoder; the encoder compresses the input data into a low-dimensional feature vector, which is a process of feature extraction of the input data. The hidden layer of the encoder can compress the input data, which contains the important features of the input data. Through training, the encoder can learn the potential representation of the input data. The decoder decompresses the low-dimensional feature vector obtained by the encoder into the same dimension as the original input data, thereby realizing data reconstruction. The decoder can restore the original input data, so that the difference between the reconstructed data and the original data is minimized. Through training, the decoder can learn how to restore the information of the original data from the low-dimensional feature vector.

[0053] Different types of movement have different positive contributions to the correction effect, and users with different vision information have different eye postures in their correction plans. By training the DBN model, a trained DBN model can be obtained. The trained DBN model can learn the positive contribution of each type of movement to pseudomyopia correction, and can also learn the eye habits corresponding to different vision information.

[0054] Furthermore, hidden features are extracted from the trained DBN model, wherein the hidden features characterize the positive contribution of each type of exercise to pseudomyopia correction, as well as the eye habits corresponding to the range of each indicator in the vision information. Specifically, the hidden features usually exist in the previous network layer of the output layer. In the embodiment of the present application, the hidden features are located in the RBM2 network layer.

[0055] The solution recommendation model is configured to match the latent features based on the initial physical information and vision information of the target user, and then determine the correction solution that suits the target user, that is, the eye use solution and the exercise solution.

[0056] Further, in one embodiment, referring to Figure 5, is a schematic diagram of the data collection process, step 30 may specifically include steps 31 to 33, wherein: Step 31: In each unit period, motion data collected by a first monitoring device is obtained. The first monitoring device is used to monitor the motion state of the target user. The motion data includes the motion type and the motion duration.

[0057] Step 32: in each unit period, obtain a monitoring image collected by the second monitoring device, and determine eye usage data, eye usage posture and eye usage duration based on the monitoring image, and the second monitoring device is used to monitor the image of the eye usage status of the target user; Step 33: In each unit period, obtain the medical condition information input by the target user.

[0058] In an embodiment of the present application, the monitoring device includes a first monitoring device and a second monitoring device, wherein the first monitoring device is used to monitor the motion status of the target user, including the type of motion and the duration of motion; the second monitoring device is used to collect monitoring images of the target user when using his eyes.

[0059] In one embodiment, the first monitoring device may be a smart wearable device capable of recording motion data, such as a smart watch / bracelet; the second monitoring device may be an image acquisition device capable of acquiring images or videos of the target user's eyes when using their eyes. The server acquires the monitoring images and determines the duration of eye use based on the time or video of the acquired images; the computer device may also determine the target user's eye use habits based on the acquired eye use images.

[0060] Further, in one of the embodiments, step 40 may specifically include: if the medical condition information indicates that the target user has a preset type of disease, the eye use data is compared with the eye use plan to determine whether to generate a first prompt information; if the medical condition information indicates that the target user does not have a preset type of disease, the eye use data is compared with the eye use plan, and the motion data is compared with the motion plan to determine whether to generate a first prompt information.

[0061] Among them, the preset type of disease can be pre-set, and this is not specifically limited in the embodiment of the present application, as long as it can meet the actual physical needs of the target user. Specifically, when the disease information input by the target user indicates that the target user has a disease of a preset type, in the current unit period, only the target user's eye data is compared with the medication plan, and the first prompt information is generated when the target user's eye use time exceeds the eye use time threshold in the eye use plan, and / or the target user's eye use posture does not conform to the eye use posture in the eye use plan. When the disease information input by the target user indicates that the target user does not have a disease of a preset type, it is necessary to compare the target user's motion information and the motion plan in the current unit period, and compare the eye use data with the eye use plan, and generate the first prompt information when the target user's motion time is less than the eye use time in the motion plan, and / or the target user's eye use time exceeds the eye use time threshold in the eye use plan, and / or the target user's eye use posture does not conform to the eye use posture in the eye use plan.

[0062] Further, in one embodiment, referring to Figure 6 , which is a step of monitoring the fault of the monitoring device, the correction method also includes steps 51 to 53, wherein: Step 51: Obtain fault information sent by any monitoring device, and determine that any monitoring device is a faulty device; Step 52: Determine the device number of the backup device corresponding to the faulty device, and send a wake-up instruction to the backup device based on the device number.

[0063] Specifically, any monitoring device (the first monitoring device or the second monitoring device) can communicate with the server through the network, and the monitoring device can execute a fault diagnosis program based on the operation log to determine whether a fault occurs. When a monitoring device fails, the monitoring device can send the device number to the server. The server searches the mapping table based on the device number of the failed device to determine the device number of the backup device, and sends a wake-up instruction to the device number of the backup device so that the backup device is awakened to monitor the physical information of the target user. Among them, the device number can be the MAC address of the device, or other identifier that can characterize the identity of the monitoring device.

[0064] Step 53: Send a second prompt message to the normal device based on the device number. The second prompt message is used to prompt the target user to switch to the backup device. The normal device is a monitoring device that is not faulty.

[0065] Specifically, a monitoring device that has not failed is a normal device. When the server sends a wake-up command to the backup device, it also sends a second prompt message to the target user's normal device to encourage the target user to switch the failed device to the backup device. The second prompt message can be in text format or in sound format. Furthermore, any monitoring device is provided with a prompt light, and the second prompt message can also be used to make the prompt light on the normal device flash, so as to achieve a prompt effect for the target user.

[0066] By automatically monitoring the fault status of the device and sending a wake-up instruction to the backup device based on the device number of the faulty device, the backup device can be awakened in advance and put into a working state; and the server can send a second prompt message to the normal device to remind the target user to switch to the backup device. This method can reduce manual intervention, so as to obtain more comprehensive and timely physical information of the target user, and improve the integrity of the data.

[0067] Further, in one embodiment, referring to Figure 7 The correction method may further include a step of conducting a questionnaire survey, specifically including step 61 and step 62, wherein: Step 61: At the end of each unit cycle, a feedback record is generated based on the physical information of the target user collected by each monitoring device and / or faulty device; the feedback record is sent to a designated device, which is a coach-end device or a parent-end device.

[0068] Step 62: After receiving the reception information of the feedback record fed back by the designated device of the target user, send a questionnaire to the designated device once every preset time period.

[0069] Specifically, the designated device of the target user can be a computer device or an address, such as an email address. At the end of each unit period, the physical information of the target user collected by each first monitoring device and each second monitoring device is analyzed to obtain the user's exercise type and exercise duration in each unit period, as well as information on eye use duration and eye posture, and generate a feedback record with this information in a preset format. The feedback record is sent to the designated device of the target user. After receiving the feedback record, the target user can send a receipt message to the server. After receiving the receipt message fed back by the target user's execution device, the server sends a questionnaire file or a link address that can point to the questionnaire to the target user's execution device once every preset time period to prompt the target user to fill in the questionnaire.

[0070] Furthermore, the method of the embodiment of the present application also includes: recording the time of receipt each time supplementary information input by the user is received, the supplementary information including sleep information and nutritional intake information; based on the time of receipt, generating a statistical table distributed along a timeline for the supplementary information received each time; establishing a mapping relationship between the supplementary information received each time and the real-time physical information and medical condition information collected from the target user.

[0071] Specifically, since nutritional intake and sleep duration can also affect human vision, when considering the user's real-time physical information and exercise information, the user's nutritional intake information and sleep information in each unit cycle should also be considered, among which the sleep information includes the average daily sleep duration; the nutritional intake information includes the daily intake of nutrients.

[0072] By associating the supplementary information with the user's real-time physical information and medical information, it is easier for doctors or coaches to better analyze the user's vision condition, thereby facilitating the doctor or coach to make targeted adjustments to the correction plan generated by the recommendation model.

[0073] The above-mentioned embodiment introduces an AI intelligent correction method for the prevention and control of pseudomyopia in adolescents based on big data from the perspective of method flow. The following embodiment introduces an AI intelligent correction device for the prevention and control of pseudomyopia in adolescents based on big data from the perspective of a virtual module or a virtual unit. Please see the following embodiment for details.

[0074] The present application embodiment provides an AI intelligent correction device for the prevention and control of pseudomyopia in adolescents based on big data, such as Figure 8 As shown, the device may specifically include a first information acquisition module 801, a correction scheme determination module 802, a second information acquisition module 803 and a comparison module 804, wherein: The first information acquisition module 801 is used to acquire the vision information and initial body information of the target user, where the vision information includes various indicators for evaluating vision, and the initial body information is the body information before the correction plan is implemented, and the body information includes movement data and eye use data; A correction scheme determination module 802 is used to process the vision information and the initial body information based on a preset scheme recommendation model to obtain a correction scheme, wherein the correction scheme includes a movement scheme and an eye use scheme; The second information acquisition module 803 is used to collect the real-time physical information and disease information of the target user in each unit period; The comparison module 804 is used to compare the real-time body information with the correction plan based on the disease information, and determine whether to generate a first prompt information based on the comparison result. The first prompt information indicates that the target user's eye information and / or movement information does not match the correction plan.

[0075] By adopting the above-mentioned technical solution, the device can collect the vision information and physical information of the target user, use the preset solution recommendation model, and train based on the historical data of multiple correction users. It can scientifically evaluate the positive contribution of different types of exercise and eye habits to the correction effect, and ensure the accuracy and effectiveness of the correction solution; and then it can generate personalized correction solutions according to the specific situation of the target user, including exercise solutions and eye solutions, to improve the correction effect; and further, in each unit cycle, the physical information and medical condition information of the target user are collected in real time through the monitoring device, and compared with the correction plan, so as to timely discover and remind the user whether the correction plan is implemented, thereby improving the implementation effect of the correction plan; at the same time, taking into account the possible disease conditions of the user, the exercise duration and other requirements in the correction plan are flexibly adjusted through the medical condition information input by the user, thereby increasing the flexibility and convenience of the correction plan; by generating the first prompt information, the user is promptly reminded to pay attention to the eye posture, control the eye duration, etc., to help the user develop good eye habits and exercise habits.

[0076] In a possible implementation, the motion data includes the motion type and the motion duration, the eye use data includes the eye use posture and the eye use duration, and the second information acquisition module is specifically used to: In each unit period, motion data collected by a first monitoring device is obtained, where the first monitoring device is used to monitor the motion state of the target user, and the motion data includes the motion type and the motion duration; In each unit period, a monitoring image collected by the second monitoring device is obtained, and eye use data, eye use posture and eye use duration are determined based on the monitoring image, and the second monitoring device is used to monitor the image of the eye use state of the target user; In each unit period, the medical condition information input by the target user is obtained.

[0077] In one possible implementation, the comparison module 804 is specifically configured to: If the disease condition information indicates that the target user has a preset type of disease, the eye use data is compared with the eye use plan to determine whether to generate the first prompt information; If the disease condition information indicates that the target user does not have a preset type of disease, the eye use data is compared with the eye use plan, and the exercise data is compared with the exercise plan to determine whether to generate the first prompt information.

[0078] In a possible implementation, the device further includes a fault processing module, which is specifically configured to: Obtaining fault information sent by any monitoring device, and determining any monitoring device as a faulty device; Determine the device number of the backup device corresponding to the failed device, and send a wake-up instruction to the backup device based on the device number; A second prompt message is sent to the normal device based on the device number, where the second prompt message is used to prompt the target user to switch to the backup device, and the normal device is a monitoring device that is not faulty.

[0079] In a possible implementation manner, the method further includes a feedback collection module, and the feedback collection module is specifically used to: At the end of each unit cycle, a feedback record is generated based on the physical information of the target user collected by each monitoring device and / or the faulty device; the feedback record is sent to a designated device, which is a coach-end device or a parent-end device; After receiving the receiving information for the feedback record fed back by the designated device of the target user, a questionnaire is sent to the designated device once every preset time period.

[0080] In a possible implementation, the device further includes a model training module, and the model training module is specifically used to: Constructing a training set, the training set includes multiple training samples, each training sample includes a historical correction scheme, a historical vision index and a correction score, wherein the correction score represents the change of the vision index after adopting the historical correction scheme; Construct a positive sample group and a negative sample group, wherein the positive sample group includes two different positive samples, and the negative sample group includes two different negative samples; Training is performed based on the initial feature extraction model of each positive sample group and each negative sample group to obtain a feature extraction model, and feature extraction is performed on each positive sample and each negative sample based on the feature extraction model to obtain positive correction features and negative correction features corresponding to each positive sample; The initial network model is trained based on the positive correction features and the negative correction features to obtain a solution recommendation model.

[0081] In one possible implementation, the model training module is also used to: Two positive samples with matching historical vision indicators are constructed into a positive sample group, and two negative samples with matching historical indicators are constructed into a negative sample group; Two positive samples with matching correction scores are constructed into a positive sample group, and two negative samples with matching correction scores are constructed into a negative sample group. In a possible implementation, the device further includes a statistical module, which is specifically used to: recording the receiving time each time supplementary information input by the user is received, wherein the supplementary information includes sleep information and nutritional intake information; Based on the receiving time, generating a statistical table distributed along a timeline for the supplementary information received each time; A mapping relationship is established between the supplementary information received each time and the real-time physical information and condition information of the target user collected.

[0082] In the embodiment of the present application, an AI intelligent correction system for the prevention and control of pseudomyopia in adolescents based on big data is provided, and the system includes a monitoring device, a server, and a prompt device, wherein: A monitoring device is used to collect vision information and initial physical information of the target user, and send the collected information to a server; the vision information includes various indicators for evaluating vision, the initial physical information is physical information without implementing a correction plan, and the physical information includes movement data and eye use data; The server is used to receive the vision information and initial physical information of the target user, and process the vision information and the initial physical information based on the preset solution recommendation model to obtain a correction solution, wherein the correction solution includes an exercise solution and an eye use solution; The server is further configured to collect real-time physical information and disease information of the target user in each unit period; compare the real-time physical information with the correction plan based on the disease information, and determine whether to generate first prompt information to be sent to the prompt device based on the comparison result, wherein the first prompt information indicates that the eye information and / or movement information of the target user does not match the correction plan; The prompting device performs corresponding prompting in response to the first prompting information and / or the second prompting information sent by the server; the second prompting information is used to prompt the target user to switch to the backup device, and the normal device is a monitoring device that is not faulty.

[0083] The monitoring device includes a first monitoring device, a second monitoring device and a backup device; wherein the first monitoring device is arranged at the user end, and the first monitoring device is used to monitor the motion state of the target user, and the motion data includes the motion type and the motion duration; the first monitoring device sends the collected motion data to the server; The second monitoring device is arranged at the user end, and is used to monitor the image of the target user's eye status, and the second monitoring device sends the collected image to the server.

[0084] Furthermore, each first monitoring device and each second monitoring device respectively corresponds to a same standby device, and the standby device is associated with the device code of the corresponding monitoring device in a one-to-one matching manner. The standby device can communicate with the server, and the standby device is awakened in response to the awakening instruction of the server.

[0085] Furthermore, the server is also used to execute any method step in the above-mentioned AI intelligent correction method embodiment for the prevention and control of pseudomyopia in adolescents based on big data.

[0086] The schematic structure of the server 900 is as follows Fig. 9 As shown, the server 900 includes: a processor 901 and a memory 903. The processor 901 and the memory 903 are connected, such as through a bus 902. Optionally, the server 900 may also include a transceiver 904. It should be noted that in actual applications, the transceiver 904 is not limited to one, and the structure of the server 900 does not constitute a limitation on the embodiments of the present application.

[0087] The processor 901 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It may implement or execute various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. The processor 901 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, and the like.

[0088] The bus 902 may include a path to transmit information between the above components. The bus 902 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus 902 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Fig. 9 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0089] The memory 903 may be a ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, a RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, optical disk storage (including compressed optical disk, laser disk, optical disk, digital versatile disk, Blu-ray disk, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto.

[0090] The memory 903 is used to store the application code for executing the solution of the present application, and the execution is controlled by the processor 901. The processor 901 is used to execute the application code stored in the memory 903 to implement the contents shown in the above method embodiment. Fig. 9 The server shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0091] An embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer-readable storage medium is run on a computer, the computer can execute the corresponding content in the aforementioned method embodiment.

[0092] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0093] The above are only some implementation methods of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. An AI intelligent correction method for the prevention and control of pseudomyopia in adolescents based on big data, characterized in that: Executed by a server, the method includes: Obtaining vision information and initial physical information of the target user, wherein the vision information includes various indicators for evaluating vision, and the initial physical information is physical information before a correction plan is implemented, and the physical information includes movement data and eye use data; Processing the vision information and the initial body information based on a preset solution recommendation model to obtain a correction solution, wherein the correction solution includes an exercise solution and an eye use solution; In each unit period, the real-time physical and medical information of the target user is collected; Based on the medical condition information, the real-time body information is compared with the correction plan, and based on the comparison result, it is determined whether to generate a first prompt information, wherein the first prompt information indicates that the target user's eye information and / or motion information does not match the correction plan.

2. The method according to claim 1, characterized in that The exercise data includes exercise type and exercise duration, the eye use data includes eye use posture and eye use duration, and the real-time physical information and disease information of the target user collected in each unit cycle includes: In each unit period, motion data collected by a first monitoring device is acquired, where the first monitoring device is used to monitor the motion state of the target user, and the motion data includes a motion type and a motion duration; In each unit period, a monitoring image collected by a second monitoring device is obtained, and eye use data, eye use posture and eye use duration are determined based on the monitoring image, where the second monitoring device is used to monitor the image of the target user's eye use status; In each unit period, the medical condition information input by the target user is obtained.

3. The method according to claim 2, characterized in that The comparing the real-time body information with the correction plan based on the condition information, and determining whether to generate the first prompt information based on the comparison result, includes: If the disease condition information indicates that the target user has a disease of a preset type, comparing the eye use data with the eye use plan to determine whether to generate a first prompt information; If the medical condition information indicates that the target user does not have a preset type of disease, the eye use data is compared with the eye use plan, and the exercise data is compared with the exercise plan to determine whether to generate the first prompt information.

4. The method according to claim 2, characterized in that: The method further comprises: Obtaining fault information sent by any monitoring device, and determining that any monitoring device is a faulty device; Determine a device number of a standby device corresponding to the faulty device, and send a wake-up instruction to the standby device based on the device number; A second prompt message is sent to the normal device based on the device number, where the second prompt message is used to prompt the target user to switch to the backup device, and the normal device is a monitoring device that is not faulty.

5. The method according to claim 4, characterized in that The method further comprises: At the end of each unit cycle, a feedback record is generated based on the physical information of the target user collected by each monitoring device and / or the faulty device; the feedback record is sent to a designated device, which is a coach-end device or a parent-end device; After receiving the reception information of the feedback record fed back by the designated device of the target user, a questionnaire is sent to the designated device every preset time period.

6. The method according to claim 1, characterized in that The training steps of the solution recommendation model include: Constructing a training set, wherein the training set includes a plurality of training samples, each of the training samples includes a historical correction scheme, a historical vision index, and a correction score, wherein the correction score represents the change of the vision index after adopting the historical correction scheme; Constructing a positive sample group and a negative sample group, wherein the positive sample group includes two different positive samples, and the negative sample group includes two different negative samples; Training is performed based on the initial feature extraction model of each positive sample group and each negative sample group to obtain a feature extraction model, and feature extraction is performed on each positive sample and each negative sample based on the feature extraction model to obtain a positive correction feature and a negative correction feature corresponding to each positive sample; The initial network model is trained based on the positive correction features and the negative correction features to obtain a solution recommendation model.

7. The method according to claim 6, characterized in that The constructing of the positive sample group and the negative sample group includes any of the following: Two positive samples with matching historical vision indicators are constructed into a positive sample group, and two negative samples with matching historical indicators are constructed into a negative sample group; Two positive samples with matching correction scores are constructed into a positive sample group, and two negative samples with matching correction scores are constructed into a negative sample group.

8. The method according to any one of claims 1 to 7, characterized in that: The method further comprises: recording the receiving time each time supplementary information input by the user is received, wherein the supplementary information includes sleep information and nutritional intake information; Based on the receiving time, generating a statistical table distributed along a timeline for the supplementary information received each time; A mapping relationship is established between the supplementary information received each time and the real-time physical information and condition information of the target user collected.

9. An AI intelligent correction system for adolescent pseudomyopia prevention and control based on big data, characterized in that: The system includes a monitoring device, a server, a prompting device and: A monitoring device is used to collect vision information and initial physical information of the target user, and send the collected information to a server; the vision information includes various indicators for evaluating vision, the initial physical information is physical information without implementing a correction plan, and the physical information includes movement data and eye use data; The server is used to receive the vision information and initial physical information of the target user, and process the vision information and the initial physical information based on the preset solution recommendation model to obtain a correction solution, wherein the correction solution includes an exercise solution and an eye use solution; The server is further configured to collect real-time physical information and disease information of the target user in each unit period; compare the real-time physical information with the correction plan based on the disease information, and determine whether to generate first prompt information to be sent to the prompt device based on the comparison result, wherein the first prompt information indicates that the eye information and / or movement information of the target user does not match the correction plan; The prompting device performs a prompting in response to the first prompting information sent by the server.

10. An AI intelligent correction device for the prevention and control of pseudomyopia in adolescents based on big data, characterized in that: include: A first information acquisition module is used to acquire vision information and initial physical information of the target user, wherein the vision information includes various indicators for evaluating vision, and the initial physical information is physical information without implementing a correction plan, and the physical information includes movement data and eye use data; A correction scheme determination module, used for processing the vision information and the initial body information based on a preset scheme recommendation model to obtain a correction scheme, wherein the correction scheme includes an exercise scheme and an eye use scheme; The second information acquisition module is used to collect the real-time physical information and disease information of the target user in each unit period; A comparison module is used to compare the real-time body information with the correction plan based on the disease information, and determine whether to generate a first prompt information based on the comparison result, wherein the first prompt information indicates that the target user's eye information and / or movement information does not match the correction plan.

Images