A Myopia Intervention Compliance Management System Based on Eye Use Behavior Monitoring
By using an eye behavior monitoring system to dynamically adjust myopia intervention strategies, the problem of existing systems being unable to identify short-term changes in lifestyle rhythms is solved, thus improving user compliance and the effectiveness of intervention measures.
Patent Information
- Application Number
- CN202611632677.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-21
- Publication Date
- 2026-11-17
AI Technical Summary
Existing myopia intervention management systems cannot identify the impact of short-term changes in lifestyle on adherence and lack dynamic adaptation capabilities, leading to decreased user adherence and reminder fatigue.
The myopia intervention compliance management system based on eye use behavior monitoring employs data collection, time period segmentation, scoring mapping, and correlation analysis modules to dynamically adjust intervention prompt strategies, quantify the correlation between eye use behavior and intervention response, identify key time periods, and make differentiated adjustments.
This improved users' acceptance and willingness to implement intervention measures, enhanced overall compliance, and ensured the reliability of calculation results and the efficiency of resource utilization.
Smart Images

Figure CN123120330A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of medical and health management technology, specifically to a myopia intervention compliance management system based on eye behavior monitoring. Background Technology
[0002] Myopia has become a major problem affecting the visual health of teenagers. The effectiveness of existing myopia intervention methods (such as reminders about viewing distance, control of viewing time, and correction of viewing posture) depends heavily on users' long-term adherence. Therefore, how to monitor and improve users' adherence to intervention measures is a core technical issue in the field of myopia intervention and management.
[0003] Currently, mainstream myopia intervention management systems typically employ a linear "monitoring-recording-feedback" model, using the same analysis and intervention strategies across all time periods. However, such systems suffer from the following technical shortcomings in practical applications: First, users' eye-use behavior and intervention responses are not stable but are significantly affected by various short-term changes in lifestyle. Existing systems struggle to identify the impact of these short-term changes on compliance and to quantify the degree of compliance change across different time periods. Second, if user compliance declines, existing systems often simply increase the frequency of reminders. This approach easily leads to user fatigue and cannot be finely adjusted based on the actual differences in user performance across different time periods. Summary of the Invention
[0004] To address the technical problems of existing myopia intervention management systems failing to recognize the impact of short-term lifestyle changes on adherence and lacking dynamic adaptability in intervention strategies, this invention provides a myopia intervention adherence management system based on eye use behavior monitoring. The specific technical solution adopted is as follows: This invention proposes a myopia intervention compliance management system based on eye use behavior monitoring. The system includes: The data acquisition module is used to collect users' eye use status parameters, intervention response parameters, and environmental parameters; The time period segmentation module is used to set an analysis period and divide the analysis period into multiple equal-length time periods as behavior segments. The scoring mapping module is used to map eye use status parameters to eye use scores according to a preset first scoring rule; and to map intervention response parameters to intervention response scores according to a preset second scoring rule. The correlation analysis module is used to divide each behavior segment into time units; filter out valid time units with intervention response scores; determine the segment correlation degree based on the correlation between eye use scores and response scores of all valid time units within the behavior segment; determine the historical correlation degree based on the historical correlation between eye use scores and response scores of all valid time units within a preset number of behavior segments before the current behavior segment; and determine the interference weight of the current behavior segment based on the difference between the segment correlation degree and the historical correlation degree. The intervention adjustment module is used to dynamically adjust the intervention prompt strategy based on the interference weight and environmental parameters.
[0005] Furthermore, the eye use status parameter acquisition process includes: By using smart monitoring glasses to detect users’ eye-use behavior, each continuous eye-use process is marked as an eye-use event, and each eye-use event has a start time, end time and duration. Within each eye-use event, eye-use status parameters are collected according to a preset sampling frequency; among them, eye-use status parameters include at least one of the following: eye-use distance, eye-use posture, blinking frequency, and continuous eye-use duration.
[0006] Furthermore, the intervention response parameter acquisition process includes: When the system issues an intervention prompt, mark the start of an intervention event; if it detects that the user has responded to the intervention prompt or the waiting time exceeds the preset response time threshold, mark the end of the intervention event. Within each intervention event, intervention response parameters are collected; these parameters include at least one of the following: intervention response determination result, intervention response duration, and intervention response completion rate.
[0007] Furthermore, the effective time unit filtering process includes: For each behavior segment, the behavior segment is divided into multiple consecutive time units according to a preset time interval; Determine whether an intervention response score exists in each time unit; mark time units with intervention response scores as valid time units; mark time units without intervention response scores as invalid time units; invalid time units will not participate in the determination of segment correlation and historical correlation.
[0008] Furthermore, each eye use score is obtained by mapping each eye use state parameter to a score value within the same preset score range according to its corresponding preset first scoring rule; The system is also used for: Obtain the eye use scores for all sampling points of all eye use events within each valid time unit; For each sampling point, the eye use scores of that sampling point are weighted and summed according to a preset first weight set to obtain the instantaneous eye use score of that sampling point; for each eye use event, the arithmetic mean of the instantaneous eye use scores of all sampling points within that eye use event is calculated as the event eye use score of that eye use event. For each valid time unit, the arithmetic mean of the event eye use scores of all eye use events within that valid time unit is calculated, and this is used as the unit eye use score for that valid time unit.
[0009] Furthermore, the scoring of each intervention response involves mapping each intervention response parameter to a score value within the same preset score range according to its corresponding preset second scoring rule. The preset score range is the same as the preset score range of the eye use score. The system is also used for: For each valid time unit, obtain the intervention response scores for each intervention event within that valid time unit; For each intervention event, the scores of each intervention response within that intervention event are weighted and summed according to a preset second weight set to obtain the instantaneous intervention score for that intervention event; Calculate the arithmetic mean of the instantaneous intervention scores of all intervention events within the effective time unit, and use it as the unit intervention score for the effective time unit.
[0010] Furthermore, the process of determining the correlation degree of the segments includes: The unit intervention score for each effective time unit is obtained by analogy with the aforementioned method of determining the unit intervention score for effective time units; For each behavioral segment, extract the unit eye use scores of all valid time units within the behavioral segment, arrange them in chronological order to form an eye use score sequence; extract the unit intervention scores of the same group of valid time units within the same behavioral segment, arrange them in the same chronological order to form a response score sequence; Calculate the Pearson correlation coefficient between the eye use rating sequence and the response rating sequence as the segment correlation degree of the behavioral segment.
[0011] Furthermore, the process of determining the historical correlation degree includes: The unit intervention score for each effective time unit is obtained by analogy with the aforementioned method of determining the unit intervention score for effective time units; For the current behavior segment, extract the unit eye use scores of all valid time units within a preset number of previous behavior segments, arrange them in chronological order to form a historical eye use score sequence; extract the unit intervention scores of the same group of valid time units, arrange them in the same chronological order to form a historical response score sequence; The Pearson correlation coefficient between the historical eye use score sequence and the historical response score sequence is calculated as the historical correlation degree. Further, the interference weight determination process includes: For each behavior segment, if the number of valid time units within the behavior segment is less than a preset threshold, its interference weight is reset to zero. If the number of valid time units within a behavior segment is not less than a preset threshold, calculate the absolute difference between the historical correlation degree and the segment correlation degree as the behavior segment's difference degree; for the current behavior segment, calculate the sum of the difference degrees of all behavior segments with a number of valid time units not less than a preset threshold within a preset number of behavior segments before the current behavior segment, as the historical difference degree sum. Determine if the sum of historical differences is zero; if it is zero, reset the interference weight of all behavior segments to zero; if it is not zero, divide the difference of the current behavior segment by the sum of historical differences to obtain the interference weight of the current behavior segment.
[0012] Furthermore, based on the interference weights and environmental parameters, the intervention prompt strategy is dynamically adjusted, including: For each behavior segment, obtain the interference weight of the behavior segment and the environmental parameters within the behavior segment; wherein, the environmental parameters include at least illuminance and flicker value; If the illuminance is greater than or equal to the preset illuminance threshold and the flicker value is less than the preset flicker threshold, the environmental parameters are determined to be within the preset normal range; otherwise, the environmental parameters are determined to be outside the preset normal range. The difference between historical correlation and fragment correlation is calculated as the deviation direction value; Based on whether the interference weight, deviation direction value, and environmental parameters are within the preset normal range, a corresponding intervention prompt adjustment scheme is matched to achieve adaptive adjustment of the intervention prompt strategy.
[0013] The present invention has the following beneficial effects: This invention divides the analysis period into multiple equal-length time segments as behavioral fragments, thereby calculating the fragment correlation degree of each behavioral fragment and the historical correlation degree based on a historical sliding window. The interference weight of each behavioral fragment is determined based on the difference between these two values. The interference weight quantifies the difference between the correlation degree between user eye-use behavior and intervention response within the corresponding behavioral fragment and the correlation degree within recent historical time periods. This allows for precise identification of key time periods of declining user compliance, providing data support for subsequent intervention adjustments. Through an effective time unit screening mechanism, only time units with intervention response scores are included in subsequent calculations, avoiding interference from time units without intervention events. This ensures the reliability of the calculation results and effectively improves the utilization efficiency of system computing resources, allowing limited intervention resources to be concentrated on the truly important time periods. The intervention prompt strategy is dynamically adjusted based on the interference weight of different behavioral fragments and environmental parameters. Different behavioral fragments will undergo differentiated dynamic adjustment strategies, thereby improving user acceptance and willingness to implement intervention measures, thus enhancing overall compliance. Attached Figure Description
[0014] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0015] Figure 1 This is a schematic diagram of a myopia intervention compliance management system based on eye behavior monitoring, provided as an embodiment of the present invention. Detailed Implementation
[0016] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a myopia intervention compliance management system based on eye behavior monitoring proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.
[0017] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.
[0018] The following description, in conjunction with the accompanying drawings, details a specific scheme for a myopia intervention compliance management system based on eye behavior monitoring provided by the present invention.
[0019] Please see Figure 1 The diagram illustrates a structural schematic of a myopia intervention compliance management system based on eye use behavior monitoring, according to an embodiment of the present invention. The system includes: The data acquisition module 101 is used to collect the user's eye use status parameters, intervention response parameters, and environmental parameters.
[0020] Eye usage status parameters are used to describe the user's behavioral state during eye use.
[0021] As a preferred implementation, the user's eye-use behavior is detected by smart monitoring glasses, and each continuous eye-use process is marked as an eye-use event. Each eye-use event has a start time, an end time, and a duration. Within each eye-use event, eye-use status parameters are collected according to a preset sampling frequency. The eye-use status parameters include at least one of the following: eye-use distance, eye-use posture, blinking frequency, and continuous eye-use duration.
[0022] An eye-use event refers to a continuous process of eye use by a user, such as a complete process of doing homework, reading a book, or using an electronic device, which has a clear start time, end time, and duration.
[0023] Start time: The timestamp of when the eye-use event began.
[0024] End Time: The timestamp indicating when the eye-use event ended.
[0025] Duration: End time minus start time, usually in minutes, representing the cumulative duration of continuous eye use.
[0026] In real-life scenarios, users' eye-use behavior is not continuous but consists of multiple independent processes. Therefore, to facilitate subsequent evaluation of the quality of each eye-use process, user eye-use behavior can be structured through event-based segmentation. Thus, user eye-use behavior is detected and marked as eye-use events.
[0027] It should be noted that mature smart wearable devices on the market (such as smart clips and smart glasses) already have built-in distance sensors, posture sensors, and eye-tracking sensors, which can detect parameters such as eye distance, eye posture, and blink frequency in real time, and the duration of continuous eye use can be accumulated through a timer. Therefore, the acquisition of eye state parameters involved in this solution can all be achieved based on existing sensor technologies and can be directly applied here.
[0028] It should be noted that the preset sampling frequency refers to the time interval at which the system collects eye-use status parameters within an eye-use event. Its specific value can be determined based on industry experience, and this embodiment does not impose a specific limitation. For example, a typical preset sampling frequency is once per minute.
[0029] In essence, within a single eye-use event, the system continuously collects eye-use status parameters at a preset sampling frequency. Each collection is called a sampling point, and each sampling point yields a set of eye-use status parameters (including at least one of the following: eye-use distance, eye-use posture, blinking frequency, and continuous eye-use duration). Therefore, an eye-use event contains multiple sampling points, and each sampling point corresponds to a set of eye-use status parameters.
[0030] In a preferred implementation, when the system issues an intervention prompt, it marks the start of an intervention event; when it detects that the user has responded to the intervention prompt or the waiting time exceeds a preset response duration threshold, it marks the end of the intervention event; within each intervention event, intervention response parameters are collected; wherein, the intervention response parameters include at least one of the following: intervention response determination result, intervention response duration, and intervention response completion rate.
[0031] An intervention event refers to a complete process from when the system issues an intervention prompt to when the user responds, or when the response times out (i.e., the waiting time exceeds the preset response duration threshold).
[0032] It is understandable that, unlike eye-use events, an intervention event corresponds to a single, independent "cue-response" process and does not involve multiple sampling points. Therefore, within an intervention event, the system collects a set of intervention response parameters (including at least one of the following: intervention response determination result, intervention response duration, and intervention response completion rate).
[0033] Intervention prompts are reminder signals issued by the system when it detects that a user has poor eye-use behavior (such as using the eyes at too close a distance, using the eyes for too long continuously, or blinking too infrequently).
[0034] The intervention prompts can take the form of voice broadcasts, vibration alerts, interface pop-ups, or a combination of the above, to guide users to adjust their eye-use behavior (such as "please look up and gaze into the distance" or "please adjust your posture"), and there are no specific limitations.
[0035] Specifically, the system can use sensors built into the smart monitoring glasses (such as distance sensors, posture sensors, and eye-tracking sensors) to determine whether the user has performed the eye-use behavior required by the intervention prompt. For example, when the pre-prompt is "Please look up and gaze into the distance," the system uses the distance sensor to detect whether the distance between the user's eyes and the target being gazed at has changed from near to far; if the intervention prompt is "Please adjust your posture," the system uses the posture sensor to detect whether the user's head angle has returned to a normal position. The above detection processes all use existing sensor technology and will not be elaborated further here.
[0036] It should be noted that the preset response time threshold refers to the maximum time the system waits for a user response after issuing an intervention prompt. Its specific value can be adjusted based on the user's age, the type of intervention prompt, or the usage scenario; this embodiment does not impose a specific limitation. For example, the preset response time threshold is set to 30 seconds.
[0037] Intervention response determination result: Record whether the user responded to the intervention prompt. It is a binary variable used to assess whether the user noticed the intervention prompt and was willing to cooperate.
[0038] Intervention response duration: This records the duration for which the user continues to perform the response after responding to the intervention prompt, usually in seconds. For example, if the system issues an intervention prompt "Please look into the distance for 2 minutes," and the user actually looks into the distance for 45 seconds, then the intervention response duration is 45 seconds.
[0039] Intervention response completion rate: The ratio of the actual duration of the user's response to the duration required in the intervention prompt provided by the system, usually expressed as a percentage. For example, if the system issues an intervention prompt "Please look into the distance for 2 minutes," which requires 120 seconds, and the user responds for 45 seconds, then the intervention response completion rate is 37.5%.
[0040] Environmental parameters are used to describe the physical environmental conditions in which a user is using their eyes.
[0041] As a preferred option, environmental parameters include illuminance and flicker value. Illuminance reflects the brightness of ambient light and is measured in lux (Lux); flicker value reflects the fluctuation of the light source and is expressed as a percentage. Specifically, illuminance and flicker value can be collected by ambient light sensors and flicker sensors, respectively, which are applications of existing sensors and will not be elaborated on here.
[0042] The time period segmentation module 102 is used to set an analysis period and divide the analysis period into multiple time periods of equal length as behavior segments.
[0043] The analysis period refers to the time range within which the system conducts overall compliance analysis, covering multiple complete life cycles of the user. Its value is determined based on the user's actual situation, and this embodiment does not impose a specific limitation. For example, the analysis period can be preset to 180 days. This analysis period can typically include various daily life states of the user, such as school weeks, weekends, exam weeks, holidays, etc.
[0044] Because equal-length time periods facilitate horizontal comparisons between different behavioral segments, and this division method is simple and clear, the analysis period is divided using equal-length time periods. Specifically, using the natural week as the basic unit, the 180-day analysis period is divided into 26 behavioral segments, with each behavioral segment corresponding to one week.
[0045] The scoring mapping module 103 is used to map eye use status parameters to eye use scores according to a preset first scoring rule, and to map intervention response parameters to intervention response scores according to a preset second scoring rule.
[0046] It should be noted that the preset first scoring rules are different for different types of eye use parameters (i.e., eye distance, eye posture, blinking frequency, and continuous eye use duration).
[0047] Similarly, the preset second scoring rules are different for different types of intervention response parameters (i.e., intervention response judgment result, intervention response duration, and intervention response completion rate).
[0048] Because parameters such as eye distance, eye posture, and blink frequency have different dimensions and measurement ranges, they cannot be directly used for subsequent calculations or comparisons. Therefore, they need to be mapped to the same preset score range. Specifically, each eye use score maps each eye use state parameter to a score value within the same preset score range according to its corresponding preset first scoring rule.
[0049] Similarly, as an example, each intervention response score is obtained by mapping each intervention response parameter to a score value within the same preset score range according to its corresponding preset second score rule. The preset score range is the same as the preset score range of the eye use score.
[0050] It should be noted that the specific range of the preset score can be determined based on actual needs, and this embodiment does not impose a specific limitation. For example, the typical range of the preset score is 0 to 100.
[0051] It should be noted that the preset first and second scoring rules can be pre-set based on ophthalmological medical knowledge and clinical experience, and this embodiment does not impose specific limitations. Specifically, these rules can be stored in the system in the form of configuration files or database tables, and can be adjusted according to specific use cases.
[0052] For example, the preset scoring rules for each eye use index include the following: First, scoring for eye use distance: a preset ideal eye use distance is set. If the actual eye use distance reaches or exceeds the preset ideal eye use distance, the score is 100 points; if the actual eye use distance is less than the ideal distance, the score is equal to the actual distance divided by the preset ideal eye use distance and then multiplied by 100. Second, scoring for eye use posture: a preset mapping relationship between posture state and score is set. Different posture states correspond to different scores, and the more upright the posture, the higher the score. Third, scoring for blinking frequency: a preset ideal blinking frequency range is set. If the actual blinking frequency is within the preset ideal blinking frequency range, the score is... The score is 100 points. If the measured blink frequency deviates from the preset ideal blink frequency range, the score is equal to 100 minus the degree of deviation (such as the absolute difference between the measured blink frequency and the median of the preset ideal blink frequency range) multiplied by the preset first deduction coefficient, with a minimum of 0 points. Fourth, the score is based on continuous eye use duration: the preset ideal continuous eye use duration is set. If the measured continuous eye use duration is less than or equal to the preset ideal continuous eye use duration, the score is 100 points. Conversely, if the measured continuous eye use duration exceeds the preset ideal continuous eye use duration, the score is equal to 100 minus the excess continuous eye use duration multiplied by the preset second deduction coefficient, with a minimum of 0 points.
[0053] For example, the preset scoring rules for each intervention response indicator include the following: First, scoring based on the intervention response judgment result: if the user responds to the intervention prompt, the score is full marks; otherwise, if the user does not respond to the intervention prompt, the score is zero marks. Second, scoring based on the intervention response duration: a preset ideal response duration is set. If the actual response duration reaches or exceeds the preset ideal response duration, the score is full marks; if the actual response duration is less than the preset ideal response duration, the score is equal to the actual response duration divided by the preset ideal response duration, and then multiplied by the full mark (i.e., the maximum value within the preset score range). Third, scoring based on the intervention response completion rate: the score is equal to the response completion rate multiplied by the full mark.
[0054] It should be noted that the parameters mentioned in the above examples, such as the preset ideal eye distance, the mapping relationship between preset posture state and score, the preset ideal blink frequency range, the preset deduction coefficient, the preset ideal continuous eye use duration, and the preset ideal response duration, can all be preset based on ophthalmological medical knowledge and clinical practice experience. In actual applications, they can be adjusted according to individual user differences or different usage scenarios. This embodiment does not impose any specific limitations. For example, the preset ideal viewing distance is 33 cm, derived from classic ophthalmological research on accommodative needs; the preset mapping relationship between posture and score is: upright posture corresponds to 100 points, slight misalignment corresponds to 70 points, and severe misalignment corresponds to 30 points; the preset ideal blinking frequency range is 15 to 20 times per minute, taken from normal human eye physiological indicators; the preset first deduction coefficient for blinking frequency is 8, and the preset second deduction coefficient for continuous viewing duration is 2, both of which are empirical values to ensure that the greater the deviation from the ideal value, the more reasonable the score decrease; the preset ideal continuous viewing duration is 20 minutes, following the common "20-20-20" eye care rule; the preset ideal response duration is 120 seconds, which is the clinically recommended duration for distant relaxation.
[0055] In summary, the scoring rules described above are essentially mathematical functions or conditional statements, which can be implemented using if-else conditional branches or linear calculation formulas in computer programs, and will not be elaborated further here.
[0056] The correlation analysis module 104 is used to divide each behavioral segment into time units; filter out valid time units with intervention response scores; determine the segment correlation degree based on the correlation between eye use scores and response scores of all valid time units within the behavioral segment; determine the historical correlation degree based on the historical correlation between eye use scores and response scores of all valid time units within a preset number of previous behavioral segments; and determine the interference weight of the current behavioral segment based on the difference between the segment correlation degree and the historical correlation degree.
[0057] In this embodiment, for each behavioral segment, the behavioral segment is divided into multiple consecutive time units according to a preset time interval; it is determined whether there is an intervention response score in each time unit; the time unit with the intervention response score is marked as a valid time unit; the time unit without the intervention response score is marked as an invalid time unit; wherein, the invalid time unit will not participate in the determination of the segment correlation degree and the determination of the historical correlation degree.
[0058] It should be noted that the specific value of the preset time interval can be adjusted according to actual needs, and this embodiment does not impose a specific limitation. For example, generally speaking, dividing the time interval into days can match the user's daily life rhythm (such as studying during the day and resting at night), and a day usually contains multiple eye-use events and possible intervention events. Therefore, the preset time interval is often set to 24 hours, that is, with the calendar day as the basic unit.
[0059] In addition, if the number of valid time units within a certain behavior segment is less than a preset threshold, the segment correlation and interference weight of that behavior segment will not be calculated, and its interference weight will be set to zero.
[0060] It should be noted that the specific value of the preset quantity threshold can be set based on industry experience, and this embodiment does not impose a specific limitation. A typical value for the preset quantity threshold is 2.
[0061] It's important to understand that since each valid time unit may contain multiple eye use events and multiple intervention events, directly using the raw data for calculation would result in an excessively large data volume. Furthermore, the raw data contains random fluctuations between sampling points, which would introduce unnecessary noise if used directly. Therefore, it is necessary to first aggregate all eye use scores within each valid time unit into a unit eye use score, and aggregate all intervention response scores within the same time unit into a unit intervention score, to unify the time granularity and provide a regular data sequence for subsequent calculations.
[0062] In a preferred embodiment, the eye use scores of all sampling points for all eye use events within each valid time unit are obtained; for each sampling point, the eye use scores of the sampling point are weighted and summed according to a preset first weight set to obtain the instantaneous eye use score of the sampling point; for each eye use event, the arithmetic mean of the instantaneous eye use scores of all sampling points within the eye use event is calculated as the event eye use score of the eye use event; for each valid time unit, the arithmetic mean of the event eye use scores of all eye use events within the valid time unit is calculated as the unit eye use score of the valid time unit.
[0063] Since different eye-use parameters have varying degrees of impact on visual health—for example, eye distance is more critical than blinking frequency—weights can be used to reflect the differences in importance among these parameters. Specifically, a preset first weight set is a set of weight values assigned to each eye-use parameter.
[0064] It should be noted that the specific values of the preset first weight set can be determined based on ophthalmological medical knowledge and clinical experience, and this embodiment does not impose specific limitations. For example, viewing distance is the most important factor affecting the development of myopia, so it is given the highest weight of 0.4; viewing posture directly affects the balance of force on both eyes, so it is also given the second highest weight of 0.3; too low blinking frequency can lead to tear film rupture and eye fatigue, which has a certain impact on the development of myopia, so it is given a weight of 0.2; although continuous viewing time is important, users can alleviate it through multiple short breaks, so it is given a relatively low weight of 0.1.
[0065] Since eye use behavior may fluctuate at different sampling points within a single eye use event (e.g., occasionally adjusting posture and then reverting to poor posture), taking the average value can reflect the overall quality of the eye use event, thus avoiding evaluation bias caused by accidental fluctuations at a single sampling point.
[0066] As a preferred implementation, for each effective time unit, the intervention response scores of all intervention events within the effective time unit are obtained; for each intervention event, the intervention response scores of the intervention event are weighted and summed according to a preset second weight set to obtain the instantaneous intervention score of the intervention event; the arithmetic mean of the instantaneous intervention scores of all intervention events within the effective time unit is calculated as the unit intervention score of the effective time unit.
[0067] Since different intervention response parameters contribute differently to compliance assessment—for example, the intervention response determination directly reflects whether a user is willing to cooperate and is more critical than the duration of the intervention response—weights can be used to reflect the differences in importance among the various intervention response parameters. Specifically, a pre-defined second weight set is a set of weight values assigned to each intervention response parameter.
[0068] It should be noted that the specific values of the preset second weight set can be based on the differences in the importance of intervention response parameters to compliance assessment, and this embodiment does not impose specific limitations. For example, the intervention response judgment result directly reflects whether the user is willing to cooperate with the intervention prompts, and is the most basic and core basis for judging compliance, so it is given the highest weight of 0.4; the two indicators of intervention response duration and intervention response completion are highly correlated (i.e., intervention response completion is calculated based on intervention response duration), and both reflect the degree to which the user implements the intervention measures, so they are each given the same weight of 0.3.
[0069] In this embodiment, the unit intervention score for each effective time unit is obtained; for each behavior segment, the unit eye use score for all effective time units within the behavior segment is extracted and arranged in chronological order to form an eye use score sequence; the unit intervention scores for the same group of effective time units within the same behavior segment are extracted and arranged in the same chronological order to form a response score sequence; the Pearson correlation coefficient between the eye use score sequence and the response score sequence is calculated as the segment correlation degree of the behavior segment.
[0070] It is understandable that the eye use rating sequence and the response rating sequence are of equal length, and the data at corresponding positions come from the same time unit.
[0071] In actual calculation of the Pearson correlation coefficient, there may be extreme cases where all values in the eye use score sequence or response score sequence are completely equal (e.g., a user's eye use score is 100 points in all valid time units). In this case, the standard deviation of the corresponding sequence will be zero, and the denominator of the Pearson correlation coefficient will be zero, leading to calculation failure. Therefore, to avoid system errors in this situation, the variance of the two sequences is checked before calculation: if the variance of either sequence is zero, the segment correlation degree of that behavior segment is directly set to a constant (e.g., 1 if both sequences are constant sequences, 0 if only one sequence is a constant sequence); if the variances of both sequences are not zero, the Pearson correlation coefficient is calculated as the segment correlation degree. The above processing method is a common anomaly handling rule in the case of single data, and is only a preferred example. Technicians in this field can adjust it according to actual needs, which will not be elaborated here.
[0072] Segment correlation is used to quantify the degree of correlation between the quality of a user's visual behavior and the quality of the intervention response within a specific behavioral segment. Its value ranges from -1 to +1. A positive segment correlation indicates a positive correlation between visual behavior quality and intervention response quality; that is, better user response leads to better visual behavior. A negative segment correlation indicates a negative correlation; that is, better user response may result in worse visual behavior, which is an anomaly. The closer the absolute value of the segment correlation is to +1, the stronger the correlation. A value close to zero indicates no significant correlation between visual behavior and intervention response.
[0073] In this embodiment, the unit intervention score for each valid time unit is obtained; for the current behavior segment, the unit eye use scores of all valid time units within a preset number of behavior segments preceding the behavior segment are extracted and arranged in chronological order to form a historical eye use score sequence; the unit intervention scores of the same group of valid time units are extracted and arranged in the same chronological order to form a historical response score sequence; the Pearson correlation coefficient between the historical eye use score sequence and the historical response score sequence is calculated as the historical correlation.
[0074] The current behavior segment refers to the segment of behavior that the system is currently processing or that requires intervention and adjustment. For example, if the system is analyzing eye usage data for this week (e.g., week 5) and determining the intervention strategy for this week, then this week is the current behavior segment.
[0075] Historical correlation is used to quantify the degree of correlation between the quality of user eye behavior and the quality of intervention response within a preset number of behavior segments prior to the current behavior segment, with a value ranging from negative one to positive one.
[0076] It should be noted that the preset number of behavioral segments can be a fixed value (e.g., the previous 4 weeks, or one month), or it can be all historical segments before the current segment. Preferably, the preset number of behavioral segments is 4, that is, the 4 behavioral segments before the current behavioral segment (usually about 4 weeks) are used as the historical sliding window. The length of this historical sliding window can ensure that there is enough data to calculate reliable correlation, and can also reflect the recent behavioral change trend of users in a timely manner.
[0077] In the early stages of the analysis period, such as the first few weeks, the number of historical behavioral segments preceding the current behavioral segment may be insufficient to meet the preset number. In this case, the system cannot calculate the complete sum of historical correlation and historical dissimilarity. Therefore, the system can directly set the interference weight of the current behavioral segment to zero and temporarily refrain from differentiated intervention adjustments. The complete interference weight calculation will only be implemented after sufficient historical data has been accumulated.
[0078] It's important to note that, unlike segment relevance, historical relevance is a value calculated based on all valid time units within a preset number of previous behavioral segments (i.e., the historical sliding window). It represents the user's average performance over recent historical periods and serves as a baseline for judging whether the current behavioral segment deviates from recent typical performance. Specifically, if the segment relevance of a particular behavioral segment differs significantly from its historical relevance, it indicates a significant deviation between the eye-response relationship within that segment and the user's historical typical performance. This suggests that the behavioral segment is more likely to have been affected by short-term changes in lifestyle (such as exam week or holidays).
[0079] Similarly, in the actual calculation of the Pearson correlation coefficient, an extreme case may occur where all values in the historical eye use score series or the historical response score series are completely equal. In this case, the standard deviation of one series is zero, the denominator of the Pearson correlation coefficient is zero, and it cannot be calculated normally. Therefore, to avoid system errors, the variance of the two series is checked first before calculation: if the variance of either series is zero, the historical correlation is directly set to a constant (for example, set to 1 if both series are constant series, and set to 0 if only one series is a constant series); if the variance of both series is not zero, the Pearson correlation coefficient is calculated as the historical correlation.
[0080] It is important to understand that the correlation between user eye-use behavior and intervention response varies across different behavioral segments, and this difference may stem from various short-term changes in lifestyle. Therefore, the correlation between segments alone cannot directly determine which behavioral segments are affected. Consequently, it is necessary to calculate the deviation (i.e., the degree of difference) between the correlation of each segment and its historical correlation, and then normalize the degree of difference for each segment to obtain an interference weight. This allows for the quantification of the relative severity of short-term interference to each behavioral segment throughout the entire analysis period, providing a quantitative basis for the dynamic adjustment of subsequent intervention strategies.
[0081] In a preferred implementation, for each behavior segment, if the number of valid time units within the behavior segment is less than a preset threshold, its interference weight is reset to zero; if the number of valid time units within the behavior segment is not less than the preset threshold, the absolute difference between the historical correlation and the segment correlation is calculated as the difference degree of the behavior segment; for the current behavior segment, the sum of the difference degrees of all behavior segments with a number of valid time units not less than the preset threshold within a preset number of behavior segments before the current behavior segment is calculated as the historical difference degree sum; it is determined whether the historical difference degree sum is zero; if it is zero, the interference weight of all behavior segments is reset to zero; if it is not zero, the difference degree of the current behavior segment is divided by the historical difference degree sum to obtain the interference weight of the current behavior segment.
[0082] Since both fragment correlation and historical correlation range from -1 to +1, their absolute difference directly reflects the degree of deviation between the fragment and the whole, and is unaffected by positive or negative directions. For example, when the historical correlation is 0.8 and the fragment correlation is 0.3, the absolute difference is 0.5; when the fragment correlation is -0.5, the absolute difference is 1.3, the latter being larger, consistent with the judgment that negatively correlated behavioral fragments deviate more severely.
[0083] The interference weight is the degree of difference after local normalization based on the historical sliding window, and its value ranges from zero to one. For each current behavior segment, its interference weight reflects the relative degree of interference of the segment among all valid segments in the historical sliding window, and the sum of the interference weights of all valid segments in the historical sliding window is 1.
[0084] It should be noted that after each behavior segment completes its calculation, its difference value is stored in the system. To calculate the interference weight of the current behavior segment, the difference values of a preset number of behavior segments preceding the current behavior segment are read, and these values are summed to obtain the historical difference sum. Therefore, the difference of past behavior segments is derived based on the historical window of their respective calculation times.
[0085] The interference weight is used to quantify the relative severity of short-term interference for each behavioral segment throughout the analysis period. A higher interference weight for a behavioral segment indicates a greater deviation between its segment correlation and historical correlation; that is, a more significant difference between the eye-use-response relationship within that segment and the user's historical routine performance. This suggests that the behavioral segment is more likely to have been subject to strong short-term interference (such as the impact of exam weeks, holidays, or other changes in lifestyle rhythm), and should be given more attention in subsequent intervention adjustments. Conversely, a lower interference weight for a behavioral segment indicates a smaller deviation between its segment correlation and historical correlation; that is, a more consistent eye-use-response relationship within that segment and the user's historical routine performance. This suggests that the behavioral segment has not been subject to significant short-term interference, and the current intervention strategy can be maintained.
[0086] The intervention adjustment module 105 is used to dynamically adjust the intervention prompt strategy based on the interference weight and environmental parameters.
[0087] As an example, for each behavioral segment, the interference weight of the behavioral segment and the environmental parameters within the behavioral segment are obtained. The environmental parameters include at least illuminance and flicker value. If the illuminance is greater than or equal to a preset illuminance threshold and the flicker value is less than a preset flicker threshold, the environmental parameters are determined to be within a preset normal range; otherwise, the environmental parameters are determined to be outside the preset normal range. The difference between the historical correlation and the segment correlation is calculated as the deviation direction value. Based on whether the interference weight, the deviation direction value, and the environmental parameters are within the preset normal range, the corresponding intervention prompt adjustment scheme is matched to achieve adaptive adjustment of the intervention prompt strategy.
[0088] The deviation direction value reflects the direction in which the performance of the current behavior segment deviates from the historical baseline.
[0089] Specifically, if the interference weight is higher than a preset first threshold, the deviation direction value is positive, and the environmental parameters are within a preset normal range, the frequency of intervention prompts is increased; if the interference weight is higher than a preset first threshold, the deviation direction value is positive, and the environmental parameters exceed a preset normal range, environmental improvement suggestions are sent first; if the interference weight is higher than the first threshold and the deviation direction value is negative (i.e., the current behavior segment performs better than the overall baseline), the frequency of intervention prompts is maintained or reduced, and encouraging prompts are optionally sent; if the interference weight is lower than a preset first threshold and the environmental parameters exceed a preset normal range, environmental improvement suggestions are sent along with encouraging prompts; if the interference weight is lower than a preset first threshold and the environmental parameters are within a preset normal range, the frequency of intervention prompts is maintained or reduced. It should be noted that the specific values of the preset illumination threshold and the preset flicker threshold can be set based on common standards in the field of lighting engineering, and this embodiment does not impose specific limitations. For example, the preset illumination threshold is usually 300 to 500 lux, and the light is usually too dim when it is below 300 lux; since the human eye can hardly perceive flicker depth below 3%, some sensitive people may feel uncomfortable at 3% to 5%, and flicker is obvious when it exceeds 5%, the preset flicker threshold can be set to 5%.
[0090] It should be noted that the specific value of the preset first threshold can be determined based on the range of interference weights and practical application experience, and this embodiment does not impose a specific limitation. For example, it is known that if the interference weight of a certain behavior segment is greater than 0.5, it means that the difference of the behavior segment exceeds half of the sum of the historical differences of all non-abnormal segments, indicating that the deviation of the behavior segment from the overall baseline is relatively significant. This is already a high-interference segment that needs to be focused on, so the preset first threshold can be set to 0.5.
[0091] It should be understood that the above-mentioned intervention strategy adjustment schemes (such as increasing the frequency of prompts, sending suggestions for environmental improvement, and attaching encouraging prompts) are all exemplary descriptions of this embodiment. In practical applications, the specific intervention prompt adjustment methods can be flexibly configured according to the actual application scenario, and those skilled in the art can implement various modified adjustment strategies, which will not be elaborated here.
[0092] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0093] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.
Claims
1. A myopia intervention compliance management system based on eye use behavior monitoring, characterized in that, The system includes: The data acquisition module is used to collect users' eye use status parameters, intervention response parameters, and environmental parameters; The time period segmentation module is used to set an analysis period and divide the analysis period into multiple equal-length time periods as behavior segments. The scoring mapping module is used to map eye use status parameters to eye use scores according to a preset first scoring rule; and to map intervention response parameters to intervention response scores according to a preset second scoring rule. The correlation analysis module is used to divide each behavior segment into time units; filter out valid time units with intervention response scores; determine the segment correlation degree based on the correlation between eye use scores and response scores of all valid time units within the behavior segment; determine the historical correlation degree based on the historical correlation between eye use scores and response scores of all valid time units within a preset number of behavior segments before the current behavior segment; and determine the interference weight of the current behavior segment based on the difference between the segment correlation degree and the historical correlation degree. The intervention adjustment module is used to dynamically adjust the intervention prompt strategy based on the interference weight and environmental parameters.
2. The myopia intervention compliance management system based on eye use behavior monitoring according to claim 1, characterized in that, The process of collecting eye use status parameters includes: By using smart monitoring glasses to detect users’ eye-use behavior, each continuous eye-use process is marked as an eye-use event, and each eye-use event has a start time, end time and duration. Within each eye-use event, eye-use status parameters are collected according to a preset sampling frequency; among them, eye-use status parameters include at least one of the following: eye-use distance, eye-use posture, blinking frequency, and continuous eye-use duration.
3. A myopia intervention compliance management system based on eye use behavior monitoring according to claim 2, characterized in that, The intervention response parameter acquisition process includes: When the system issues an intervention prompt, mark the start of an intervention event; if it detects that the user has responded to the intervention prompt or the waiting time exceeds the preset response time threshold, mark the end of the intervention event. Within each intervention event, intervention response parameters are collected; these parameters include at least one of the following: intervention response determination result, intervention response duration, and intervention response completion rate.
4. A myopia intervention compliance management system based on eye use behavior monitoring according to claim 1, characterized in that, The effective time unit filtering process includes: For each behavior segment, the behavior segment is divided into multiple consecutive time units according to a preset time interval; Determine whether an intervention response score exists in each time unit; mark time units with intervention response scores as valid time units; mark time units without intervention response scores as invalid time units; invalid time units will not participate in the determination of segment correlation and historical correlation.
5. A myopia intervention compliance management system based on eye use behavior monitoring according to claim 3, characterized in that, Each eye use score is obtained by mapping each eye use status parameter to a score value within the same preset score range according to its corresponding preset first scoring rule. The system is also used for: Obtain the eye use scores for all sampling points of all eye use events within each valid time unit; For each sampling point, the eye use scores of that sampling point are weighted and summed according to a preset first weight set to obtain the instantaneous eye use score of that sampling point; for each eye use event, the arithmetic mean of the instantaneous eye use scores of all sampling points within that eye use event is calculated as the event eye use score of that eye use event. For each valid time unit, the arithmetic mean of the event eye use scores of all eye use events within that valid time unit is calculated, and this is used as the unit eye use score for that valid time unit.
6. A myopia intervention compliance management system based on eye use behavior monitoring according to claim 5, characterized in that, The scoring of each intervention response is achieved by mapping each intervention response parameter to a score value within the same preset score range according to its corresponding preset second scoring rule. The preset score range is the same as the preset score range of the eye use score. The system is also used for: For each valid time unit, obtain the intervention response scores for each intervention event within that valid time unit; For each intervention event, the scores of each intervention response within that intervention event are weighted and summed according to a preset second weight set to obtain the instantaneous intervention score for that intervention event; Calculate the arithmetic mean of the instantaneous intervention scores of all intervention events within the effective time unit, and use it as the unit intervention score for the effective time unit.
7. A myopia intervention compliance management system based on eye use behavior monitoring according to claim 6, characterized in that, The process of determining the correlation degree of the segments includes: For each behavioral segment, extract the unit eye use scores of all valid time units within the behavioral segment, arrange them in chronological order to form an eye use score sequence; extract the unit intervention scores of the same group of valid time units within the same behavioral segment, arrange them in the same chronological order to form a response score sequence; Calculate the Pearson correlation coefficient between the eye use rating sequence and the response rating sequence as the segment correlation degree of the behavioral segment.
8. A myopia intervention compliance management system based on eye use behavior monitoring according to claim 6, characterized in that, The process for determining the historical correlation includes: For the current behavior segment, extract the unit eye use scores of all valid time units within a preset number of previous behavior segments, arrange them in chronological order to form a historical eye use score sequence; extract the unit intervention scores of the same group of valid time units, arrange them in the same chronological order to form a historical response score sequence; The Pearson correlation coefficient between the historical eye use score series and the historical response score series is calculated as the historical correlation degree.
9. A myopia intervention compliance management system based on eye use behavior monitoring according to claim 1, characterized in that, The interference weight determination process includes: For each behavior segment, if the number of valid time units within the behavior segment is less than a preset threshold, its interference weight is reset to zero. If the number of valid time units within a behavior segment is not less than a preset threshold, calculate the absolute difference between the historical correlation degree and the segment correlation degree as the behavior segment's difference degree; for the current behavior segment, calculate the sum of the difference degrees of all behavior segments with a number of valid time units not less than a preset threshold within a preset number of behavior segments before the current behavior segment, as the historical difference degree sum. Determine if the sum of historical differences is zero; if it is zero, reset the interference weight of all behavior segments to zero; if it is not zero, divide the difference of the current behavior segment by the sum of historical differences to obtain the interference weight of the current behavior segment.
10. A myopia intervention compliance management system based on eye use behavior monitoring according to claim 1, characterized in that, Based on the interference weight and environmental parameters, the intervention prompt strategy is dynamically adjusted, including: For each behavior segment, obtain the interference weight of the behavior segment and the environmental parameters within the behavior segment; wherein, the environmental parameters include at least illuminance and flicker value; The difference between historical correlation and fragment correlation is calculated as the deviation direction value; If the illuminance is greater than or equal to the preset illuminance threshold and the flicker value is less than the preset flicker threshold, the environmental parameters are determined to be within the preset normal range; otherwise, the environmental parameters are determined to be outside the preset normal range. Based on whether the interference weight, deviation direction value, and environmental parameters are within the preset normal range, a corresponding intervention prompt adjustment scheme is matched to achieve adaptive adjustment of the intervention prompt strategy.
