An eye movement tracking interaction method for myopia prevention and control analysis of children

Abstract

The application discloses an eye movement tracking interaction method for myopia prevention and control analysis of children, and particularly relates to the field of eye movement tracking analysis, and aims to solve the problems that it is difficult to identify the abnormal state of micro-vision adjustment when continuously watching animation and interactive content, and it is impossible to accurately locate the high-risk eye content segment in the existing myopia prevention and control process of children; the application detects the line-of-sight landing point stable state in the watching process of children, collects corneal reflection offset information and pupil change information after stable fixation is formed, and performs independent component analysis, phase correlation analysis and phase-locked analysis on the fixation micro-motion change and adjustment correlation fluctuation, thereby constructing a micro-motion-adjustment coupling coordination coefficient, and identifying the continuous low-coordination fixation interval; the corresponding time interval is further mapped to the current playing content time axis, a high-risk eye segment labeling label bound with the content position is generated, and the fine degree and content positioning ability of myopia prevention and control analysis of children are improved.

Description

Technical Field

[0001] This invention relates to the field of eye-tracking analysis technology, and more specifically, to an eye-tracking interactive method for myopia prevention and control analysis in children. Background Technology

[0002] With children spending extended periods of time using electronic display devices such as tablets, learning machines, and smart TVs, close-range, sustained viewing of animation, interactive courses, and digital reading content has gradually become an important part of children's daily learning and entertainment. While watching animation and interactive content, children's eyes typically maintain a prolonged state of focused attention, which can easily lead to problems such as continuously increased accommodative load, decreased fixational stability, and accumulated visual fatigue, thereby impacting myopia prevention and control.

[0003] Current myopia prevention programs for children mostly focus on viewing distance detection, ambient light monitoring, usage time statistics, or blink frequency analysis. These programs typically only provide a rough overview of external eye-use behavior and fail to reflect changes in children's micro-accommodative states during actual viewing. Especially in scenarios involving animation, dynamic interactions, and highly engaging visual content, children tend to maintain intense focus for extended periods. This can lead to a gradual loss of coordination between micro-eye movements and accommodative activity. However, existing programs generally lack the ability to continuously analyze the coupling relationship between micro-eye movements and accommodative activity, making it impossible to accurately identify specific content segments corresponding to the continuous accumulation of visual load.

[0004] Meanwhile, traditional eye-tracking analysis methods typically rely on fixation duration or gaze trajectory as the primary analytical basis. They lack effective modeling methods for the correlation between fine-grained eye-tracking behaviors such as microsaccades and drift and accommodative fluctuations. This makes it difficult for the system to accurately locate high-risk eye-use scenarios and fails to provide effective support for content optimization, viewing intervention, and high-risk segment identification in myopia prevention and control in children. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide an eye-tracking interactive method for myopia prevention and control analysis in children to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: An eye-tracking interactive method for myopia prevention and control analysis in children includes the following steps: S1. When the user's gaze falls on the animation playback interface of the display device, and the spatial dispersion of the gaze point decreases and enters a stable state, a collection trigger signal is generated. S2. Upon receiving the acquisition trigger signal, continuously capture the pupil images and corneal reflection images of both eyes, and real-time corneal reflection offset vector and pupil diameter to form a raw signal stream with timestamp alignment; S3. Perform independent component separation on the original signal stream to obtain the gaze micro-motion source signal and the accommodation wave source signal; mark micro-saccade events in the gaze micro-motion source signal and extract the displacement magnitude sequence of the drift between adjacent micro-saccades; at the same time, construct an analytical signal for the accommodation wave source signal and obtain the instantaneous phase sequence. S4. Using the peak point of the instantaneous phase sequence of the regulating wave source signal as a reference, calculate the probability density of the start and end timestamps of the micro-scanning falling in each phase interval, as the phase concentration; perform cyclic coherence operation on the drift displacement magnitude sequence and the instantaneous phase sequence of the regulating wave source signal to obtain the phase-locked value. S5. Calculate the micro-motion-adjustment coupling coordination coefficient of the current gaze window by measuring the distance between the phase concentration and the phase-locked value in the two-dimensional normalized space. S6. When the micro-motion-adjustment coupling coordination coefficient is continuously lower than the preset farsighted defocus coordination threshold within the continuous sliding time window, the start and end timestamps of the corresponding fixation interval are mapped to the timeline of the currently playing interactive content, generating a high-risk eye use segment label bound to the content location.

[0007] As a further aspect of the present invention, in step S1, generating the acquisition trigger signal specifically includes: Based on the coordinate range of the display area of ​​the animation currently playing on the display device, the user's gaze point is tracked and detected at a preset frequency at low speed. When the gaze point is located within the display area of ​​the animation playback interface, the continuous gaze point trajectory within the preset sliding time window is extracted, and a spatial dispersion sequence is formed based on the degree of discrete distribution of the gaze point trajectory in the two-dimensional plane. Perform trend analysis on the spatial dispersion sequence to obtain the dispersion change rate sequence. When the dispersion change rate changes from negative to positive and the subsequent continuous fluctuation range does not exceed the preset stability threshold, the corresponding time is marked as a local stable time and a data acquisition trigger signal is generated.

[0008] As a further aspect of the present invention, in step S2, forming the original signal stream with timestamp alignment specifically includes: When a trigger signal is detected, the system uses an infrared light source to alternately illuminate the left and right eye illumination modules to capture the corresponding corneal reflection image in a time-division manner. Pupil boundary fitting is performed on the captured left and right eye images respectively, the major and minor axis parameters of the pupil ellipse are extracted, and the geometric mean of the major and minor axes is used as the current pupil diameter. For corneal reflection images at the same timestamp, perform bright spot centroid localization and use the vector difference between the coordinates of the bright spot centroid and the coordinates of the pupil ellipse center as the corneal reflection offset vector; The pupil diameter, corneal reflection offset vector, and corresponding timestamp are combined into a raw signal frame and arranged in chronological order to form a timestamp-aligned raw signal stream.

[0009] As a further aspect of the present invention, in step S3, extracting the displacement modulus sequence of the drift between adjacent microscanning intervals specifically includes: Using the corneal reflection offset vector sequence and pupil diameter sequence in the original signal stream as input, independent component separation processing is performed to obtain the gaze micro-motion source signal and the accommodation fluctuation source signal; In the observation of micro-motion source signals, the starting time of the micro-saccade event is marked when the signal amplitude in the sliding window exceeds the preset rate of change threshold, and the ending time of the micro-saccade event is marked when the amplitude falls back to below the rate of change threshold. The time interval between the end of the previous microsaccade and the start of the next microsaccade is defined as the adjacent microsaccade interval. The displacement vector is accumulated point by point for the gaze micro-motion source signal within the saccade interval. The magnitude of the accumulated displacement vector is calculated as the displacement magnitude of the drift during the saccade interval. The displacement magnitudes corresponding to each adjacent microsaccade interval are arranged in chronological order to form a displacement magnitude sequence.

[0010] As a further aspect of the present invention, in step S3, the instantaneous phase sequence is obtained by performing a Hilbert transform on the regulating wave source signal to obtain an orthogonal signal, and the phase angle sequence of the complex sequence formed by the regulating wave source signal and the orthogonal signal is taken as the instantaneous phase sequence.

[0011] As a further aspect of the present invention, in S4, the phase concentration specifically includes: The phase interval between each peak point and the previous peak point in the instantaneous phase sequence is divided into a preset number of phase sub-intervals; Traverse the start and end timestamps of microscanning events, assign each microscanning start and end timestamp to the corresponding phase sub-interval according to its corresponding instantaneous phase value, count the frequency of start and end timestamps appearing in each phase sub-interval, divide by the total number of start and end timestamps of all microscanning events, and obtain the probability density of each phase sub-interval. A discrete probability density distribution is constructed with the phase sub-interval index as the horizontal axis and the probability density as the vertical axis. The reciprocal of the variance of the probability density distribution is calculated as the phase concentration.

[0012] As a further aspect of the present invention, obtaining the phase-locked value in step S4 specifically includes: When performing cyclic coherence operation on the displacement modulus sequence and the instantaneous phase sequence, the displacement modulus sequence and the instantaneous phase sequence are aligned with the same time index and divided into several data segments of equal length. The cross spectrum of the displacement modulus sequence and the instantaneous phase sequence in each segment is calculated. The cross spectrum of each segment is averaged, and the maximum amplitude of the average cross spectrum within the preset adjustment fluctuation frequency band is extracted as the phase-locked value.

[0013] As a further aspect of the present invention, in step S5, calculating the micro-motion-adjustment coupling coordination coefficient of the current gaze window specifically includes: Obtain the phase concentration and phase-locking value corresponding to the current gaze window and a preset number of previously processed gaze windows; The phase concentration of the processed gaze windows is used to form a first sequence, and the phase-locked values ​​are used to form a second sequence. The mean and standard deviation of the first and second sequences are calculated respectively. The standardized phase concentration is obtained by subtracting the mean of the first sequence from the phase concentration of the current gaze window and then dividing by the standard deviation of the first sequence. The standardized phase-locked value is obtained by subtracting the mean of the second sequence from the phase-locked value of the current gaze window and then dividing by the standard deviation of the second sequence. Two-dimensional points are constructed with standardized phase concentration as the abscissa and standardized phase-locked value as the ordinate. The reciprocal of the Euclidean distance from the corresponding two-dimensional point to the set reference coupling point is calculated as the micro-motion-adjustment coupling coordination coefficient.

[0014] As a further aspect of the present invention, in step S6, generating high-risk eye-use segment annotation tags bound to content locations specifically includes: The micro-motion-adjustment coupling coordination coefficients are arranged in the order of timestamps to form a coefficient sequence. A sliding window of fixed length is taken and moved along the time axis with a fixed step size. A low coordination event is recorded when all coefficient values ​​in the window are lower than the farsighted defocus coordination threshold at each window position. When the cumulative number of consecutive low-coordination events reaches the preset number of consecutive judgments, the earliest and latest timestamps covered by the consecutive low-coordination events are backtracked and used as the start and end timestamps of the gaze interval, respectively. The system queries the global time base of the interactive content currently being played on the display device, establishes a linear mapping relationship between the start and end timestamps and the timeline of the interactive content, and extracts the identifiers of the interactive content segments defined by the mapping relationship as high-risk eye-use segments.

[0015] The technical effects and advantages of the eye-tracking interactive method for myopia prevention and control analysis in children according to the present invention are as follows: This invention dynamically detects the stability of a child's gaze during the viewing of animation and interactive content. After a stable gaze is established, it simultaneously collects corneal reflection shift information and pupillary change information. Based on the coupling relationship between gaze micro-motion changes and accommodation-related fluctuations, it performs joint analysis, enabling the detection of microscopic visual accommodation abnormalities that are difficult to identify in traditional eye use monitoring. By extracting microsaccadic inter-period drift features, instantaneous phase features, and phase-locked features, a micro-motion-accommodation coupling coordination coefficient is constructed, which can reflect the trend of visual accommodation coordination changes in children during continuous viewing, improving the accuracy and continuity of identifying high-risk eye use states.

[0016] Meanwhile, this invention further maps low-coordination gaze intervals to the timeline of interactive content, realizing the correspondence between high-risk eye-use segments and specific animation content locations. This can accurately locate content segments that are likely to induce sustained visual tension, providing more refined data support for content optimization, viewing behavior intervention, and visual health analysis in children's myopia prevention and control. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of an eye-tracking interactive method for myopia prevention and control analysis in children, according to the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1

[0020] Figure 1 This invention provides an eye-tracking interactive method for myopia prevention and control analysis in children, which includes the following steps: S1. When the user's gaze falls on the animation playback interface of the display device, and the spatial dispersion of the gaze point decreases and enters a stable state, a collection trigger signal is generated. S2. Upon receiving the acquisition trigger signal, continuously capture the pupil images and corneal reflection images of both eyes, and real-time corneal reflection offset vector and pupil diameter to form a raw signal stream with timestamp alignment; S3. Perform independent component separation on the original signal stream to obtain the gaze micro-motion source signal and the accommodation wave source signal; mark micro-saccade events in the gaze micro-motion source signal and extract the displacement magnitude sequence of the drift between adjacent micro-saccades; at the same time, construct an analytical signal for the accommodation wave source signal and obtain the instantaneous phase sequence. S4. Using the peak point of the instantaneous phase sequence of the regulating wave source signal as a reference, calculate the probability density of the start and end timestamps of the micro-scanning falling in each phase interval, as the phase concentration; perform cyclic coherence operation on the drift displacement magnitude sequence and the instantaneous phase sequence of the regulating wave source signal to obtain the phase-locked value. S5. Calculate the micro-motion-adjustment coupling coordination coefficient of the current gaze window by measuring the distance between the phase concentration and the phase-locked value in the two-dimensional normalized space. S6. When the micro-motion-adjustment coupling coordination coefficient is continuously lower than the preset farsighted defocus coordination threshold within the continuous sliding time window, the start and end timestamps of the corresponding fixation interval are mapped to the timeline of the currently playing interactive content, generating a high-risk eye use segment label bound to the content location.

[0021] In step S1, a data acquisition trigger signal is generated.

[0022] After the animation begins playing, the display device first reads the actual display area of ​​the current animation playback region within the display interface and converts this display area into the corresponding screen coordinate area. For full-screen playback scenarios, the entire display area is directly used as the effective animation area; for split-screen playback, floating playback, or playback scenarios with a menu bar, the actual video frame boundary is extracted as the animation interface display area, thus avoiding misidentification of the menu bar, subtitle edges, progress bar area, or blank borders as effective viewing areas. During animation playback, the display device performs low-speed tracking detection of gaze points at a preset frequency. The low-speed tracking detection uses a low-frame-rate infrared eye-tracking pre-detection method, and the detection frequency is set according to the normal eye movement speed of children. In this embodiment, the low-speed tracking detection frequency is set to 18 to 25 times per second to ensure that the stable convergence process of the gaze can be identified, while reducing the power consumption and invalid data accumulation caused by continuous high-speed acquisition. During the detection process, the display device periodically acquires the coordinates of the gaze point corresponding to the current moment and determines whether the gaze point is located within the display area of ​​the animation interface. When the gaze point is within the animation interface for multiple consecutive sampling periods, the continuous gaze point trajectory within the corresponding time period is extracted. To avoid accidental triggering caused by short-term saccades, blinking, head tilting, or posture adjustments in children, this embodiment uses a sliding time window method to extract continuous gaze trajectories. The length of the sliding time window is set based on the time required for a child to form a normal stable gaze. In this embodiment, the sliding time window length is set to 0.6 seconds, and the sliding step size is set to 0.1 seconds. Subsequently, the discrete distribution of the coordinates of all gaze points within the sliding time window in the two-dimensional plane is calculated to form a spatial dispersion sequence. The spatial dispersion does not use a single coordinate deviation, but rather simultaneously counts the overall distribution range of the gaze trajectory in both the horizontal and vertical directions to avoid interference from small unidirectional jitters on stability determination. In specific processing, the discrete range of the horizontal coordinate and the discrete range of the vertical coordinate of all gaze points within the sliding time window are calculated separately, and then the discrete results in the two directions are jointly statistically analyzed to obtain the spatial dispersion corresponding to the current moment.

[0023] A continuous trend analysis is performed on the spatial dispersion sequence to obtain the corresponding dispersion change rate sequence. During the trend analysis, the direction of spatial dispersion change corresponding to adjacent sampling times is compared in chronological order, and the continuous change results are recorded. When the child's gaze gradually focuses on a fixed target in the animation, the spatial dispersion will continuously decrease, and the dispersion change rate will be continuously negative. When the gaze has converged and entered a stable fixation phase, the spatial dispersion decrease process ends, and the dispersion change rate gradually approaches zero and turns positive. In this embodiment, when the dispersion change rate is detected to change from negative to positive, the corresponding time is initially determined as a local stable candidate time. To avoid misjudgment by children in short pauses, momentary fixations, or occasional low fluctuations, the stability of the continuous fluctuation range after the candidate time is further verified. Specifically, starting from the candidate time, the spatial dispersion values ​​are extracted for multiple consecutive sampling periods, and the maximum fluctuation range of spatial dispersion during this period is calculated. When the continuous fluctuation range does not exceed a preset stability threshold, it is determined that the current gaze has entered a stable fixation state, and the corresponding time is officially marked as a local stable time. The preset stability threshold is set based on the range of minute eye movement fluctuations during a child's normal gaze. By pre-collecting stable gaze samples from 20 children under standard reading conditions, the 95th percentile of the spatial dispersion fluctuation range during the stable gaze phase is statistically analyzed. This statistical result is used as the stability threshold, ensuring that the threshold is based on actual children's viewing behavior. After confirming the local stability moment, a data acquisition trigger signal is generated.

[0024] In S2, a timestamp-aligned original signal stream is formed.

[0025] Upon detecting the acquisition trigger signal, the display device immediately switches to high-speed synchronous acquisition mode and initiates the binocular time-sharing exposure acquisition process. In this embodiment, independently arranged infrared illumination areas are set inside the display device's frame, with the left infrared illumination area corresponding to the left eye illumination module and the right infrared illumination area corresponding to the right eye illumination module. To avoid superimposed interference from corneal reflection bright spots of the left and right eyes within the same exposure cycle, the infrared light source employs an alternating illumination method for time-sharing exposure. Specifically, in the first exposure cycle, only the left eye illumination module is illuminated, while the right eye illumination module is turned off, allowing the image sensor to acquire only the corneal reflection image corresponding to the left eye region; in the next exposure cycle, the left eye illumination module is turned off, and the right eye illumination module is illuminated, allowing the image sensor to acquire the corneal reflection image corresponding to the right eye region. By using the alternating exposure method for the left and right eyes, the problem of bright spot overlap and reflection crosstalk caused by the simultaneous appearance of corneal reflection bright spots on both sides in the same frame image is avoided. Considering that children may slightly move their heads and blink momentarily while watching animation, this embodiment uses a short-exposure continuous acquisition method for the image sensor to ensure that each frame remains sufficiently clear during the exposure period, avoiding motion blur that could distort the pupil boundary. After acquiring the left and right eye images, pupil boundary fitting processing is performed on each image separately. Specifically, infrared grayscale enhancement is first applied to the image to enhance the brightness difference between the pupil and iris regions, and then the pupil edge contour is extracted based on the brightness gradient change. Due to eyelid occlusion and eyelash interference during children's viewing, this embodiment does not directly use a circular boundary but instead uses an elliptical boundary for fitting. During the fitting process, continuous boundary filtering is performed on the extracted edge contour points to remove isolated noise points and short boundary fragments. Elliptical parameter fitting is then performed on the remaining boundary points to obtain the major and minor axis parameters corresponding to the pupil ellipse. Considering that the pupil image is prone to perspective distortion when a child's head is slightly turned, this embodiment uses the geometric mean of the major and minor axes as the current pupil diameter, thereby reducing the impact of unidirectional stretching deformation on the pupil size measurement results. A stability screening is performed on the pupil boundary results of consecutive frames. When the changes in the major and minor axes between consecutive frames exceed a preset range, the current frame is identified as a boundary distortion frame and discarded. The preset range is set based on the natural rate of pupil change during normal viewing of animation by children. In this embodiment, a statistical distribution is established by pre-collecting normal viewing samples of children, and the 95% stable range is used as the basis for boundary distortion screening.

[0026] After pupil boundary fitting, bright spot centroid localization processing is performed on corneal reflection images acquired at the same time stamp. In this embodiment, infrared illumination forms a stable high-brightness reflection area on the corneal surface, and the brightness of the corresponding area in the image is significantly higher than that of the surrounding iris area. In the specific processing, brightness threshold segmentation is first performed on the corneal reflection image to extract continuous areas with brightness higher than the set bright spot threshold, and area filtering is performed on the continuous areas to remove noise bright spots and edge reflective areas with too small area. The bright spot threshold is set according to the infrared illumination power and the dynamic range of the image sensor. In this embodiment, corneal reflection images under different ambient brightness are pre-acquired to statistically analyze the brightness distribution of the real corneal bright spot area, and the gray value corresponding to the stable clustering interval in the brightness distribution is used as the bright spot segmentation threshold. After the bright spot area is extracted, brightness-weighted centroid calculation is performed on all pixels inside the bright spot area to obtain the current bright spot centroid coordinates. At the same time, the center coordinates of the pupil ellipse at the corresponding time are read, and the vector difference between the bright spot centroid coordinates and the pupil ellipse center coordinates is calculated. This vector difference is used as the corneal reflection offset vector. Because children may move their heads slightly during viewing, the absolute coordinates of the bright spot are not used directly. Instead, the relative offset between the bright spot and the pupil center is used to reduce the impact of overall head displacement on the visual analysis results. After calculating the offset vector, the pupil diameter, corneal reflection offset vector, and acquisition timestamp at the current moment are combined into a single raw signal frame. The timestamp is generated using a unified hardware clock, and the left and right eye images and corresponding bright spot information are marked using the same clock source to avoid phase shifts in subsequent signals caused by misalignment of acquisition times between the left and right eyes. Subsequently, all raw signal frames are arranged in chronological order to form a raw signal stream with consecutive timestamps aligned. To avoid signal interruptions caused by children's momentary occlusion, rapid head turning, or continuous blinking, this embodiment performs continuity detection on abnormal jump frames in the raw signal stream. When the interval between consecutive timestamps exceeds twice the set acquisition period, the corresponding time period is marked as an acquisition interruption interval.

[0027] In step S3, the displacement modulus sequence of the drift between adjacent microscanning intervals is extracted.

[0028] After the original signal stream is formed, the corneal reflection offset vector sequence and pupil diameter sequence are read in a unified timestamp order, and synchronization time alignment processing is performed on the two types of sequences. Mean removal processing is performed on the horizontal component, vertical component, and pupil diameter sequence of the corneal reflection offset vector sequence, and the amplitude of each sequence is scaled to a unified numerical range. After normalization, the horizontal offset component, vertical offset component, and pupil diameter sequence are combined to form a joint input matrix, and independent component separation operation is performed on the joint input matrix. Specifically, covariance whitening processing is first performed on the joint input matrix to reduce the correlation between the input components. Then, the unmixing matrix is ​​updated iteratively. In each iteration, the statistical independence between the output components is calculated. When the change in independence between two consecutive iterations is less than a set convergence threshold, the iteration stops and the corresponding independent component is output. The convergence threshold is set based on the stable range of the change in independence during continuous iterations. In this embodiment, eye movement samples of children watching animation normally are pre-collected, the distribution of the change in independence between two adjacent iterations in the stable iteration stage is statistically analyzed, and the value corresponding to the 95% stable interval is used as the convergence threshold. After independent component separation, multiple candidate fluctuation components are obtained. Subsequently, frequency and continuous fluctuation statistics are performed on each candidate fluctuation component. Specifically, the number of local peak occurrences, the average time interval between adjacent peaks, and the duration of continuous fluctuations are counted for each candidate fluctuation component within a unit time. When the number of local peaks of a candidate fluctuation component within a consecutive 1-second time window exceeds a preset peak count threshold, and the time interval between adjacent peaks is consistently less than a preset interval threshold, the corresponding candidate fluctuation component is identified as a gaze micro-motion source signal. The preset peak count threshold is set based on the statistical results of micro-saccade events under normal stable gaze conditions in children. In this embodiment, normal viewing samples are pre-collected to statistically analyze the distribution of the number of local peaks related to micro-saccades within a unit time under stable gaze conditions, and the upper boundary of the stable clustering interval in the distribution is used as the peak count threshold. The adjacent peak time interval threshold is set based on the statistical results of the duration of normal micro-saccades in children. In this embodiment, the statistical median corresponding to the average interval of normal micro-saccade events is used as the interval threshold. Meanwhile, when a candidate fluctuation component exhibits periodic continuous fluctuations across multiple consecutive time windows, and the time interval between adjacent peaks does not exceed a set periodic stability threshold, the corresponding candidate fluctuation component is identified as the accommodative fluctuation source signal. The periodic stability threshold is set based on the continuous oscillation period distribution during a child's normal accommodative activity. In this embodiment, pupil change samples are pre-collected during normal animation viewing to statistically analyze the range of continuous oscillation period changes. The fluctuation range corresponding to the stable distribution interval is used as the periodic stability threshold, thereby preventing short-term random noise from being misidentified as an accommodative fluctuation signal. After completing the gaze micro-motion source signal identification, micro-salivation event detection is performed on this signal.In the specific processing, a fixed-length sliding window is continuously moved along the time axis, and the displacement change between adjacent sampling points is calculated within each sliding window. When the displacement change of multiple consecutive sampling points exceeds the preset rate of change threshold, the corresponding time is marked as the start time of the micro-scanning event; when the displacement change of multiple subsequent consecutive sampling points falls below the rate of change threshold again, the corresponding time is marked as the end time of the micro-scanning event.

[0029] After completing the microsaccade event detection, the continuous time period between the end of the previous microsaccade event and the start of the next microsaccade event is defined as the adjacent microsaccade interval. Subsequently, continuous displacement accumulation processing is performed on the gaze micro-motion source signal within each saccade interval. The corresponding displacement vector is read point by point in chronological order, and vector accumulation is performed sequentially to obtain the cumulative displacement trajectory corresponding to the current saccade interval. Two-dimensional displacement direction information is retained, and after all displacement vectors have been accumulated, the magnitude corresponding to the final cumulative displacement vector is calculated. This magnitude is used as the displacement magnitude of the drift in the current saccade interval. Then, the displacement magnitudes corresponding to each saccade interval are arranged in chronological order to form a displacement magnitude sequence. To avoid abnormally large drift magnitudes caused by children's head posture adjustments, Hilbert transform processing is performed on the aforementioned accommodation wave source signal. Specifically, for each sampling point of the accommodation wave source signal, a corresponding orthogonal signal is generated, and then the accommodation wave source signal is combined with the corresponding orthogonal signal to form a complex analytical sequence. Subsequently, the phase angle changes corresponding to the complex analytical sequence are read in chronological order, and continuous phase angles are arranged to form an instantaneous phase sequence.

[0030] In step S4, the phase-locked value is obtained.

[0031] Phase interval division processing is performed on the instantaneous phase sequence corresponding to the adjustment fluctuation source signal. Since the adjustment fluctuation source signal exhibits periodic oscillation changes during continuous animation viewing, the phase interval is dynamically divided according to the continuous oscillation period. Specifically, the local phase peak points in the instantaneous phase sequence are traversed, and the continuous phase change interval between the current peak point and the previous peak point is defined as a complete adjustment cycle. The phase interval corresponding to each complete adjustment cycle is divided into a preset number of phase sub-intervals. The number of phase sub-intervals is set according to the phase change resolution requirements within the adjustment fluctuation cycle; specifically, 16 phase sub-intervals are used to divide each complete adjustment cycle, ensuring that the distribution differences of micro-scanning events in different phase stages can be distinguished. Subsequently, the start and end timestamps of all micro-scanning events obtained above are traversed, and the corresponding timestamp is assigned to its respective phase sub-interval based on the instantaneous phase value corresponding to each timestamp. Specifically, the corresponding sampling point is first located in the instantaneous phase sequence based on the timestamp, then the phase value of the corresponding sampling point is read, and it is determined which phase sub-interval the current phase value is located within. After classification, the number of start and end timestamps of micro-saccades within each phase sub-interval is counted. Since local noise or short-term abnormal movements during animation viewing can cause abnormally dense micro-saccade events, continuous short-interval repetitive micro-saccade events are further merged. When the time interval between two adjacent micro-saccade events is less than a preset merging interval, the corresponding events are merged into the same continuous micro-saccade event before being included in frequency statistics. The preset merging interval is set based on the statistical results of the duration of normal micro-saccade events in children. By statistically analyzing the interval distribution of normal micro-saccade events, the lower boundary of the stable distribution is used as the basis for event merging, thereby avoiding distortion of the phase distribution caused by local abnormal high-frequency fluctuations. After frequency statistics are completed, the frequency of the timestamps corresponding to each phase sub-interval is divided by the total number of start and end timestamps of all micro-saccade events to obtain the probability density of the corresponding phase sub-interval. Subsequently, a discrete probability density distribution is constructed using the phase sub-interval index as the horizontal arrangement order and the corresponding probability density as the vertical statistical value. Considering the differences in the number of saccades among different children watching animation, this embodiment does not directly use frequency values, but instead uses probability density for normalization statistics, thereby reducing the impact of differences in viewing time and total number of events on the phase concentration analysis results. After constructing the probability density distribution, the variance of the discrete distribution is calculated for the probability density corresponding to all phase sub-intervals. When saccades are mainly concentrated in a few phase sub-intervals, the probability density distribution shows obvious concentration, and the distribution variance decreases accordingly; when saccades occur randomly in multiple phase intervals, the probability density distribution tends to be discrete, and the distribution variance increases accordingly. This embodiment further takes the reciprocal of the probability density distribution variance as the phase concentration, so that the more concentrated the saccades are in a fixed phase area, the greater the corresponding phase concentration.

[0032] Cyclic coherence analysis was performed on the displacement modulus sequence and instantaneous phase sequence to extract the lock-in correlation between the microscan drift process and the regulation fluctuation process. First, the displacement modulus sequence and instantaneous phase sequence were aligned using a unified timestamp, and data segments corresponding to missing timestamps or interrupted acquisition intervals were deleted. Then, the time-aligned displacement modulus sequence and instantaneous phase sequence were divided into multiple continuous data segments of fixed length. The segment length was set based on statistical results of normal regulation fluctuation cycles in children, using the time length covering multiple complete regulation cycles as the length of a single data segment. After segmentation, spectral transformation was performed on the displacement modulus sequence and instantaneous phase sequence within each segment, and the cross-spectrum between the two types of sequences was calculated. Averaging was performed on the cross-spectrums corresponding to all data segments to reduce the impact of local anomalies on the lock-in analysis results. Subsequently, the spectral amplitude variation of the average cross-spectrum within the preset regulation fluctuation frequency band was read. The preset regulatory fluctuation frequency band is set based on the statistical results of the oscillation frequency of children's normal regulatory activities. By pre-collecting regulatory fluctuation samples when children continuously watch animation, the dominant frequency distribution corresponding to continuous regulatory oscillation is statistically analyzed, and a stable clustering frequency band is used as the range of the regulatory fluctuation frequency band. After completing the frequency band selection, the maximum amplitude of the average cross spectrum within the corresponding regulatory fluctuation frequency band is extracted as the phase-locked value. When there is a stable synchronous relationship between the change in displacement modulus and the change in regulatory fluctuation, the average cross spectrum will form a significant concentrated peak within the regulatory fluctuation frequency band, and the corresponding phase-locked value will increase; when there is no stable synchronous relationship between the two types of changes, the energy distribution of the average cross spectrum within the corresponding frequency band tends to be dispersed, and the corresponding phase-locked value will decrease.

[0033] In step S5, the micro-motion-adjustment coupling coordination coefficient of the current gaze window is calculated.

[0034] After calculating the phase concentration and phase-locked value, the continuous fixation phase is divided into multiple fixation windows according to time sequence. The phase concentration and phase-locked value corresponding to the current fixation window and a preset number of previously processed fixation windows are then read. Since there are individual differences among children watching animation, a dynamic historical window normalization method is used to standardize the current window. Specifically, the phase concentration corresponding to the historically processed fixation windows is first arranged in chronological order to form a first sequence, and the corresponding phase-locked values ​​are arranged to form a second sequence. The number of historical windows is set based on the statistical requirements for stable fluctuations during normal continuous viewing by children. In this embodiment, the number of historical windows is set to the most recent 30 consecutive effective fixation windows, ensuring that the statistical results reflect the overall fluctuation level of the current viewing phase while avoiding excessive influence of earlier states on the current results due to a long historical span. Subsequently, statistical analysis is performed on the first and second sequences, calculating their respective means and standard deviations. The standardized phase concentration is obtained by subtracting the mean of the first sequence from the phase concentration of the current gaze window and dividing by the standard deviation of the first sequence. Simultaneously, the standardized phase-locked value is obtained by subtracting the mean of the second sequence from the phase-locked value of the current gaze window and dividing by the standard deviation of the second sequence. Considering that some children may experience small fluctuations during continuous viewing over a short period, potentially resulting in a standard deviation close to zero, a lower standard deviation constraint is set. When the standard deviation corresponding to either the first or second sequence is lower than the preset lower limit, the preset lower limit is used instead of the actual standard deviation. The preset lower limit is set based on the minimum fluctuation range of the stable phase in normal viewing samples. In this embodiment, the continuous fluctuation distribution of children during stable viewing phases is statistically analyzed, and the minimum stable fluctuation range in the stable distribution is used as the lower standard deviation. After standardization, a two-dimensional point corresponding to the current gaze window is constructed using the standardized phase concentration as the x-axis and the standardized phase-locked value as the y-axis. Since the phase concentration reflects the concentrated distribution of micro-saccade events in the accommodation phase, and the phase-locked value reflects the degree of synchronous correlation between the drift process and accommodation fluctuations, the two-dimensional point can simultaneously reflect the micro-motion-accommodation coupling state corresponding to the current gaze window.

[0035] After constructing the current 2D point, a reference coupling point is further determined, and the micro-motion-adjustment coupling coordination coefficient corresponding to the current window is calculated based on the reference coupling point. Specifically, all 2D points corresponding to previously processed gaze windows are traversed, and the stability of the phase concentration and phase-locked value for each historical window across multiple consecutive windows is statistically analyzed. Since even if a single window exhibits a large phase concentration and phase-locked value simultaneously, it may originate from local abnormal spikes. Therefore, this embodiment does not directly select the window with the single maximum value as the reference window, but instead uses a continuous stable coordination determination method to determine the historical window with the highest coordination degree. Specifically, for each historical window, the number of times the phase concentration and phase-locked value are simultaneously higher than the corresponding historical average in subsequent consecutive windows is counted. When a historical window maintains a phase concentration and phase-locked value simultaneously higher than the historical average in subsequent consecutive windows, and the fluctuation range of the corresponding 2D point in the consecutive windows does not exceed a preset stable fluctuation threshold, the corresponding historical window is determined to be a stable coordination window. The stability fluctuation threshold is set based on the fluctuation range of two-dimensional points during the continuous stable coordination phase in the normal viewing samples. This is achieved by statistically analyzing the distribution of two-dimensional coordinate changes during the normal stable viewing phase in children, and using the fluctuation range corresponding to the stable clustering interval as the stability threshold. Subsequently, within all stable coordination windows, the offset between the corresponding two-dimensional points and the historical mean center point is statistically analyzed, and the stable coordination window with the longest continuous stable maintenance time is selected as the window with the highest degree of coordination. The two-dimensional point corresponding to this window is determined as the reference coupling point. After determining the reference coupling point, the Euclidean distance between the current two-dimensional point and the reference coupling point is calculated, and the reciprocal of the corresponding Euclidean distance is used as the micro-motion-accommodation coupling coordination coefficient corresponding to the current gaze window. When the current window's two-dimensional point is close to the reference coupling point, it indicates that the current micro-saccade distribution state and accommodation fluctuation state are close to the historical stable coordination state, and the corresponding coordination coefficient increases; when the current window's two-dimensional point deviates from the reference coupling point, it indicates that the synchronization relationship between the current micro-motion and accommodation decreases, and the corresponding coordination coefficient decreases.

[0036] In step S6, high-risk eye-use segment labels are generated and bound to the content location.

[0037] After calculating the micro-motion-adjustment coupling coordination coefficient, the coordination coefficients corresponding to the continuous gaze phase are arranged according to the timestamp order to form a continuous coefficient sequence. A sliding window is divided according to a fixed time length and moves continuously along the time axis with a fixed step size. The sliding window length is set based on the continuous change time of the child's accommodation activity during continuous gaze. In this embodiment, by pre-collecting samples of coordination coefficient changes during continuous animation viewing, the duration distribution of continuous low coordination states is statistically analyzed, and the time length corresponding to stable low coordination states is used as the sliding window length. The sliding step size is set according to the resolution of coordination state changes. In this embodiment, a fixed time interval smaller than the sliding window length is used as the window movement step size to ensure continuous overlapping areas between adjacent windows, avoiding the omission of local short-term low coordination states during window switching. Subsequently, all coordination coefficient values ​​within the corresponding time period are read window by window, and it is determined whether the coordination state in the current window is continuously lower than the hyperopic defocus coordination threshold. The hyperopic defocus coordination threshold is set based on the statistical results of coordination coefficients under normal stable viewing conditions in children. In this embodiment, the distribution of coordination coefficients during normal animation viewing without significant visual fatigue is pre-collected. The lower boundary of the coordination coefficients during stable viewing is statistically determined, and the corresponding statistical results are used as the farsighted defocus coordination threshold. When all coordination coefficients within a window remain below the farsighted defocus coordination threshold, the current window is recorded as a low coordination event. Since children may experience localized brief recovery during continuous animation viewing, the intervals between adjacent low coordination windows are continuously merged. When the time interval between two low coordination windows is less than a preset continuous interval threshold, the corresponding windows are merged into the same continuous low coordination event.

[0038] When the cumulative number of consecutive low-coordination events reaches the preset consecutive judgment count, the corresponding time period is determined as a sustained fixation interval for hyperopic defocus type. The preset consecutive judgment count is set based on the formation process of a child's sustained visual tension state. Subsequently, backtracking is performed on all time intervals covered by the current consecutive low-coordination events, reading the earliest and latest timestamps corresponding to the consecutive low-coordination events, and using them as the start and end timestamps of the current fixation interval, respectively. Since the display device continuously records the correspondence between the playback progress time and the system acquisition time during animation playback, the global time reference information corresponding to the current interactive content is further read. In specific processing, the system timestamp of the playback start and the starting position of the animation timeline are recorded synchronously when the animation starts playing, and the time offset changes corresponding to pause, drag, speed up playback, and jump operations are continuously recorded during playback. When the user is detected to perform drag playback, pause and resume, or speed up switching, the correspondence table between the system time and the animation timeline is immediately updated to ensure that subsequent time mapping always corresponds to the actual playback position. After completing the global time base reading, the previously obtained start and end timestamps are mapped to the corresponding position intervals in the current animation content timeline, and the interactive content segment identifiers within the corresponding time intervals are extracted as high-risk eye strain segments. The interactive content segment identifiers are determined based on the pre-defined segment numbers of the animation content. During the resource loading phase, the animation content is segmented according to scene transition points, camera change points, and the boundaries of continuous interactive segments, and a unique segment identifier number is generated for each segment. After completing the time mapping, the animation segment numbers covered by the corresponding time interval are read, and the corresponding segments are marked as high-risk eye strain segments. To avoid segment positioning offsets caused by time mapping boundary errors, a neighboring frame check is further performed on the animation timeline positions corresponding to the start and end timestamps. When the time mapping position is near the boundary of two animation segments, the playback time information corresponding to several consecutive frames before and after the boundary is read, and the final segment assignment is re-determined based on the time coverage ratio.

[0039] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0040] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0041] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0042] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0043] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0044] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0045] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0046] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0047] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An eye-tracking interactive method for myopia prevention and control analysis in children, characterized in that, Includes the following steps: S1. When the user's gaze falls on the animation playback interface of the display device, and the spatial dispersion of the gaze point decreases and enters a stable state, a collection trigger signal is generated. S2. Upon receiving the acquisition trigger signal, continuously capture the pupil images and corneal reflection images of both eyes, and real-time corneal reflection offset vector and pupil diameter to form a raw signal stream with timestamp alignment; S3. Perform independent component separation on the original signal stream to obtain the gaze micro-motion source signal and the accommodation wave source signal; mark micro-saccade events in the gaze micro-motion source signal and extract the displacement magnitude sequence of the drift between adjacent micro-saccades; at the same time, construct an analytical signal for the accommodation wave source signal and obtain the instantaneous phase sequence. S4. Using the peak point of the instantaneous phase sequence of the regulating wave source signal as a reference, calculate the probability density of the start and end timestamps of the micro-scanning falling in each phase interval, as the phase concentration; perform cyclic coherence operation on the drift displacement magnitude sequence and the instantaneous phase sequence of the regulating wave source signal to obtain the phase-locked value. S5. Calculate the micro-motion-adjustment coupling coordination coefficient of the current gaze window by measuring the distance between the phase concentration and the phase-locked value in the two-dimensional normalized space. S6. When the micro-motion-adjustment coupling coordination coefficient is continuously lower than the preset farsighted defocus coordination threshold within the continuous sliding time window, the start and end timestamps of the corresponding fixation interval are mapped to the timeline of the currently playing interactive content, generating a high-risk eye use segment label bound to the content location.

2. The eye-tracking interactive method for myopia prevention and control analysis in children according to claim 1, characterized in that, In step S1, generating the acquisition trigger signal specifically includes: Based on the coordinate range of the display area of ​​the animation currently playing on the display device, the user's gaze point is tracked and detected at a preset frequency at low speed. When the gaze point is located within the display area of ​​the animation playback interface, the continuous gaze point trajectory within the preset sliding time window is extracted, and a spatial dispersion sequence is formed based on the degree of discrete distribution of the gaze point trajectory in the two-dimensional plane. Perform trend analysis on the spatial dispersion sequence to obtain the dispersion change rate sequence. When the dispersion change rate changes from negative to positive and the subsequent continuous fluctuation range does not exceed the preset stability threshold, the corresponding time is marked as a local stable time and a data acquisition trigger signal is generated.

3. The eye-tracking interactive method for myopia prevention and control analysis in children according to claim 1, characterized in that, In step S2, forming the original signal stream with timestamp alignment specifically includes: When a trigger signal is detected, the system uses an infrared light source to alternately illuminate the left and right eye illumination modules to capture the corresponding corneal reflection image in a time-division manner. Pupil boundary fitting is performed on the captured left and right eye images respectively, the major and minor axis parameters of the pupil ellipse are extracted, and the geometric mean of the major and minor axes is used as the current pupil diameter. For corneal reflection images at the same timestamp, perform bright spot centroid localization and use the vector difference between the coordinates of the bright spot centroid and the coordinates of the pupil ellipse center as the corneal reflection offset vector; The pupil diameter, corneal reflection offset vector, and corresponding timestamp are combined into a raw signal frame and arranged in chronological order to form a timestamp-aligned raw signal stream.

4. The eye-tracking interactive method for myopia prevention and control analysis in children according to claim 1, characterized in that, In step S3, extracting the displacement modulus sequence of the drift between adjacent microscanning intervals specifically includes: Using the corneal reflection offset vector sequence and pupil diameter sequence in the original signal stream as input, independent component separation processing is performed to obtain the gaze micro-motion source signal and the accommodation fluctuation source signal; In the observation of micro-motion source signals, the starting time of the micro-saccade event is marked when the signal amplitude in the sliding window exceeds the preset rate of change threshold, and the ending time of the micro-saccade event is marked when the amplitude falls back to below the rate of change threshold. The time interval between the end of the previous microsaccade and the start of the next microsaccade is defined as the adjacent microsaccade interval. The displacement vector is accumulated point by point for the gaze micro-motion source signal within the saccade interval. The magnitude of the accumulated displacement vector is calculated as the displacement magnitude of the drift during the saccade interval. The displacement magnitudes corresponding to each adjacent microsaccade interval are arranged in chronological order to form a displacement magnitude sequence.

5. The eye-tracking interactive method for myopia prevention and control analysis in children according to claim 4, characterized in that, In step S3, the instantaneous phase sequence is obtained by performing a Hilbert transform on the regulating wave source signal to obtain an orthogonal signal, and the phase angle sequence of the complex sequence formed by the regulating wave source signal and the orthogonal signal is taken as the instantaneous phase sequence.

6. The eye-tracking interactive method for myopia prevention and control analysis in children according to claim 1, characterized in that, In S4, the phase concentration specifically includes: The phase interval between each peak point and the previous peak point in the instantaneous phase sequence is divided into a preset number of phase sub-intervals; Traverse the start and end timestamps of microscanning events, assign each microscanning start and end timestamp to the corresponding phase sub-interval according to its corresponding instantaneous phase value, count the frequency of start and end timestamps appearing in each phase sub-interval, divide by the total number of start and end timestamps of all microscanning events, and obtain the probability density of each phase sub-interval. A discrete probability density distribution is constructed with the phase sub-interval index as the horizontal axis and the probability density as the vertical axis. The reciprocal of the variance of the probability density distribution is calculated as the phase concentration.

7. The eye-tracking interactive method for myopia prevention and control analysis in children according to claim 1, characterized in that, In step S4, obtaining the phase-locked value specifically includes: When performing cyclic coherence operation on the displacement modulus sequence and the instantaneous phase sequence, the displacement modulus sequence and the instantaneous phase sequence are aligned with the same time index and divided into several data segments of equal length. The cross spectrum of the displacement modulus sequence and the instantaneous phase sequence in each segment is calculated. The cross spectrum of each segment is averaged, and the maximum amplitude of the average cross spectrum within the preset adjustment fluctuation frequency band is extracted as the phase-locked value.

8. The eye-tracking interactive method for myopia prevention and control analysis in children according to claim 1, characterized in that, In step S5, calculating the micro-motion-adjustment coupling coordination coefficient of the current gaze window specifically includes: Obtain the phase concentration and phase-locking value corresponding to the current gaze window and a preset number of previously processed gaze windows; The phase concentration of the processed gaze windows is used to form a first sequence, and the phase-locked values ​​are used to form a second sequence. The mean and standard deviation of the first and second sequences are calculated respectively. The standardized phase concentration is obtained by subtracting the mean of the first sequence from the phase concentration of the current gaze window and then dividing by the standard deviation of the first sequence. The standardized phase-locked value is obtained by subtracting the mean of the second sequence from the phase-locked value of the current gaze window and then dividing by the standard deviation of the second sequence. Two-dimensional points are constructed with standardized phase concentration as the abscissa and standardized phase-locked value as the ordinate. The reciprocal of the Euclidean distance from the corresponding two-dimensional point to the set reference coupling point is calculated as the micro-motion-adjustment coupling coordination coefficient.

9. The eye-tracking interactive method for myopia prevention and control analysis in children according to claim 1, characterized in that, In step S6, generating high-risk eye-use segment annotation labels that are bound to the content location specifically includes: The micro-motion-adjustment coupling coordination coefficients are arranged in the order of timestamps to form a coefficient sequence. A sliding window of fixed length is taken and moved along the time axis with a fixed step size. A low coordination event is recorded when all coefficient values ​​in the window are lower than the farsighted defocus coordination threshold at each window position. When the cumulative number of consecutive low-coordination events reaches the preset number of consecutive judgments, the earliest and latest timestamps covered by the consecutive low-coordination events are backtracked and used as the start and end timestamps of the gaze interval, respectively. The system queries the global time base of the interactive content currently being played on the display device, establishes a linear mapping relationship between the start and end timestamps and the timeline of the interactive content, and extracts the identifiers of the interactive content segments defined by the mapping relationship as high-risk eye-use segments.

Images