A method and apparatus for assessing vision health

Abstract

This application provides a method and device for vision health assessment. The method includes: Step S1, collecting eye movement data of the subject observing a visual target displayed on a vision test screen, with the visual target displayed at different positions on the screen for a preset duration; Step S2, calculating multiple preset vision status indicators based on the eye movement data, including fixation stability indicators, visual response indicators, pupillary reaction time indicators, fixation-state pupillary amplitude index indicators, and fixation-state pupillary variability index indicators; Step S3, inputting the preset vision status indicators into a trained binary classification model to obtain the subject's vision health assessment results. The technical solution of this application improves the accuracy and reliability of vision testing, not only capturing minute fixation fluctuations but also revealing the patterns of eye movements, enhancing the sensitivity of early screening, and providing a reference for clinical intervention.

Description

Technical Field

[0001] This invention relates to the fields of visual data processing and vision health screening technology, specifically to a vision health assessment method and device. Background Technology

[0002] Amblyopia is a type of functional visual impairment related to visual development, clinically characterized primarily by reduced best-corrected visual acuity and binocular vision dysfunction. Current amblyopia eye-tracking analysis often suffers from the following problems: First, it relies on a single assessment dimension. Some amblyopia screening and diagnosis methods over-rely on static visual acuity charts, which only reflect the static resolution of the macula and fail to cover tracking and control impairments in amblyopic patients. This single assessment method easily leads to misdiagnosis of borderline amblyopia or visual developmental delay. Second, it lacks objective quantitative indicators of visual dynamics. Clinical observations show micro-tremors or drift in amblyopic eyes, but current technology lacks the ability to accurately quantify the spatial dispersion of fixation. Furthermore, assessments in dynamic environments often rely on physician observation, lacking detailed analysis of eye movement trajectories, resulting in assessment results lacking objective dynamic evidence. Summary of the Invention

[0003] In view of this, this application provides a method and device for assessing visual health. The technical solution of this application provides multi-dimensional visual status indicators, breaking through the limitations of traditional methods that rely on subjective human judgment or single-dimensional static indicators, and providing a new solution for the early screening and clinical assessment of amblyopia and visual function abnormalities in children.

[0004] Specifically, this application is implemented through the following technical solution: According to a first aspect of the embodiments of this specification, a method for assessing vision health is provided, comprising: Step S1: Collect eye movement data of the subject observing the visual targets displayed on the vision test screen. The visual targets are displayed for a preset duration at different display positions on the vision test screen. Step S2: Calculate multiple preset visual acuity state indices based on the eye movement data. These indices include fixation stability, visual response, pupillary reaction time, pupillary amplitude index, and pupillary variability index. Specifically: the fixation stability index characterizes the subject's fixation control state; the visual response index characterizes the subject's localization efficiency of the visual target stimulus; the pupillary reaction time index characterizes the time from the appearance of the visual target to the onset of pupillary constriction; the pupillary amplitude index characterizes the amplitude of pupillary diameter fluctuation during fixation; and the pupillary variability index characterizes the dispersion of pupillary diameter during fixation. Step S3: Input the preset visual state index into the trained binary classification model to obtain the visual health assessment results of the subject.

[0005] According to a second aspect of the embodiments of this specification, a vision health assessment device is provided, the device comprising: The data acquisition unit is used to collect eye movement data of the subjects observing the visual targets displayed on the visual acuity test screen. The visual targets are displayed for a preset duration at different display positions on the visual acuity test screen. The index calculation unit calculates multiple preset visual acuity state indices based on the eye movement data. These indices include fixation stability, visual response, pupillary reaction time, pupillary amplitude index, and pupillary variability index. Specifically: the fixation stability index characterizes the subject's fixation control state; the visual response index characterizes the subject's localization efficiency of the visual target stimulus; the pupillary reaction time index characterizes the time from the appearance of the visual target to the onset of pupillary constriction; the pupillary amplitude index characterizes the amplitude of pupillary diameter fluctuation during fixation; and the pupillary variability index characterizes the dispersion of pupillary diameter during fixation. The result feedback unit is used to input the preset visual state index into the trained binary classification model to obtain the visual health assessment result of the subject.

[0006] According to a third aspect of the embodiments of this specification, an electronic device is provided, including a processor; and a computer-readable storage medium storing computer program instructions that, when executed by the processor, cause the processor to perform the method described in the first aspect.

[0007] According to a fourth aspect of the embodiments of this specification, a computer-readable storage medium is provided having a computer program stored thereon, the computer program being executed by a processor of the method described in the first aspect.

[0008] The embodiments of this application have achieved at least the following technical effects: This invention provides a vision health assessment solution based on eye-tracking data, offering a new technological approach for the early screening and clinical evaluation of amblyopia and visual function abnormalities in children. First, the solution establishes a data collection, indicator calculation, and result feedback process. Compared to traditional methods where inconsistencies arise due to differences in operator experience, this application improves the accuracy and reliability of vision testing while supporting data comparability across institutions and devices. Second, it quantifies multi-dimensional indicators. Utilizing high-precision eye-tracking technology, the solution analyzes and processes the subjects' eye-movement behavior, calculating indicators such as visual reaction time and pupil diameter fluctuations. This not only captures subtle fixation fluctuations, latency differences, and trajectory irregularities but also reveals patterns in eye movements, providing accurate and objective data support for the early detection of visual abnormalities. Finally, through standardized processing and a binary classification model, the solution synergistically uses multiple key indicators for amblyopia risk prediction. Compared to single-dimensional analysis, this solution can capture subtle but clinically significant eye-movement differences, improving the sensitivity of early screening, providing quantification and reference for clinical intervention, and possessing the potential for widespread application in various child populations, including early screening, treatment follow-up, and intervention effect evaluation. In summary, this approach combines multi-dimensional indicator quantification with predictive models, improving the accuracy of assessments of amblyopia and visual function abnormalities in children and providing a basis for early detection, intervention, and clinical decision-making in amblyopia. Attached Figure Description

[0009] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Some specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings in an exemplary and non-limiting manner. The same reference numerals in the drawings indicate the same or similar parts or components. Those skilled in the art should understand that these drawings are not necessarily drawn to scale. In the drawings: Figure 1 This is a flowchart illustrating a vision health assessment method according to an exemplary embodiment of this application; Figure 2 This is a schematic flowchart illustrating a vision health assessment method according to an exemplary embodiment of this application; Figure 3 This is a schematic diagram illustrating the data acquisition process in an exemplary embodiment of this application; Figure 4 This is a schematic diagram illustrating the results of the gaze stability index in an exemplary embodiment of this application; Figure 5 This is a schematic diagram illustrating the results of a visual response index in an exemplary embodiment of this application; Figure 6 This is a schematic diagram illustrating the results of three pupil-related indicators in an exemplary embodiment of this application; Figure 7This is a schematic diagram illustrating the calculation results of five indicators in an exemplary embodiment of this application; Figure 8 This is a block diagram illustrating an electronic device according to an exemplary embodiment of this application; Figure 9 This is a block diagram illustrating a vision health assessment device according to an exemplary embodiment of this application. Detailed Implementation

[0010] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0011] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0012] To address the inconsistency in vision health assessments caused by differences in operator experience, this application improves the accuracy and reliability of amblyopia screening through a standardized and objective process, including unified data collection, pre-defined indicator calculation, and classification model prediction. Furthermore, compared to single-dimensional analysis, this application provides a vision health assessment scheme based on multi-dimensional vision status indicators. This scheme can capture subtle but clinically significant eye movement differences, enhance the sensitivity of early screening, provide quantification and reference for clinical intervention, and has the potential for widespread application across different child populations, thus contributing to increased coverage of screening and amblyopia rehabilitation monitoring at the grassroots level.

[0013] like Figure 1 As shown, the vision health assessment method of this application embodiment includes the following steps: Step S1: Collect eye movement data of the subject observing the visual targets displayed on the vision test screen. The visual targets are displayed for a preset duration at different display positions on the vision test screen.

[0014] Step S2: Calculate multiple preset visual state indicators based on the eye movement data. The multiple preset visual state indicators include fixation stability index, visual response index, pupillary response time index, fixation state pupillary amplitude index, and fixation state pupillary variation index.

[0015] Wherein: the fixation stability index is used to characterize the subject's fixation control state; the visual response index is used to characterize the subject's localization efficiency of the visual target stimulus; the pupillary reaction time index is used to characterize the starting time from the appearance of the visual target to the subject's pupil starting to constrict; the fixation state pupillary amplitude index is used to characterize the fluctuation amplitude of the pupil diameter during the subject's fixation of the visual target; and the fixation state pupillary variability index is used to characterize the dispersion of the pupil diameter during the subject's fixation of the visual target.

[0016] Step S3: Input the preset visual state index into the trained binary classification model to obtain the visual health assessment results of the subject.

[0017] The vision health assessment method of this application, through eye movement data acquisition and processing, calculates multi-dimensional indicators. This not only captures subtle fixation fluctuations, latency differences, and trajectory irregularities, but also reveals patterns in eye movements, providing detailed and objective data support for the early detection of visual abnormalities and improving the accuracy and reliability of vision assessment. Furthermore, by using a binary classification model, multiple key indicators are synergistically used for amblyopia risk prediction, enhancing the sensitivity of early screening, providing quantification and reference for clinical intervention, and possessing the potential for widespread application in different child populations.

[0018] The embodiments described in this specification will now be described in detail.

[0019] See Figure 2 The vision health assessment method in this embodiment includes four steps: 1. Data acquisition, 2. Data preprocessing, 3. Index calculation, and 4. Result feedback. The process of each of the four steps is described below.

[0020] I. Data Collection In general, data acquisition involved using an eye tracker to photograph and record the subjects' eye movements. Before the experiment, the eye tracker underwent routine calibration to ensure the acquisition accuracy met the requirements of subsequent index calculations. During the experiment, the eye tracker continuously tracked the subjects' gaze direction and fixation position in real time, recording continuous eye movement trajectory data and corresponding fixation point data, thereby achieving high-precision capture and quantitative representation of minute changes in eye movement. The acquired raw data was organized and saved according to a preset field structure, preferably output in CSV format. CSV data can include timestamps, left and right eye fixation point coordinates, fixation status / validity indicators, etc., to facilitate direct retrieval and data preprocessing, segmentation, and index calculation in subsequent processes, achieving seamless integration between data acquisition and subsequent calculations.

[0021] Eye tracker: An eye tracker is used to collect data on the eye movements of subjects under specific experimental paradigms.

[0022] In this embodiment, the data acquisition environment was strictly standardized to minimize confounding variables. The experiment was conducted in an acoustically isolated laboratory with constant indoor lighting to avoid fluctuations in pupillary reflexes and physiological arousal caused by changes in ambient light.

[0023] See Figure 3 The subject sat directly in front of the binocular eye tracker, with their head firmly secured using an adjustable chin rest 301 and forehead rest 302. This was done to decouple eye movement within the head from head movement in space; this embodiment focuses only on eye movement within the head, and head fixation ensures that the vast majority of recorded fixation point changes are attributable to eye rotation rather than head displacement. The distance between the subject and the vision testing screen 303 was controlled at 55 cm, and the eye tracker sampling rate was set to 200 Hz.

[0024] Two-point calibration: Before the formal experiment began, a precise, personalized calibration procedure was performed on the subjects using a two-point calibration paradigm. The eye-tracking system recorded raw pupil-corneal reflex data at each fixation point and constructed a nonlinear mapping model that mapped the pupil-corneal reflex feature space to the screen coordinate system.

[0025] Experimental Paradigm: The experimental paradigm, or visual target, is specifically designed based on the purpose of visual acuity assessment. In this embodiment, the visual target is displayed at different positions on the visual acuity testing screen for a preset duration, for example, 3 seconds at one position. After the experiment officially begins, the subject is asked to sequentially look at the four preset positions on the screen where the experimental paradigm appears. The eye tracker records the entire process in real time and stores the data in CSV format. The experimental paradigm is specifically represented by the stimulus image presented on the eye tracker, see [link to relevant documentation]. Figure 3 In this embodiment of the invention, to improve the cooperation of the subjects, mainly children, in vision testing, a cartoon image 304 is used as a visual target.

[0026] Data Storage: The raw data acquired by the eye tracker mainly includes gaze trajectory and pupil diameter sequences. For example, key fields in a frame of eye-tracking data include: timestamp, validity marker, trigger signal, gaze coordinates for each eye, and pupil diameter for each eye. The acquired data is stored in CSV format.

[0027] II. Data Preprocessing The main purpose of data preprocessing is to segment eye-tracking data, filter out noise, and fill in data to improve data quality and facilitate subsequent indicator calculations.

[0028] Data Segmentation: For the CSV format eye-tracking data collected in the aforementioned data acquisition steps, the eye-tracking location field, validity flag field, and trigger signal field are read according to preset CSV field rules. Based on the trigger signals, the gaze trajectory data and pupil diameter sequence are segmented to obtain data segments that correspond one-to-one with the experimental task segments. For example, if the experimental task here is to sequentially gaze at four preset positions of the experimental paradigm on the screen, then the data segments corresponding one-to-one with the experimental task segments are the four data segments where the subject gazes at the experimental paradigm at the four preset positions.

[0029] In this embodiment, data segmentation involves segmenting the gaze trajectory and pupil diameter sequence separately to obtain gaze trajectory segments and pupil diameter sub-sequences corresponding to each display position. Combined with... Figure 3 The preset positions are the top, bottom, left, and right of the screen. After segmentation, the gaze trajectory segments and pupil diameter subsequences corresponding to these four positions are obtained respectively. For example, the gaze trajectory segment and pupil diameter subsequence corresponding to the top, and the gaze trajectory segment and pupil diameter subsequence corresponding to the left.

[0030] Missing data imputation: After completing frame-by-frame quality verification and segmentation, missing data caused by device failure, incomplete segments, or field parsing failure is uniformly processed. In accordance with the numerical missing data specification, the corresponding eye-tracking coordinates, indicators, and derived analysis result fields are filled with NaN (Not a Number) and the field structure is kept consistent in subsequent indicator calculation and result output, and they do not participate in numerical calculation.

[0031] Noise Removal: For the primary noise source in eye-tracking data—blinking—noise is identified by real-time monitoring of the validity marker field (valid) and changes in pupil diameter. Noise Marking: When a valid value is detected to momentarily drop to zero accompanied by pupil data loss, the eye tracker accurately pinpoints the start and end frames of the blink and marks them as invalid data. Interpolation Compensation: For short-term data gaps, a linear regression model based on the eyeball center rotation angle is used for interpolation to maintain data continuity. Boundary Anomaly Correction: To further improve data quality, this application integrates a boundary anomaly correction algorithm. This algorithm automatically identifies and removes noise points at coordinate abrupt changes by monitoring whether the gaze point exceeds the physical screen boundary and combining the instantaneous motion velocity between adjacent sampling points. This dual verification mechanism effectively prevents the impact of spatial coordinate drift caused by device calibration errors or lighting interference on the results.

[0032] III. Indicator Calculation Based on data preprocessing, this step mainly involves calculating various preset visual state indicators based on eye movement data.

[0033] The embodiments of this invention include various preset visual state indicators, including a fixation stability index, a visual response index, a pupillary reaction time index, a fixation-state pupillary amplitude index, and a fixation-state pupillary variability index. Specifically: the fixation stability index characterizes the subject's fixation control state; the visual response index characterizes the subject's localization efficiency of the visual target stimulus; the pupillary reaction time index characterizes the starting time from the appearance of the visual target to the subject's pupil beginning to constrict; the fixation-state pupillary amplitude index characterizes the fluctuation amplitude of the pupil diameter during the subject's fixation of the visual target; and the fixation-state pupillary variability index characterizes the dispersion of the pupil diameter during the subject's fixation of the visual target.

[0034] The calculation of the indicators here includes: dividing the gaze trajectory and pupil diameter sequence into a preset number of gaze trajectory and pupil diameter sequences in chronological order; segmenting the gaze trajectory and pupil diameter sequence of the first number of sequences to obtain gaze trajectory segments and pupil diameter subsequences corresponding to each display position; calculating the gaze stability index and visual response index corresponding to each gaze trajectory segment, and using the average of the gaze stability index and the average of the visual response index corresponding to each gaze trajectory segment as the final gaze stability index; calculating the pupil amplitude index and the pupil variation index under gaze state based on the pupil diameter sequence; calculating the pupil reaction time index corresponding to each pupil diameter subsequence, and using the average of the pupil reaction time index corresponding to each pupil diameter subsequence as the final pupil reaction time index.

[0035] It should be noted that this embodiment presupposes two rounds. In the specific experimental process, the continuous eye movement data stream can be divided into round 1 and round 2 data segments in ascending order of time. Within each round, it can be further divided into sub-segments in four directions: up, down, left, and right, according to the order of directional presentation. Finally, each data segment is attached with two attributes: round number (1 / 2) and direction label, so that eye movement indicators can be calculated separately for each round. Because this embodiment mainly assesses amblyopia, and the test subjects are children, who are active and do not follow instructions, the visual acuity test results become worse as the test progresses. Therefore, the data from the first round is the best and can best distinguish between amblyopia and normal vision. The following indicator calculations in this embodiment are all based on the fixation trajectory and pupil diameter sequence of the first round.

[0036] (a) Calculation of gaze stability index.

[0037] This indicator aims to assess the neuromuscular control precision of subjects when maintaining fixation in different spatial orientations. By mapping the dispersion of the fixation point to a 3×3 spatial matrix, it can identify whether subjects have visual instability, fixation tremor, or spatial perception bias in specific visual field directions. This improves the targeting and accuracy of amblyopia detection.

[0038] Specifically, the design involves collecting eye-tracking data in two rounds. Each round includes data segments corresponding to the four gaze directions: up, down, left, and right. In each round, gaze stability indices are calculated for each of the four gaze directions. The average of these four gaze stability indices is then used as the final gaze stability index. Two rounds yield two final gaze stability indices. In this embodiment, the final gaze stability index calculated in the first round is used as the initial gaze level for subsequent model input and vision health assessment.

[0039] In theory, after the eyes of a normal person become familiar with the visual targets and experimental tasks in the first round of testing, the fixation stability in the second round is better than that in the first round. However, the opposite is true for amblyopic patients; the longer the fixation time, the less focused their attention becomes. Based on this, this embodiment collects data from two rounds and calculates the final fixation stability index for both rounds. The fixation stability indices of these two rounds are then compared and displayed together. If the index in the second round is worse than that in the first round, it indicates a higher risk of amblyopia in terms of fixation stability.

[0040] It is understandable that in practical applications, only one round of eye movement data can be collected and five preset visual state indicators can be calculated to obtain a visual health assessment result. This embodiment of the invention is not limited to collecting data from two rounds.

[0041] In this embodiment, the fixation stability index is calculated based on the preprocessed eye movement data. The fixation trajectory data is divided into eight fixation trajectory segments in the first and second test rounds according to a preset trigger code. The first and second test rounds each contain four fixation directions: up, down, left, and right. Based on the selected target eye (left or right eye), the horizontal and vertical fixation coordinates corresponding to the valid data frames are extracted in each fixation trajectory segment to calculate the fixation stability index (FSI) for the corresponding fixation trajectory segment. The FSI is determined based on the dispersion of the position of each valid fixation point relative to the average fixation center of the fixation trajectory segment. The smaller the value, the more stable the fixation.

[0042] In this embodiment, the gaze stability index is calculated through the following steps: The gaze stability index corresponding to each gaze trajectory segment is calculated using formula (1). : (1) in, This represents the x-coordinate of the i-th frame of the gaze trajectory segment. This represents the ordinate of the i-th frame of the gaze trajectory segment. This represents the average x-coordinate of all frames in the gaze trajectory segment. This represents the average vertical coordinate of all frames in the gaze trajectory segment; This indicates the total number of frames in the gaze trajectory segment.

[0043] The final gaze stability index is obtained by calculating the arithmetic mean of the gaze stability indices corresponding to each gaze trajectory segment.

[0044] For example, the first round includes gaze stability indices corresponding to four directions: up, down, left, and right. The gaze stability index for the up direction is... The gaze stability index in the lower position is The fixation stability index for the left side is The fixation stability index for the right side is After calculating the fixation stability index (FSI) for these four directions, the mean FSI for all directions in the first round is calculated. : (2) Will As the final indicator of gaze stability.

[0045] To improve the readability and visualization of vision test results, this embodiment maps the fixation stability FSI values ​​corresponding to each fixation direction in the first and second test rounds to the top, bottom, left, and right positions of the four-eye position matrix, respectively, to construct the stability heatmaps for the first and second rounds. At the same time, the FSI values ​​at each matrix position are colored and rendered based on a unified color scale, and numerical labels are superimposed at the corresponding positions to characterize the spatial distribution characteristics of the subject's stability in different fixation directions and the retest consistency differences between the two rounds of testing.

[0046] Combination Figure 4 Heatmaps, a visualization tool, visually represent the density and distribution of a subject's fixation points. Heatmaps reflect fixation stability. A healthy, stable fixation should have the fixation point highly concentrated within a small area, appearing as a well-defined orange-red area on the heatmap. Conversely, unstable fixation (such as nystagmus) will appear as a large, diffuse, warm-colored area on the heatmap. Heatmaps also reflect spatial perception biases. By observing whether the heatmap is biased to one side (e.g., always to the left), it can be inferred whether there is a systematic shift in the subject's perception of the spatial center.

[0047] like Figure 4 As shown, the two-round fixation stability heatmap is constructed using nine eye positions, i.e., the subject's left eye looks up, center, down, upper left, directly left, upper right, directly right, lower left, and lower right. In this embodiment, fixation stability indices are calculated and labeled for four of these positions: up, down, directly left, and directly right. The heatmap of the first round (up, down, left, right) nine-eye stability (left eye) is used for illustration. Figure 4 As shown, the fixation stability index for the upper position is 40.8, for the left position it is 124.6, for the right position it is 109.5, and for the lower position it is 24.2. Figure 4 As shown in the heatmap, the second pass (up, down, left, right) of fixation stability (left eye) across nine eye positions reveals the following: the fixation stability index value for the upper position is 126.8, for the left position it is 49.1, for the right position it is 121.7, and for the lower position it is 21.4. This indicates that for subjects fixing their gaze in different positions, an index value less than 150 is considered normal; the lower the value, the more stable the fixation.

[0048] In this embodiment of the application, the average value of the FSI in the four fixation directions during the first test round is defined as the initial fixation level. The initial level here refers to the initial level relative to the second round of testing. After calculating the initial fixation level, it is combined with normative anchor point mapping (percentage algorithm) and displayed in the form of a circular scoring chart to assess the subject's basic fixation control ability. The lower the value, the more stable the basic fixation; the higher the converted percentage score, the better the initial fixation control level.

[0049] In this embodiment, the initial gaze level index is also mapped to a score through a normative anchor point for intuitive display. The specific mapping formula is as follows: Score=max(0,min(100,100-20 (3) Where V is the subject's final gaze stability index, This indicates the 5th percentile in the group with normal vision, meaning approximately 5% of the samples have a value no higher than this. This represents the 80th percentile in the normal vision group, meaning approximately 80% of the samples have a value no higher than this. Note: In the normal vision group, the 5th and 80th percentiles are determined by sorting all values ​​of the fixation stability index from smallest to largest.

[0050] See also Figure 4 , Figure 4A circular scoring chart is displayed, dynamically showing the score calculated from the initial fixation level index through a scoring ring. In this example, the subject's initial fixation level score is 87, and the scoring ring is displayed in green. Here, the ring adopts a segmented slice structure, and the ring color automatically changes according to the score. If the score is ≥80, the ring is displayed in green; if the score is ≥60, the ring is displayed in yellow; if the score is <60, the ring is displayed in red. This intuitively reflects the subject's initial fixation ability, improves the sensitivity of early screening, provides quantification and reference for clinical intervention, and has the potential to be widely applied in different child groups, including early screening, treatment follow-up, and intervention effect evaluation.

[0051] (ii) Calculation of visual response index.

[0052] This indicator aims to assess the processing speed of visual stimuli in the central nervous system and the efficiency of eye movement initiation in subjects. By calculating the distribution of visual reaction time in different spatial orientations, it quantitatively analyzes the agility of visual response, spatial symmetry, and attentional stability.

[0053] Visual response index calculation involves extracting valid fixation points from each fixation trajectory segment of segmented data, and calculating the reaction time corresponding to each fixation trajectory segment, which is the time difference between the appearance of the visual target stimulus and the start of the subject's eye movement. It can be understood that the visual target stimulus here is an event in visual acuity testing that can elicit a pupillary response, such as the appearance or disappearance of a visual target that requires the subject's fixation.

[0054] In this embodiment, the visual response index is calculated through the following steps: Calculate the instantaneous velocity corresponding to each frame in the gaze trajectory segment. : = (4) in, This represents the change in the x-coordinate between the i-th frame and the (i+1)-th frame. This represents the change in the ordinate between the i-th frame and the (i+1)-th frame. This represents the time difference between the i-th frame and the (i+1)-th frame; For example, after the visual target stimulus appears, the initial interval of the stimulus's duration is selected as the analysis window. Let the sampling frequency be... (e.g., 200Hz), the inter-frame time interval is t= Obtain the two-dimensional gaze point coordinate sequence within this analysis window. =( The instantaneous motion velocity of the eyeball in a two-dimensional screen coordinate system is calculated frame by frame using Euclidean distance and first-order forward difference. .

[0055] Filter the instantaneous velocity corresponding to each frame in the gaze trajectory segment, and when the instantaneous velocity of multiple consecutive frames... When all values ​​exceed the preset speed threshold and the time interval between the first and last frames in the consecutive frames is not less than the preset time threshold, the sampling time of the first frame in the consecutive frames is determined as the visual response start time.

[0056] Here, visual responses such as saccades are defined by setting a threshold for the speed at which saccade events are triggered. and minimum duration threshold The speed threshold is, for example, 8000 (px / s), and the minimum duration threshold is, for example, 5 milliseconds. The processor sequentially iterates through the instantaneous motion speed sequence calculated above. And perform condition judgment: when the instantaneous motion speed of k consecutive frames is detected to be greater than the speed threshold (i.e. > When the time interval between k consecutive frames is greater than or equal to 1, it is determined that the subject has initiated rapid saccades to capture the visual target. Note: The time interval between consecutive k frames is greater than or equal to 1. .

[0057] The time difference between the start time of the visual response and the time of the appearance of the target is calculated, and this time difference is used as the visual response index for the gaze trajectory segment. For example, the index of the first frame in a continuous frame sequence that meets the aforementioned dual threshold conditions of velocity and time is extracted. The timestamp corresponding to this frame marks the start time of the scanning action. The difference between the start time of a saccade event and the presentation time of the visual target stimulus in that location is calculated in real time to derive a visual response index.

[0058] In the specific implementation process, since the aforementioned analysis window strictly presents the time based on the viewpoint, Since it is captured at the zero point of time, the above calculation process can be directly implemented by mapping the first frame index to physical time. The equivalent transformation equation is: Latency= (5) in, This represents the index of the first frame in a continuous frame sequence. t represents the inter-frame time interval. Multiply by 1000 to convert the time unit to milliseconds (ms).

[0059] After calculating the visual response index corresponding to the four gaze trajectory segments in the first round, the arithmetic mean of the visual response index corresponding to each gaze trajectory segment is calculated to obtain the final visual response index.

[0060] For example, a gaze trajectory segment includes segments in four directions: up, down, left, and right. The corresponding visual response indices are as follows: the visual response index for the up direction is... The visual response index in the lower position is The visual response index for the left side is The visual response index for the right side is After calculating the visual response indices for these four directions, the mean of the visual response indices for all directions is then calculated. The formula is as follows: (6) Among them, the visual response index in the upper position is The visual response index in the lower position is The visual response index for the left side is The visual response index for the right side is .

[0061] In other words, after calculating the visual response indexes in the four directions, the arithmetic mean is taken to obtain the final visual response index.

[0062] See Figure 5 In this embodiment, the scores of the visual response index are displayed in the form of a circular rating chart to intuitively represent the overall visual response speed of the subject. The lower the visual response index, the faster the response. The higher the percentage score of the converted visual response index, the better the visual response. Figure 5 The visual response score of the subject was 77 points, and the ring was displayed in yellow.

[0063] The algorithm for mapping the visual response index to a percentage score here also uses the norm-modulated anchor point mapping. For detailed steps, please refer to the mapping process of the aforementioned gaze stability index. The algorithm is the same and will not be repeated here.

[0064] (III) Calculation of pupillary reaction time index.

[0065] This indicator aims to assess the pupillary response speed of subjects during fixation tasks, thereby helping to determine their visual stimulus processing, neural regulatory response, and the stability of autonomic nervous function under fixation conditions.

[0066] The pupillary reaction time index is calculated for each pupil diameter subsequence after processing based on the pupil diameter sequence. Based on target stimulus event locking analysis, the pupillary reaction time is extracted by baseline correction before stimulation and response window detection after stimulation. The starting point of the first sustained satisfaction of the pupillary constriction amplitude and velocity thresholds is then obtained, and the average of multiple pupillary reaction times is taken to characterize the timeliness of pupillary visual response.

[0067] In this step, the pupillary reaction time index corresponding to each pupil diameter subsequence is calculated as follows: Pre-stimulation baseline correction: Set a baseline time window before the stimulus occurs, with the stimulus trigger time as the zero point. Seconds, extract the pupil diameter sequence within the window and use the median as the baseline pupil level for this stimulus: The onset time of visual target stimulation is determined based on the trigger signal of the visual target stimulus; The pupil diameter subsequence is analyzed using the following formula. The sampling times of each pupil diameter are time-aligned to obtain the alignment time τ: (7) in, Indicates the onset time of the visual target stimulus. This represents the sampling time for each pupil diameter. The sampling time for the pupil diameter is the target stimulus initiation time; taking the target stimulus initiation time as the zero point, the alignment time is taken as... The pupil diameter within a second is used as the baseline time window; The median pupil diameter within this baseline time window is calculated using the following formula to obtain the baseline pupil diameter value: (8) in, Indicates the baseline value of pupil diameter. Represents the pupil diameter subsequence. Indicates the baseline time window, This indicates the calculation of the median.

[0068] Subsequently, the pupil diameter sequence within the post-stimulation response window was subtracted from the baseline pupil level to obtain the relative baseline change.

[0069] Specifically, for the pupil diameter subsequence Baseline correction was performed on each pupil diameter to obtain the change in pupil diameter. : (9) in, Represents the pupil diameter subsequence. This represents the baseline value of the pupil diameter. This represents the change in pupil diameter relative to the baseline before stimulation. The starting point of the constriction response is then identified based on this relative change and its rate of change, thereby eliminating the influence of individual differences in baseline pupil size and instantaneous state on the calculation of pupillary reaction time.

[0070] Using the change in pupil diameter Calculate the rate of change of pupil diameter. : (10) in, This represents the difference in pupil diameter. The difference in time taken for the pupil diameter to change; Determine the velocity threshold based on velocity fluctuations within the baseline time window. : (11) in, This represents the standard deviation of the rate of change within the baseline time window, and max() indicates taking the maximum value.

[0071] Post-stimulus response window detection: The search time window is set after the stimulus is triggered, with the stimulus trigger time as the zero point. Extract the baseline-corrected pupil change sequence within this window. The rate of change is calculated. Within this search time window, the pupillary change is determined point by point to determine whether it meets the conditions for an effective pupillary constriction response. When at a certain time point, the pupillary change amplitude exceeds the baseline fluctuation threshold, the pupillary change rate exceeds the velocity threshold and the direction is pupillary constriction, and the duration of this state is not less than [a certain value]. When the first time point of the continuous segment is determined, the starting point of the pupillary response is defined.

[0072] Specifically, taking the onset time of the visual target stimulus as zero, the alignment time is selected as... The pupil diameter within a second range is used as the search time window. Within the search time window, the starting time that satisfies the pupil constriction response condition is found by using the following criteria: (12) in, Represents the absolute value of the rate of change. Indicates the speed threshold. This represents the absolute value of the change in pupil diameter. This represents the standard deviation of pupil dilation within the search time window. Indicates a negative speed threshold. This indicates that the pupil is moving in the direction of contraction and the rate of change exceeds the velocity threshold. .

[0073] Starting from this initial time, the pupil diameter is checked frame by frame to see if it meets the condition for pupil constriction. If it does, the search continues to the next frame until the pupil diameter no longer meets the condition for pupil constriction, thus obtaining a continuous segment. When the length of the continuous segment is not less than 0.05 seconds, the first time point of the continuous segment is determined as the pupil reaction time index. (13) in, Indicator of pupillary reaction time This indicates taking the minimum value. Indicates the search time window. express The pupil diameter at all time points within the continuous segment satisfies , This indicates that the condition for the pupil to contract has been met.

[0074] It should be noted that the aforementioned equation (8) represents the condition for a pupillary constriction response. A pupillary constriction response occurs if and only if this condition is met. ,otherwise In other words, a time point is considered the starting time for a pupillary constriction response only if three conditions are met: the amplitude of pupillary change exceeds the baseline fluctuation threshold, the speed of pupillary change exceeds the speed threshold, and the direction is constriction. Only when all time points exceeding 0.05 seconds from this point meet the pupillary constriction response condition is this continuous segment used to determine the pupillary reaction time index.

[0075] As mentioned above, this embodiment collects visual targets presented in four directions: up, down, left, and right, corresponding to four pupil diameter subsequences. Correspondingly, four pupil reaction time indices are calculated for each of these four pupil diameter subsequences. The final pupil reaction time index is obtained by calculating the arithmetic mean of the pupil reaction time indices corresponding to each pupil diameter subsequence.

[0076] (iv) Calculation of pupil amplitude index under fixation state.

[0077] The fixational pupil relative amplitude index is based on percentile normalization, and the pupil diameter amplitude fluctuation is characterized by the ratio of the difference between the 95th percentile and the 5th percentile pupil diameter to the median pupil diameter.

[0078] This embodiment calculates the pupillary amplitude index under fixation state through the following steps: Extract all pupil diameters labeled as fixation states from the pupil diameter sequence to construct a fixation state pupil diameter sequence: Where n represents the total number of pupil diameters, This represents the pupil diameter in the nth fixation state; Calculate the 5th percentile of the pupil diameter sequence during this fixation state. and the 95th percentile These serve as the lower and upper limits of the normal range of pupil fluctuation, respectively. According to the 5th percentile and the 95th percentile Calculate the difference between the median pupil diameter and the median pupil diameter. ; Median pupil diameter As a normalization benchmark, the pupillary amplitude index under fixation is obtained by normalization using the following formula. : (14) in, The pupillary amplitude index represents the pupillary amplitude during fixation. This represents the 95th percentile of the pupil diameter sequence during this fixation state. This represents the 5th percentile of the pupil diameter sequence during this fixation state. This represents the median pupil diameter.

[0079] here, Indicates a higher pupil diameter level. This represents a lower pupil diameter level, and the difference between the two values ​​is used to characterize the range of pupil amplitude fluctuations. Compared to directly using the maximum and minimum values, this method can reduce the influence of extreme outliers on the results; further dividing by the median pupil diameter can weaken the individual differences in baseline pupil size, making pupil amplitude fluctuations comparable among different subjects.

[0080] (v) Calculation of pupil variation index under fixation state.

[0081] The fixational pupil variability index is based on the coefficient of variation and is characterized by the ratio of the standard deviation of the pupil diameter to the mean.

[0082] In this embodiment, the pupillary variation index under fixation is calculated through the following steps: Extract all pupil diameters labeled as fixation states from the pupil diameter sequence to construct a fixation state pupil diameter sequence: Where n represents the total number of pupil diameters, This represents the diameter of the nth pupil; Calculate the average pupil diameter of the pupil diameter sequence under this fixation state. : (15) Where n represents the total number of pupil diameters in the pupil diameter sequence for this gaze state. This represents the diameter of the i-th pupil; Calculate the standard deviation of pupil diameter : (16) Where n represents the total number of pupil diameters in the pupil diameter sequence for this gaze state. This represents the diameter of the i-th pupil. Indicates the average pupil diameter; Normalization is performed using the following formula, and the normalized result is used as the pupillary variability index (PVI_stare) for fixation: (17) in, The standard deviation of pupil diameter is represented by the standard deviation of the pupil diameter. This represents the average pupil diameter.

[0083] Figure 6 The results of three pupil-related indicators in this embodiment are illustrated. These three pupil-related indicators are: pupil reaction time index, pupil amplitude index under fixation, and pupil variation index under fixation.

[0084] This embodiment completes the tricolor risk stratification based on the 20th and 80th percentiles of the normal vision group, and realizes the norm localization of each pupil index and visualization of the risk level by superimposing subject index value markers on the horizontal threshold band.

[0085] exist Figure 6 In the diagram, green indicates normal vision, yellow indicates developmental fluctuations / overlapping areas, red indicates a higher risk of amblyopia, and the black vertical line represents the current subject's value. Figure 6 As shown, the subject's pupillary reaction time was 596 ms. The fixational pupillary amplitude index was 17.29%, and the fixational pupillary variability index was 6.16%.

[0086] IV. Results Feedback.

[0087] In this step, the five visual state indicators calculated above are input into the trained binary classification model to obtain the subject's visual health assessment results. This step aims to fuse isolated eye movement indicators through multi-source data and machine learning modeling, ultimately generating a comprehensive amblyopia risk assessment result with high clinical reference value.

[0088] Model training involves establishing a feature matrix and label vector based on training samples labeled with normal / amblyopia. The five indicator features are ranked according to their discriminative power, and the three features with the highest contribution are selected. Then, hierarchical cross-validation is used to obtain the amblyopia prediction probability of the training samples, and the optimal classification threshold is determined based on the prediction probability and the true label. Finally, the final binary classification model is trained using all training samples, and the amblyopia risk probability is output to the subjects. When the probability is greater than or equal to the optimal classification threshold, the risk of amblyopia is judged to be high; otherwise, the risk of amblyopia is judged to be low.

[0089] First, feature data matrix construction and standardization: To eliminate the dimensional differences between different indicators and prevent large numerical features from dominating the machine learning model excessively, feature vectors are placed into the feature data matrix X, and standardization is performed. Let... is the j-th eigenvalue of the i-th sample in the matrix, and the standardized eigenvalue The calculation formula is: (18) Wherein, is the arithmetic mean of the feature in the norm sample population, is the standard deviation of the feature in the norm sample population. After processing, a standardized feature matrix Z with a mean of 0 and a variance of 1 is obtained.

[0090] Secondly, PCA dimensionality reduction and feature fusion: Since there may be serious data collinearity among the extracted eye movement index features, the Principal Component Analysis (PCA for short) method is used here to perform dimensionality reduction and fusion on the standardized feature matrix Z. By calculating the covariance matrix of the matrix and performing eigenvalue decomposition, the first k principal components whose cumulative contribution rate reaches a preset threshold are extracted. This is equivalent to orthogonally transforming the original correlated m-dimensional features into k-dimensional (k<m) completely independent new feature vectors F= , ,..., . This step effectively filters out redundant noise and generates a more representative dimensionality-reduced feature matrix.

[0091] Finally, SVM amblyopia probability prediction and risk grading: input the dimensionality-reduced feature vector F into a pre-trained Support Vector Machine (SVM for short) binary classification model. Different from the traditional hard classification (only output yes or no), the SVM model of this embodiment outputs a continuous probability value P through probability mapping methods such as Platt Scaling [0,1], which is the probability that the subject has amblyopia or abnormal visual development, and performs amblyopia prediction and risk grading based on this probability, so as to realize quantitative and visual amblyopia risk assessment.

[0092] In combination with Figure 7 , Figure 7 illustrates the calculation results of five visual status indicators in this embodiment, which are respectively: initial fixation level indicator, unit: point, result: 87, reference range [80, 100]; visual response indicator, unit: point, result: 77, reference range [80, 100]; pupil response latency indicator, that is, the aforementioned pupil response time indicator, unit: ms, result: 596 ms, reference range [0, 349]; relative amplitude index of pupil in fixation state, unit: %, result: 17.29%, reference range [0, 15]; pupil variation index in fixation state, unit: %, result: 6.16%, reference range [0, 5].

[0093] This embodiment outputs the subject's amblyopia probability and risk. For example, based on five core visual state indicators calculated from eye-tracking data, a binary classification model is used for classification. The current results indicate that the subject's amblyopia risk is high, and the subject's amblyopia probability is 0.546. Outputting the amblyopia probability and risk provides a quantitative and visualized amblyopia prediction assessment, which is helpful for the clinical diagnosis of amblyopia screening.

[0094] Based on the above embodiments of this application, the technical solution of this application has achieved at least the following technical effects: First, this technical solution establishes a standardized process for data collection, indicator calculation, and result feedback. Compared to the inconsistencies in results caused by differences in operator experience in traditional methods, the vision health assessment solution proposed in this application improves accuracy and reliability, while also supporting data comparability across institutions and devices. Second, in terms of multi-dimensional indicator quantification, this solution processes and analyzes the subjects' eye movement behavior, calculating indicators such as visual reaction time and pupil diameter fluctuations. This not only captures subtle fixation fluctuations, pupillary latency differences, and trajectory irregularities, but also reveals the patterns of eye movements, providing refined and objective data support for the early detection of visual abnormalities. Finally, this solution synergistically uses multiple vision status indicators for amblyopia risk prediction. Compared to single-dimensional analysis, this solution can capture subtle but clinically significant eye movement differences, improve the sensitivity of early screening, provide quantification and reference for clinical intervention, and has the potential for widespread application in different child groups, including early screening, treatment follow-up, and intervention effect evaluation. In summary, this approach combines multi-dimensional quantitative indicators with predictive models to achieve a comprehensive, objective, and reproducible assessment of amblyopia and visual function abnormalities in children, providing a basis for early detection, intervention, and clinical decision-making.

[0095] Figure 8 This is a schematic diagram of an electronic device illustrated in this specification according to an exemplary embodiment. Please refer to... Figure 8 At the hardware level, the device includes a processor 802, an internal bus 804, a network interface 806, memory 808, a hardware acceleration device 810, and non-volatile memory 812, and may also include other hardware required for its functions. One or more embodiments of this application can be implemented in software, for example, the processor 802 reads the corresponding computer program from the non-volatile memory 812 into memory 808 and then runs it. Of course, in addition to software implementation, one or more embodiments of this application do not exclude other implementation methods, such as logic devices or a combination of hardware and software, etc. That is to say, the execution subject of the above processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0096] Figure 9This is a block diagram illustrating an exemplary embodiment of the present application of a vision health assessment device, which can be applied to, for example... Figure 8 The electronic device shown implements the technical solution of this application. The vision health assessment device 900 includes: The data acquisition unit 901 is used to acquire eye movement data of the visual target displayed on the visual acuity test screen observed by the subject. The visual target is displayed for a preset duration at different display positions on the visual acuity test screen. The index calculation unit 902 calculates multiple preset visual acuity state indices based on the eye movement data. These indices include a fixation stability index, a visual response index, a pupillary reaction time index, a fixation-state pupillary amplitude index, and a fixation-state pupillary variability index. Specifically: the fixation stability index characterizes the subject's fixation control state; the visual response index characterizes the subject's localization efficiency of the visual target stimulus; the pupillary reaction time index characterizes the time from the appearance of the visual target to the onset of pupillary constriction; the fixation-state pupillary amplitude index characterizes the amplitude of pupillary diameter fluctuation during fixation; and the fixation-state pupillary variability index characterizes the dispersion of pupillary diameter during fixation. The result feedback unit 903 is used to input the preset visual state index into the trained binary classification model to obtain the visual health assessment result of the subject.

[0097] The eye-tracking data includes a gaze trajectory and a pupil diameter sequence. The index calculation unit 902 is used to sequentially divide the gaze trajectory and pupil diameter sequence into preset rounds of gaze trajectory and pupil diameter sequences in chronological order. The gaze trajectory and pupil diameter sequence of the first round are segmented to obtain gaze trajectory segments and pupil diameter sub-sequences corresponding to each display position. The gaze stability index and visual response index corresponding to each gaze trajectory segment are calculated, with the average gaze stability index and the average visual response index corresponding to each gaze trajectory segment used as the final gaze stability index. Based on the pupil diameter sequence, the pupil amplitude index and pupil variation index under gaze conditions are calculated. The pupil reaction time index corresponding to each pupil diameter sub-sequence is calculated, with the average pupil reaction time index corresponding to each pupil diameter sub-sequence used as the final pupil reaction time index.

[0098] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this application according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0099] Accordingly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the methods described in any of the above embodiments.

[0100] Accordingly, embodiments of this application also provide a computer program product configured to perform the methods described in any of the above embodiments.

[0101] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer, which can take the form of a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email sending and receiving device, game console, tablet computer, wearable device, or any combination of these devices.

[0102] In a typical configuration, a computer includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0103] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0104] Computer-readable media, including both permanent and non-permanent, removable and non-removable media, can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random-access memory (SRAM), dynamic random-access memory (DRAM), other types of random-access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, disk storage, quantum memory, graphene-based storage media or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0105] While this specification contains numerous specific implementation details, these should not be construed as limiting the scope of any invention or the scope of the claims, but rather are primarily intended to describe features of specific embodiments of a particular invention. Certain features described in the various embodiments herein may also be implemented in combination in a single embodiment. Conversely, various features described in a single embodiment may also be implemented separately in various embodiments or in any suitable sub-combination. Furthermore, while features may function in certain combinations as described above and even initially claimed in this way, one or more features from a claimed combination may be removed from that combination in some cases, and a claimed combination may refer to a sub-combination or a variation thereof.

[0106] Similarly, although the operations are depicted in a specific order in the accompanying drawings, this should not be construed as requiring these operations to be performed in the specific order shown or sequentially, or requiring all illustrated operations to be performed to achieve the desired result. In some cases, multitasking and parallel processing may be advantageous. Furthermore, the separation of various system modules and components in the above embodiments should not be construed as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software product or packaged into multiple software products.

[0107] Thus, specific embodiments of the subject matter have been described. Other embodiments are within the scope of the appended claims. In some cases, the actions recited in the claims may be performed in a different order and still achieve the desired result. Furthermore, the processes depicted in the drawings are not necessarily shown in a specific order or sequence to achieve the desired result. In some implementations, multitasking and parallel processing may be advantageous.

[0108] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0109] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for assessing vision health, characterized in that, include: Step S1: Collect eye movement data of the subject observing the visual targets displayed on the vision test screen. The visual targets are displayed for a preset duration at different display positions on the vision test screen. Step S2: Calculate multiple preset visual acuity indicators based on the eye movement data. These preset visual acuity indicators include fixation stability index, visual response index, pupillary reaction time index, pupillary amplitude index under fixation state, and pupillary variability index under fixation state; wherein: The gaze stability index is used to characterize the subject's gaze control status; The visual response index is used to characterize the subject's localization efficiency of the visual target stimulus; The pupillary reaction time index is used to characterize the time from the appearance of the visual target to the onset of pupillary constriction in the subject; The pupil amplitude index during fixation is used to characterize the fluctuation amplitude of the pupil diameter during the subject's fixation on the target; the pupil variability index during fixation is used to characterize the dispersion of the pupil diameter during the subject's fixation on the target. Step S3: Input the preset visual state index into the trained binary classification model to obtain the visual health assessment results of the subject.

2. The method according to claim 1, characterized in that, The eye-tracking data includes a gaze trajectory and a pupil diameter sequence, and step S2 includes: The fixation trajectory and pupil diameter sequence are sequentially divided into preset rounds of fixation trajectory and pupil diameter sequence according to time order; The gaze trajectory and pupil diameter sequence of the first round are segmented to obtain gaze trajectory segments and pupil diameter subsequences corresponding to each display position; Calculate the gaze stability index and visual response index corresponding to each gaze trajectory segment. Use the average gaze stability index corresponding to each gaze trajectory segment as the final gaze stability index and the average visual response index corresponding to each gaze trajectory segment as the final visual response index. Based on the pupil diameter sequence, the pupil amplitude index and pupil variation index under fixation state were calculated respectively. Calculate the pupillary reaction time index corresponding to each pupil diameter subsequence, and take the mean of the pupillary reaction time index corresponding to each pupil diameter subsequence as the final pupillary reaction time index.

3. The method according to claim 2, characterized in that, Calculate the gaze stability index using the following steps: The gaze stability index for each gaze trajectory segment is calculated using the following formula. : in, This represents the x-coordinate of the i-th frame of the gaze trajectory segment. This represents the ordinate of the i-th frame of the gaze trajectory segment. This represents the average x-coordinate of all frames in the gaze trajectory segment. This represents the average vertical coordinate of all frames in the gaze trajectory segment; This indicates the total number of frames in the gaze trajectory segment; The final gaze stability index is obtained by calculating the arithmetic mean of the gaze stability indices corresponding to each gaze trajectory segment.

4. The method according to claim 2, characterized in that, Calculate the visual response index using the following steps: Calculate the instantaneous velocity corresponding to each frame in the gaze trajectory segment. : = in, This represents the change in the x-coordinate between the i-th frame and the (i+1)-th frame. This represents the change in the ordinate between the i-th frame and the (i+1)-th frame. This represents the time difference between the i-th frame and the (i+1)-th frame; Filter the instantaneous velocity corresponding to each frame in the gaze trajectory segment, and when the instantaneous velocity of multiple consecutive frames... When all exceed the preset speed threshold and the time interval between the first and last frames in the consecutive frames is not less than the preset time threshold, the sampling time of the first frame in the consecutive frames is determined as the visual response start time. Calculate the time difference between the onset time of visual response and the time of appearance of the visual target; This time difference is used as an indicator of visual response to the gaze trajectory segment; The final visual response index is obtained by calculating the arithmetic mean of the visual response index corresponding to each gaze trajectory segment.

5. The method according to claim 2, characterized in that, The calculation of pupillary reaction time indices corresponding to each pupil diameter subsequence includes: The onset time of visual target stimulation is determined based on the trigger signal of the visual target stimulus; The pupil diameter subsequence is analyzed using the following formula. The sampling times of each pupil diameter are time-aligned to obtain the alignment time τ: in, Indicates the onset time of the visual target stimulus. This represents the sampling time for each pupil diameter. When the sampling time for the pupil diameter is specified, it indicates that the sampling time is the start time of the visual target stimulus. Taking the onset time of the visual target stimulus as the zero point, the alignment time is taken as... The pupil diameter within a second is used as the baseline time window; The median pupil diameter within this baseline time window is calculated using the following formula to obtain the baseline pupil diameter value: in, Indicates the baseline value of pupil diameter. Represents the pupil diameter subsequence. Indicates the baseline time window, This indicates the calculation of the median; Pupil diameter subsequence Baseline correction was performed on the pupil diameters to obtain the pupil diameter change ΔD(τ): in, Represents the pupil diameter subsequence. Indicates the baseline value of the pupil diameter; Using the change in pupil diameter Calculate the rate of change of pupil diameter. : in, This represents the difference in pupil diameter. The difference in time taken for the pupil diameter to change; Determine the velocity threshold based on velocity fluctuations within the baseline time window. : in, This represents the standard deviation of the rate of change within the baseline time window; max() indicates taking the maximum value. Taking the visual target stimulus initiation time as zero, the alignment time is selected as... The pupil diameter within a second range is used as the search time window. Within the search time window, the starting time that satisfies the pupil constriction response condition is found by using the following criteria: in, Represents the absolute value of the rate of change. Indicates the speed threshold. This represents the absolute value of the change in pupil diameter. This represents the standard deviation of pupil dilation within the search time window. Indicates a negative speed threshold. This indicates that the pupil is moving in the direction of contraction and the rate of change exceeds the velocity threshold. ; Starting from this initial time, the pupil diameter is checked frame by frame to see if it meets the condition for pupil contraction. If it does, the next frame is checked until the pupil diameter no longer meets the condition for pupil contraction, thus obtaining a continuous segment. When the length of a continuous segment is not less than 0.05 seconds, the first time point of the continuous segment is determined as the pupillary reaction time index. in, Indicator of pupillary reaction time This indicates taking the minimum value. Indicates the search time window. express The pupil diameter at all time points within the continuous segment satisfies , This indicates that the condition for pupil constriction has been met; The final pupil reaction time index is obtained by calculating the arithmetic mean of the pupil reaction time index corresponding to each pupil diameter subsequence.

6. The method according to claim 2, characterized in that, The pupillary amplitude index during fixation is calculated using the following steps: Extract all pupil diameters labeled as fixation states from the pupil diameter sequence to construct a fixation state pupil diameter sequence: Where n represents the total number of pupil diameters, This represents the pupil diameter in the nth fixation state; Calculate the 5th percentile of the pupil diameter sequence during this fixation state. and the 95th percentile These serve as the lower and upper limits of the normal range of pupil fluctuation, respectively. According to the 5th percentile and the 95th percentile Calculate the difference between the median pupil diameter and the median pupil diameter. ; Median pupil diameter As a normalization benchmark, the pupillary amplitude index under fixation is obtained by normalization using the following formula: in, The pupillary amplitude index represents the pupillary amplitude during fixation. This represents the 95th percentile of the pupil diameter sequence during this fixation state. This represents the 5th percentile of the pupil diameter sequence during this fixation state. This represents the median pupil diameter.

7. The method according to claim 2, characterized in that, The pupillary variability index under fixation is calculated using the following steps: Extract all pupil diameters labeled as fixation states from the pupil diameter sequence to construct a fixation state pupil diameter sequence: Where n represents the total number of pupil diameters, This represents the diameter of the nth pupil; Calculate the average pupil diameter of the pupil diameter sequence under this fixation state. : Where n represents the total number of pupil diameters in the pupil diameter sequence for this gaze state. This represents the diameter of the i-th pupil; Calculate the standard deviation of pupil diameter : Where n represents the total number of pupil diameters in the pupil diameter sequence for this gaze state. This represents the diameter of the i-th pupil. Indicates the average pupil diameter; Normalization is performed using the following formula, and the normalized result is used as the pupillary variability index (PVI_stare) for fixation: in, The standard deviation of pupil diameter is represented by the standard deviation of the pupil diameter. This represents the average pupil diameter.

8. A vision health assessment device, characterized in that, include: The data acquisition unit is used to collect eye movement data of the subjects observing the visual targets displayed on the visual acuity test screen. The visual targets are displayed for a preset duration at different display positions on the visual acuity test screen. The index calculation unit calculates multiple preset visual acuity state indices based on the eye movement data. These indices include fixation stability, visual response, pupillary reaction time, pupillary amplitude index, and pupillary variability index. Specifically: the fixation stability index characterizes the subject's fixation control state; the visual response index characterizes the subject's localization efficiency of the visual target stimulus; the pupillary reaction time index characterizes the time from the appearance of the visual target to the onset of pupillary constriction; the pupillary amplitude index characterizes the amplitude of pupillary diameter fluctuation during fixation; and the pupillary variability index characterizes the dispersion of pupillary diameter during fixation. The result feedback unit is used to input the preset visual state index into the trained binary classification model to obtain the visual health assessment result of the subject.

9. An electronic device, characterized in that, Including the processor; A computer-readable storage medium storing computer program instructions that, when executed by the processor, cause the processor to perform the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which is executed by a processor according to any one of claims 1 to 7.

Images