A functional vision test method for self-monitoring of visual health

Abstract

The present invention provides a functional vision testing method for self-monitoring visual health, which belongs to the field of intelligent vision testing technology. The method comprises the following steps: loading a multi-task adaptive model and parameters of a self-supervised pre-training editor into a visual tester's smart device, and presenting test questions for visual acuity test, contrast sensitivity test, and color vision test respectively as required after loading; inputting the test content and operation behavior of each test question into the self-supervised pre-training editor to extract fusion features; based on all fusion features under each test and in combination with the difficulty level of each test question based on the current distance between the display screen and the visual tester, obtaining the visual acuity threshold, contrast sensitivity curve, and color vision arrangement error distribution of the visual tester, and generating a remote visual assessment report. This facilitates self-monitoring and improves the efficiency of visual assessment, facilitating remote and accurate disease assessment.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent vision testing, and in particular to a functional vision testing method for performing visual health self-monitoring. Background Art

[0002] Cataracts are a disease that causes visual impairment due to clouding of the lens. With the increasing aging of society, their efficient monitoring has become increasingly important in the field of elderly health management. Functional vision assessment, which reflects the patient's visual status in real life, is particularly important in the clinical evaluation system for cataract severity. Existing functional vision assessment systems such as visual acuity, contrast sensitivity, motion target resolution, and color vision can evaluate the human eye's visual function from multiple dimensions, including form resolution, the ability to observe details of moving objects, and color recognition, thereby comprehensively presenting the visual changes caused by cataracts. However, the assessment process is relatively complex and relies on professional examination equipment, which is inconvenient for the elderly to self-monitor and reduces the efficiency of the assessment. Therefore, it is necessary to develop an examination method based on standard quantitative indicators and combine it with portable electronic devices to achieve remote and accurate assessment of the condition.

[0003] Therefore, the present invention proposes a functional vision testing method for performing visual health self-monitoring. Summary of the Invention

[0004] The present invention provides a functional vision testing method for performing visual health self-monitoring, so as to solve the above-mentioned technical problems.

[0005] The present invention provides a functional vision testing method for performing visual health self-monitoring, comprising:

[0006] Load the multi-task adaptive model and the parameters of the self-supervised pre-training editor to the visual tester's smart device. After loading, the test questions for visual acuity test, contrast sensitivity test, and color vision test are presented as required.

[0007] Capturing the visual tester's operation behavior based on each test question based on the smart device;

[0008] Input the test content and operation behavior of each test question into the self-supervised pre-training editor to extract fusion features;

[0009] Based on all fusion features under each test and combined with the difficulty level of each test question based on the current distance between the display screen and the visual tester, the visual acuity threshold, contrast sensitivity curve and color perception error distribution of the visual tester are obtained, and a remote vision assessment report is generated.

[0010] Preferably, it also includes:

[0011] The test results of each loaded smart device are collected, and the federated learning mechanism is used to perform local training on the collected data to update the parameters of the multi-task adaptive model and the self-supervised pre-training editor.

[0012] Preferably, visual acuity testing includes:

[0013] When the visual tester triggers the visual acuity test operation on the display screen, a presentation set at a difficulty level matching the current distance is output based on the display screen, wherein the presentation set includes a plurality of visual acuity test questions at the initial recognition difficulty level, and the visual acuity test questions are related to the size of the sight mark and the speed of the sight mark movement;

[0014] The feedback result of the vision tester on each visual acuity test item is determined based on the presentation set, wherein the feedback result is related to the operation behavior.

[0015] Preferably, after determining the feedback result of the visual test subject for each visual acuity test question, the method further includes: adjusting the sight mark according to the feedback result and the actual distance after the change, specifically including:

[0016] Obtaining a correct recognition rate based on the feedback results at the difficulty level;

[0017] If the correct recognition rate is greater than or equal to a preset recognition rate, increasing the difficulty level according to the feedback result;

[0018]

[0019] in, Indicates the increased difficulty level; Indicates the current difficulty level; Indicates the visual acuity difficulty adjustment step size; Indicates that based on feedback results , auxiliary functions of key information K and value information V within the multi-task adaptive model; Indicates the preset recognition rate; represents the correct recognition rate; Indicates the rounding symbol;

[0020] Obtaining a first new set matching the actual distance after the change from the level-distance database, and conducting a visual acuity test of the visual test subject at the updated level of difficulty, wherein the first new set includes test questions related to the size of the sight mark and the speed of movement at the updated level;

[0021] If the correct recognition rate is less than the preset recognition rate, the difficulty level is lowered according to the feedback result, and a second new set matching the actual changed distance is obtained from the level-distance database to continue the visual acuity test.

[0022] Preferably, contrast sensitivity testing includes:

[0023] An initial contrast is set, and test questions with the same initial contrast are randomly presented on the display screen based on the initial contrast, and a feedback vector of each contrast test question is obtained from the visual test subject, where the feedback vector f = {Pj, tj}, where Pj represents the judgment result of the j-th contrast test question and tj represents the time taken to answer the j-th contrast test question;

[0024] Input the feedback vector f into the multi-task adaptive model, fuse the bar information corresponding to the contrast test question and the user interaction information to form a high-dimensional feature;

[0025] Dynamically updating the initial contrast according to all high-dimensional features under the initial contrast and continuing testing;

[0026]

[0027] in, Represents all high-dimensional features under the initial contrast , auxiliary functions of key information K and value information V within the multi-task adaptive model; Indicates the contrast after dynamic update; represents the initial contrast; Indicates the contrast difficulty adjustment step size.

[0028] Preferably, color vision testing should include:

[0029] Capturing the visual tester's dragging behavior of a color block for each color vision test question on a display screen, wherein the color block dragging behavior includes: a full dragging trajectory of each test color block and individual dragging trajectories involved in the full dragging trajectory, wherein the individual dragging trajectories are consistent with the number of drags;

[0030] Analyzing the dragging error probability of each test color block based on the color block dragging behavior, and generating a color gamut distribution in combination with the dragging position deviation of each test color block;

[0031] The color vision test level is updated according to the color gamut distribution, and the test is continued.

[0032] Preferably, analyzing the dragging error probability of each test color block based on the color block dragging behavior and combining the dragging position deviation of each test color block to generate a color gamut distribution includes:

[0033] The time it takes for each unit color block to pass through a point in each full drag trajectory involved in each color vision test question is counted to construct a trajectory vector ,in, Indicates the position of the second unit color block in the full drag trajectory and residence time ; Indicates the position of the mth unit color block in the full drag trajectory and residence time ;

[0034] Filtering positions whose dwell time is greater than a preset time from the trajectory vector as first positions, and splitting the entire dragging trajectory according to the first positions;

[0035] Determine the color difference between the color identifier of the unit color block at the last position in each individual dragging trajectory and the color of the test identifier, and obtain the color difference vector and position difference vector corresponding to the entire dragging trajectory;

[0036] When the number of the individual dragging tracks is 1, the color difference vector and the position difference vector are kept unchanged;

[0037] When there are multiple independent drag trajectories, determining the number of drags at each first position and the drag direction of each drag, and drawing a drag vector diagram;

[0038] determining, based on the drag vector diagram, the invalidity of dragging between the last position in each individual dragging track and the first position in the next individual dragging track, adjusting the color difference and the position difference of the last position, and updating the color difference vector and the position difference vector;

[0039] Determine the number of test color blocks in each color gamut for each color vision test question, and construct a color difference matrix based on the final color difference vector to obtain a color difference feature vector for the corresponding color gamut. Simultaneously, construct a position difference matrix based on the final position difference vector to obtain a position difference feature vector for the corresponding color gamut.

[0040] Obtaining a drag error probability of each test color block according to the color difference vector, and obtaining an error vertical value of the corresponding color gamut by combining the color difference feature vector and the position difference feature vector;

[0041] Generate color gamut distribution based on the error vertical values ​​under different color gamuts.

[0042] Preferably, a remote visual assessment report is generated, including:

[0043] Obtaining a visual acuity threshold based on all visual acuity test items at different difficulty levels and the fusion features of each visual acuity test item;

[0044] Based on all contrast sensitivity test items at different difficulty levels of the contrast sensitivity test and the fusion features of each contrast sensitivity test item, a contrast sensitivity curve is obtained;

[0045] Averaging each color gamut error vertical value in the color gamut distribution of color vision test questions at different difficulty levels of the color vision test to obtain a color vision arrangement error distribution;

[0046] A remote vision assessment report is generated based on the visual acuity threshold, the contrast sensitivity curve, and the color perception error distribution.

[0047] Compared with the prior art, the present invention has the following advantages:

[0048] By loading a multi-task adaptive model and editor onto a smart device, visual monitoring can be performed across three dimensions: visual acuity, contrast sensitivity, and color vision. This facilitates self-monitoring, improves the efficiency of visual assessment, and facilitates remote and accurate assessment of medical conditions. Furthermore, within visual acuity testing, the present invention provides a method for self-testing and assessing motion-related visual discrimination, addressing the need for self-monitoring of motion-related visual function in simulated special scenarios by elderly individuals with driving needs in an aging society.

[0049] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0050] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0052] Figure 1 This is a flow chart of a functional vision testing method for performing visual health self-monitoring in an embodiment of the present invention;

[0053] Figure 2 This is a structural diagram of the full-length dragging trajectory in an embodiment of the present invention;

[0054] Figure 3 This is an implementation diagram of step 1 in an embodiment of the present invention;

[0055] Figure 4 4 is a specific flow chart of visual assessment in an embodiment of the present invention. DETAILED DESCRIPTION

[0056] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0057] The present invention provides a functional vision test method for performing self-monitoring of visual health, such as Figure 1 As shown, including:

[0058] Step 1: Load the multi-task adaptive model and the parameters of the self-supervised pre-training editor to the visual tester's smart device. After loading, the test questions for visual acuity test, contrast sensitivity test, and color vision test are presented as required.

[0059] Step 2: Capturing the visual tester's operation behavior based on each test question based on the smart device;

[0060] Step 3: Input the test content and operation behavior of each test question into the self-supervised pre-training editor to extract fusion features;

[0061] Step 4: Based on all fusion features under each test and combined with the difficulty level of each test question based on the current distance between the display screen and the visual tester, the visual acuity threshold, contrast sensitivity curve, and color perception error distribution of the visual tester are obtained, and a remote visual assessment report is generated.

[0062] In this embodiment, the smart device refers to a smart phone, tablet computer, etc., which has a high-performance processor and a high-definition screen, can smoothly display various visual test questions and complex visual patterns, and is convenient for the tester to observe.

[0063] In this embodiment, the multi-task adaptive model can simultaneously handle a variety of different visual testing tasks, such as visual acuity testing, contrast sensitivity testing, and color vision testing. It can also automatically adjust the difficulty of subsequent test questions based on the test subject's actual performance during the test. For example, in a visual acuity test, if the test subject correctly answers several difficult visual items in a row, the model will automatically select a more challenging visual item, such as a smaller, faster-moving sight mark. Conversely, if the test subject repeatedly answers incorrectly, the model will reduce the difficulty of the questions and provide larger, clearer, and slower-moving sight marks.

[0064] In this embodiment, the self-supervised pre-training editor uses self-supervised learning to autonomously learn feature representations from unlabeled data, thereby extracting fused features of test content and operational behavior. It learns various features from visual test data, such as the shape and color of the sight mark image, as well as the tester's operational behavior data, to extract valuable features.

[0065] In this embodiment, the visual acuity test is used to accurately measure the visual system's ability to discern detail. An eye chart displays "E"-shaped optotypes of varying sizes and orientations. The test subject demonstrates their visual acuity by identifying the orientation of the optotypes at a certain distance from the display screen. The test system automatically calculates the size of the optotype representing a specific viewing angle at the test distance based on the test subject's distance from the display screen and presents the size of the optotype. If the test subject can clearly discern the "E" optotype representing a smaller viewing angle, their visual acuity is high; conversely, if they can only clearly discern the optotype with a larger viewing angle, their visual acuity is low.

[0066] In this embodiment, the contrast sensitivity test is used to evaluate the visual system's ability to distinguish objects of different contrasts. A grating pattern with a relatively high contrast (e.g., 80%) is first displayed. At this time, the difference in brightness between the bars in the pattern is obvious, making it easier for the test subject to see clearly. The contrast is then gradually reduced (e.g., to 20%). The difference in brightness becomes smaller, requiring the test subject to concentrate more to distinguish the pattern. The test subject's contrast sensitivity level is assessed by how well he or she can distinguish patterns of different contrasts.

[0067] In this embodiment, a color vision test is used to measure the visual system's ability to discern colors.

[0068] In this embodiment, during the visual test, the tester interacts with the smart device through a series of actions. Common operations include clicking, such as when answering a multiple-choice question, the tester taps the correct option on the screen with their finger; sliding, when the test interface displays a large amount of content, the tester slides their finger across the screen to view different test questions; and touching, in some tests that require adjusting the size or position of the sight mark, the tester touches the screen to perform operations.

[0069] In this embodiment, the test content is the specific information contained in each test question. In the visual acuity test, the shape of the sight mark (such as the letter "E", letters, numbers, etc.), size (size corresponding to different viewing angles), and direction (up, down, left, right, etc.) are all test content; in the contrast sensitivity test, the contrast value of the stripe pattern (such as 30%, 50%, etc.) and the spatial frequency (the number of stripes per unit viewing angle) are test content; in the color vision test, the color combination of the color vision detection image (such as red-green combination, blue-yellow combination, etc.) and the arrangement of color blocks (the relative position relationship between different color blocks and the degree of deviation from the standard arrangement) constitute the test content.

[0070] In this embodiment, fused features are a collection of features formed by organically integrating multiple features of the test content and operational behavior. Taking a visual acuity test as an example, fused features may include the size of the sight mark (e.g., 1′, 5′, etc.), the time it takes the test-taker to click on the answer (e.g., 2 seconds, 5 seconds), and the accuracy of the click position (the deviation between the click position and the correct answer position). These different types of features are combined to form a comprehensive feature vector that can more comprehensively reflect the test-taker's performance and ability on the test item.

[0071] In this embodiment, the visual acuity threshold is a key indicator for measuring the visual system's resolution ability. It refers to the minimum visual angle at which a subject can accurately discern an optotype. For example, if a subject can clearly discern an optotype at a 1' visual angle during a visual acuity test, their visual acuity threshold is 1' visual angle. If they can only discern an optotype at a 2' visual angle, their visual acuity threshold is 2' visual angle. The lower the visual acuity threshold, the higher the subject's visual acuity.

[0072] In this embodiment, the contrast sensitivity curve shows the relationship between the visual system's sensitivity to visual stimuli of different spatial frequencies and contrasts. The horizontal axis represents spatial frequency (in cycles / degree, i.e., the logarithm of the number of light and dark stripes per degree of visual angle), and the vertical axis represents contrast sensitivity (the inverse of the lowest discernible contrast). By plotting the contrast sensitivity points corresponding to different spatial frequencies and connecting these points to form a curve, the contrast sensitivity function of the visual system under different conditions can be intuitively demonstrated. For example, a normal contrast sensitivity curve has higher sensitivity at medium and low spatial frequencies, and the sensitivity gradually decreases as the spatial frequency increases.

[0073] In this embodiment, the color perception error distribution refers to the range of color block hue distribution resulting from an incorrect arrangement order of a series of color blocks. The hue of each color block to be arranged corresponds to a color light with a continuously varying wavelength within the visible spectrum. Arranging all color blocks in the correct order will produce a continuous, gradient color band corresponding to the visible spectrum.

[0074] In this embodiment, the remote visual assessment report can be presented in various formats, including charts (such as a bar graph of visual acuity threshold, a line graph of contrast sensitivity curves, a radar chart of color vision error distribution, a curve graph of color vision error distribution, etc.) and textual descriptions (interpretation of each assessment result, diagnostic recommendations, etc.). For example, the report will detail whether the test subject's visual acuity threshold is within the normal range or abnormal, the difference between the contrast sensitivity curve and the normal curve, and whether the test subject has color vision impairment based on the color permutation error distribution, as well as the specific color segments affected.

[0075] In this embodiment, visual acuity is tested using an adaptive LogMAR vision test algorithm. Based on the LogMAR eye chart, the system dynamically adjusts the size and presentation order of chart symbols based on real-time patient feedback. For example, if a patient makes an error with a large symbol, the system gradually reduces the symbol size until the smallest recognizable symbol is found. This dynamic adjustment mechanism ensures personalized and accurate testing, and can provide a personalized vision assessment based on changes in a patient's visual function.

[0076] In this embodiment, contrast sensitivity testing utilizes an adaptive contrast sensitivity assessment algorithm: the system uses a Bayesian optimization algorithm to dynamically adjust contrast sensitivity testing. By analyzing patient feedback in real time, the system adjusts the contrast level of the test image to find the optimal parameter settings that best reflect the patient's contrast sensitivity. This system efficiently converges to the optimal solution within a small sampling range, ensuring accurate testing and rapid feedback.

[0077] In this embodiment, color vision is tested using a color discrimination test (Farnsworth-Munsell 100Hue color chess). Combined with a convolutional neural network (CNN), this accurately assesses a patient's color discrimination ability by analyzing their sensitivity to hue changes under different lighting conditions. An adaptive mechanism automatically adjusts the hue differences based on the patient's initial performance to ensure the test difficulty matches the patient's actual color vision level, aiding in the early detection of color vision impairment in cataract patients.

[0078] Adopting this adaptive deep learning segmentation network: Using an adaptive deep learning model, it can analyze and segment visual function data in real time, identifying different levels of visual function. This segmentation network automatically adjusts the characteristics of cataract lesions at different stages to provide personalized segmentation results. For example, the system can automatically adjust the network weights and feature extraction methods based on the different stages of a patient's visual function, helping doctors identify lesion progression and providing a reliable basis for personalized treatment plans.

[0079] In this embodiment, the implementation diagram for step 1 is as follows: Figure 3 As stated.

[0080] In this embodiment, Figure 4 The following is a specific visual assessment flowchart.

[0081] The beneficial effects of the above technical solution are: by loading the multi-task adaptive model and editor onto smart devices, visual monitoring can be performed across three dimensions: visual acuity, contrast sensitivity, and color vision testing. This facilitates self-monitoring, improves the efficiency of visual assessment, and facilitates remote and accurate assessment of medical conditions. Furthermore, during visual acuity testing, the present invention provides a method for self-testing and assessing motion-related visual discrimination, addressing the need for elderly individuals with driving needs in an aging society to self-monitor their own motion-related visual function in simulated special scenarios.

[0082] The present invention provides a functional vision testing method for performing visual health self-monitoring, further comprising:

[0083] The test results of each loaded smart device are collected, and the federated learning mechanism is used to perform local training on the collected data to update the parameters of the multi-task adaptive model and the self-supervised pre-training editor.

[0084] In this embodiment, the test results are a data set generated after the visual tester completes the visual acuity test, contrast sensitivity test, and color vision test on the smart device.

[0085] In this embodiment, local training refers to the process in which the smart device uses the test result data collected by itself to locally train the multi-task adaptive model and the self-supervised pre-training editor. For example, based on the visual acuity, contrast sensitivity and color vision test result data of 10 testers, the model parameters are adjusted to improve the prediction accuracy of the model when processing this data. This process is local training.

[0086] In this embodiment, on the smart device side, a lightweight machine learning framework (such as TensorFlowLite, PyTorchMobile) is used to load the parameters of the multi-task adaptive model and the self-supervised pre-training editor. Based on the collected test result data, the model is trained according to a preset training algorithm (such as the stochastic gradient descent algorithm). During the training process, the update amount of the model parameters (such as the change in the weight value) is recorded. To improve the efficiency and effectiveness of local training, a dynamic learning rate adjustment strategy is set on the smart device side. The learning rate is automatically adjusted according to the device's computing resources and the amount of data. When the device has sufficient computing resources and a large amount of data, the learning rate is appropriately increased to speed up the training speed; when resources are limited or the amount of data is small, the learning rate is reduced to ensure the stability of the training. At the same time, to avoid overfitting, a regularization method (such as L2 regularization) is used to constrain the model.

[0087] In this embodiment, the server receives model parameter updates uploaded by each smart device and aggregates these updates using an aggregation algorithm (such as the FedAvg algorithm). Based on this aggregated result, the global parameters of the multi-task adaptive model and the self-supervised pre-training editor are updated. Once the update is complete, the new parameters are distributed to each smart device via the network for use in the next test.

[0088] The beneficial effect of the above technical solution is: by collecting test results for local training, the accuracy of the test results in the subsequent three dimensions is further guaranteed.

[0089] The present invention provides a functional vision testing method for performing visual health self-monitoring for visual acuity testing, comprising:

[0090] When the visual tester triggers the visual acuity test operation on the display screen, a presentation set at a difficulty level matching the current distance is output based on the display screen, wherein the presentation set includes a plurality of visual acuity test questions at the initial recognition difficulty level, and the visual acuity test questions are related to the size of the sight mark and the speed of the sight mark movement;

[0091] The feedback result of the vision tester on each visual acuity test item is determined based on the presentation set, wherein the feedback result is related to the operation behavior.

[0092] In this embodiment, the vision tester is a person who participates in a visual acuity test and receives a vision assessment.

[0093] In this embodiment, the current distance refers to the actual distance between the display screen and the tester's eyes. In actual testing, this distance can be measured using a distance sensor (such as a ToF sensor) on the smart device. For example, if the smart device is placed on a table and the tester is sitting normally, the distance between the device and the eyes is approximately 40 cm. If the tester holds the device in their hand, the distance may be around 30 cm.

[0094] In this embodiment, the difficulty level represents different levels of test difficulty, determined based on the visual acuity test requirements. Generally, higher difficulty levels require smaller sight marks and faster movement, placing higher demands on the tester's vision. For example, difficulty level 1 has a larger, stationary sight mark, suitable for individuals with poor vision or first-time test takers; difficulty level 5 has a very small sight mark that moves quickly across the screen, suitable for testing individuals with excellent vision or those requiring a highly accurate vision assessment.

[0095] In this embodiment, a presentation set is a set of visual acuity test items presented to a visual acuity test subject at a specific difficulty level. For example, a presentation set at difficulty level 3 may include five "E"-shaped visual items of varying orientations, moderate size, and slow movement at a specific speed, as well as three visual items of other shapes (e.g., letters or numbers).

[0096] In this embodiment, the optotype size refers to the size of the optotype used in visual acuity testing, typically measured in units of visual angle ('). For example, on a standard logarithmic eye chart, an optotype size corresponding to 1.0 visual acuity is approximately 1' visual angle; an optotype corresponding to 0.1 visual acuity is larger, approximately 10' visual angle.

[0097] In this embodiment, the movement speed is the speed at which the sight mark moves on the display screen, and the unit is degree / second, that is, the speed at which the sight mark moves relative to the tester (center of the circle) on a circle with the tester as the center and the test distance as the radius. The commonly used low-speed movement sight mark movement speed can be set to 10 degrees / second, the medium-speed movement sight mark movement speed can be set to 30 degrees / second, the medium-high speed movement sight mark movement speed can be set to 60 degrees / second, and the high-speed movement sight mark movement speed can be set to 90 degrees / second.

[0098] In this embodiment, a distance detection and difficulty matching algorithm is written in the visual test APP of the smart device. When it is detected that the tester triggers the visual acuity test operation, the current distance between the display screen and the tester's eyes is first obtained through the distance sensor. If the device is not equipped with a distance sensor, a prompt box will pop up, requiring the tester to manually enter the distance. According to the preset distance-difficulty mapping rules, the difficulty level that matches the current distance is determined. For example, when the distance is 30-40 cm, it corresponds to difficulty level 3; when the distance is greater than 40 cm, the difficulty level is reduced by one level. Then, several visual acuity test questions at this difficulty level are retrieved from the question bank, formed into a presentation set, and displayed on the display screen in a certain order. During the display process, the size and movement speed of the sight mark are controlled by graphics rendering technology. For example, the "E" sight mark is rendered, and its size and movement trajectory are set according to the difficulty level to ensure that the sight mark moves smoothly on the screen at the set speed. In terms of difficulty matching, the proportion of correctly matching the difficulty level according to the distance is 95%. Regarding the presentation effects of the sets at different difficulty levels, the testers reported that 90% of the sight signs had reasonable size and movement speed settings, making them easy to observe and identify; 10% of the testers believed that the movement speed of some high-difficulty sight signs was too fast and difficult to distinguish.

[0099] In this embodiment, the feedback result is the test taker's response to each visual acuity test question and is closely related to the test taker's operation behavior. For example, the test taker selects the direction of the visual target (such as up, down, left, or right) by clicking on the screen, which is the feedback result.

[0100] In this embodiment, the operation behavior refers to the test-taker's interaction with the display screen when answering visual acuity test questions, including touching the screen, sliding the screen, etc. For example, the test-taker uses a finger to slide on the screen to adjust the clarity of the visual mark, or clicks the confirmation button on the screen to submit the answer, which are both considered operation behaviors.

[0101] The beneficial effects of the above technical solution are: based on the output presentation set of the current distance matching difficulty level, it can provide visual acuity test questions of appropriate difficulty according to the test subject's actual test environment, avoid questions that are too difficult or too easy, and improve the accuracy and effectiveness of the test. At the same time, the dynamic sight mark size and movement speed settings simulate a more realistic visual scene, so that the test results can better reflect the test subject's vision in real life. Relying on the presentation set to accurately determine the test subject's feedback results provides a reliable data basis for subsequent visual acuity threshold calculations, vision assessments, etc. By recording operation behaviors and feedback results, it is possible to further analyze the test subject's vision characteristics and answering habits, which helps to optimize visual acuity test questions and processes, and improve the accuracy of the test and user experience.

[0102] The present invention provides a functional vision test method for self-monitoring visual health. After determining the feedback result of the visual acuity test subject for each visual acuity test question, the method further includes: adjusting the sight mark according to the feedback result and the actual distance after the change, specifically including:

[0103] Obtaining a correct recognition rate based on the feedback results at the difficulty level;

[0104] If the correct recognition rate is greater than or equal to a preset recognition rate, increasing the difficulty level according to the feedback result;

[0105]

[0106] in, Indicates the increased difficulty level; Indicates the current difficulty level; Indicates the visual acuity difficulty adjustment step size; Indicates that based on feedback results , auxiliary functions of key information K and value information V within the multi-task adaptive model; Indicates the preset recognition rate; represents the correct recognition rate; Indicates the rounding symbol;

[0107] Obtaining a first new set matching the actual distance after the change from the level-distance database, and conducting a visual acuity test of the visual test subject at the updated level of difficulty, wherein the first new set includes test questions related to the size of the sight mark and the speed of movement at the updated level;

[0108] If the correct recognition rate is less than the preset recognition rate, the difficulty level is lowered according to the feedback result, and a second new set matching the actual changed distance is obtained from the level-distance database to continue the visual acuity test.

[0109] In this embodiment, the difficulty level is a different test difficulty level pre-set in the visual acuity test.

[0110] In this embodiment, the correct recognition rate is the ratio of the number of visual acuity test questions correctly answered by the test-taker to the total number of questions at a certain difficulty level. The calculation function is written in Python language. By traversing the database table storing the feedback results, the records of correctly answered questions are screened out, and their number is counted. The correct recognition rate is then calculated with the total number of questions at this level, and the preset recognition rate is 90%.

[0111] In this embodiment, the difficulty of the visual acuity test is increased to a higher level, making subsequent visual acuity test questions more difficult. This can be achieved by reducing the size of the sight mark, increasing its speed, and increasing the complexity of the questions. For example, when increasing from difficulty level 3 to difficulty level 4, the sight mark becomes smaller and moves faster, further increasing the visual acuity requirements of the test taker. Conditional judgment code is written in Java. When the correct recognition rate meets the conditions, the variable representing the difficulty level is incremented by 1, and the relevant function is called to reset the sight mark parameters.

[0112] In this embodiment, the level-distance database is used to store the correspondence between different difficulty levels and the actual distance between the display screen and the tester's eyes, as well as a database of test questions corresponding to each difficulty level. The database records the difficulty levels corresponding to various distance ranges, as well as test question information related to the size and movement speed of the sight mark included in each difficulty level. For example, a distance of 30-40 cm corresponds to difficulty level 3, which contains multiple sight mark test questions of specific sizes and movement speeds.

[0113] In this embodiment, the distance after actual change is the distance between the display screen and the tester's eyes that actually changes during the test. For example, during the test, the tester picks up the tablet computer originally placed on the table and brings it close to the eyes, and the distance changes from 40 cm to 30 cm.

[0114] In this embodiment, the first new set is a new set of visual acuity test questions retrieved from the level-distance database based on the actual changed distance and the increased difficulty level. This set includes several test questions related to sight mark size and movement speed at the updated level, and is used to retest the test subject's visual acuity. For example, when the difficulty level increases from 3 to 4 and the actual distance becomes 30 cm, the set of visual acuity test questions retrieved from the database that is suitable for difficulty level 4 and matches the 30 cm distance is the first new set. The level-distance data is stored in an SQLite database, and database queries are performed using the Python sqlite3 library to obtain the corresponding set of test questions. It should be noted that the principles of the second new set are similar to those of the first new set and will not be repeated here.

[0115] In this embodiment, when py≤pz<1, the difficulty is adjusted proportionally through more sophisticated calculations, taking into account the gap between the correct recognition rate and the preset recognition rate. This takes into account the situation where the tester has a certain correct recognition rate but not perfect, so that the difficulty adjustment fits the actual level of the tester and avoids excessive or insufficient adjustment. The value of py is 0.9.

[0116] The auxiliary function A(Q, K, V) comprehensively considers the feedback result Q, the key information K and the value information V of the multi-task adaptive model. It does not rely solely on the correct recognition rate, but incorporates other important information in the test process into the difficulty adjustment calculation, comprehensively considering the tester's answer situation and the internal state of the model, making the difficulty adjustment more scientific and reasonable. The value range of is (0, 1.5), and ,in, 、 、 is the weight, with values ​​of 0.4, 0.3, and 0.3. 、 、 The mapping functions are based on the comprehensive feedback result Q, the key information K within the multi-task adaptive model, and the value information V, respectively. The comprehensive feedback result Q, the key information K within the multi-task adaptive model, and the value information V are obtained from the feedback-value comparison table, the key information-value comparison table, and the value-value comparison table, respectively. The feedback-value comparison table is used to convert the test-taker's feedback result Q into a reference table for the numerical mapping value f(Q). The feedback result Q contains various information about the test-taker during the visual acuity test, such as the number of correct or incorrect answers, the response time, and the user's answering behavior (click, voice input, etc.). The key information-value comparison table is used to convert the key information K within the multi-task adaptive model into a table for the numerical mapping value f(K). The key information K comes from within the model, such as the neuron activation values ​​of specific layers, the degree of parameter correlation between different task branches, and the weight distribution of the model when processing different visual features (such as the shape and color of the sight mark). For example, the value obtained from this table is 0.3. The value-value table is a reference table that converts value information V into numerical mapping values ​​f(V). Value information V is related to the model's emphasis on different test tasks and the importance of different visual features. For example, in some cases, recognizing the shape of the sight mark in a visual acuity test may be more important than its color, or the model may focus more on contrast sensitivity testing tasks in specific test scenarios. For example, if the model focuses more on visual acuity testing tasks, and the size of the sight mark is more critical than its orientation in a visual acuity test, the matching value is 0.6.

[0117] In this embodiment, the value of γ is 1.

[0118] The beneficial effects of the above technical solution are: when the correct recognition rate reaches or exceeds the preset recognition rate, the difficulty level is increased, which can promptly adapt to the test subject's better vision performance, further explore the test subject's visual ability, and make the test results more discriminative and accurate. A new set matching the actual changed distance and the increased difficulty level is obtained from the level-distance database and retested, fully accounting for changes in the test environment and test subject's ability. When the correct recognition rate falls below the preset recognition rate, the difficulty level is lowered and a second new set is obtained for further testing, allowing for timely adjustment of test difficulty to accommodate the test subject's poorer vision performance.

[0119] The present invention provides a functional vision testing method for performing visual health self-monitoring, which is aimed at contrast sensitivity testing and includes:

[0120] An initial contrast is set, and test questions with the same initial contrast are randomly presented on the display screen based on the initial contrast, and a feedback vector of each contrast test question is obtained from the visual test subject, where the feedback vector f = {Pj, tj}, where Pj represents the judgment result of the j-th contrast test question and tj represents the time taken to answer the j-th contrast test question;

[0121] Input the feedback vector f into the multi-task adaptive model, fuse the bar information corresponding to the contrast test question and the user interaction information to form a high-dimensional feature;

[0122] Dynamically updating the initial contrast according to all high-dimensional features under the initial contrast and continuing testing;

[0123]

[0124] in, An auxiliary function representing all high-dimensional features U under the initial contrast, key information K within the multi-task adaptive model, and value information V; Indicates the contrast after dynamic update; represents the initial contrast; Indicates the contrast difficulty adjustment step size.

[0125] In this embodiment, the initial contrast is a baseline contrast value set at the beginning of a contrast sensitivity test, typically based on a population average or a preliminary assessment by the test subject. For example, in medical vision testing, the initial contrast may be set to a moderate level (e.g., 50%), meaning that the brightness difference between the test pattern and the background is 50%.

[0126] In this embodiment, the visual test content generated based on the initial contrast is typically a grating pattern with different spatial frequencies (e.g., a sinusoidal grating). For example, a test item might be a vertical grating pattern with a contrast of 50% and a spatial frequency of 3 cycles / degree. The test software randomly selects the grating direction (horizontal / vertical / diagonal) and spatial frequency (e.g., 1, 3, 6, or 12 cycles / degree) to ensure randomness and objectivity of the test.

[0127] In this example, for example, the average answer time is 3.2 seconds per question (standard deviation 1.1 seconds), the accuracy rate for low spatial frequency (1-3 cycles / degree) is 92%, the accuracy rate for high spatial frequency (12 cycles / degree) is 65%, and the answer time is positively correlated with spatial frequency (r=0.62).

[0128] In this embodiment, the physical properties of the stripe pattern in the test question include contrast, spatial frequency, direction, etc. For example, a stripe information vector can be expressed as [contrast = 40%, spatial frequency = 6 cycles / degree, direction = vertical].

[0129] In this embodiment, the user interaction information is the interaction behavior data between the tester and the system, such as the duration of answering questions, mouse trajectory, click location, etc. For example, the hesitation time of the user in judging the visibility of the bar may reflect the uncertainty of their visual perception.

[0130] In this embodiment, high-dimensional features are abstract feature representations formed by fusing bar grid information with user interaction information. For example, a neural network maps raw data into a 128-dimensional feature space, where each dimension represents a different visual or behavioral characteristic. This high-dimensional feature space captures a richer set of test-taker visual perception characteristics, providing a foundation for personalized testing.

[0131] In this embodiment, the auxiliary function B(U, K, V) comprehensively considers multiple factors and does not rely solely on the tester's judgment or a single piece of information. The high-dimensional feature U integrates the bar information and user interaction information, and can fully reflect the tester's visual perception and operation behavior of the current contrast; the key information K and value information V reflect the internal operating status of the model and the degree of attention paid to different factors. The comprehensive consideration of multiple factors makes the contrast adjustment more comprehensive and scientific, in line with actual test needs, and The value of is 1.

[0132] ,in, 、 、 is the weight, with values ​​of 0.4, 0.3, and 0.3. It is a mapping value function based on the high-dimensional feature U, and the high-dimensional feature U is matched from the feature-value comparison table. The feature-value comparison table is a reference table for converting the high-dimensional feature U into a numerical mapping value f(U).

[0133] The beneficial effects of the above technical solution are: by randomizing test questions, the tester's memory effect and expectation bias are reduced, and the dynamic update mechanism significantly improves the accuracy of threshold estimation and test efficiency.

[0134] The present invention provides a functional vision testing method for performing self-monitoring of visual health, specifically for color vision testing, comprising:

[0135] Capturing the visual tester's dragging behavior of a color block for each color vision test question on a display screen, wherein the color block dragging behavior includes: a full dragging trajectory of each test color block and individual dragging trajectories involved in the full dragging trajectory, wherein the individual dragging trajectories are consistent with the number of drags;

[0136] Analyzing the dragging error probability of each test color block based on the color block dragging behavior, and generating a color gamut distribution in combination with the dragging position deviation of each test color block;

[0137] The color vision test level is updated according to the color gamut distribution, and the test is continued.

[0138] In this embodiment, dragging a color block refers to the tester's act of dragging a color block in the color vision test on the display screen. A full drag trajectory is the complete path the color block moves on the display screen from the start to the end of the drag. For example, if a tester drags a red block from the left side of the screen to the right side, the path from the left starting point to the right end point is the full drag trajectory. Individual drag trajectories are the trajectories formed by each independent drag action within the full drag trajectory. For example, if a tester drags the red block from the left side to the right side in three steps, the trajectory formed by each drag action is an individual drag trajectory, and the number of individual drag trajectories is the same as the number of drags. Taking a web-based color vision test program developed based on HTML5 and JavaScript as an example, by monitoring the mousedown (mouse press), mousemove (mouse move), and mouseup (mouse lift) events, the coordinate changes of the color block on the display screen are recorded, thereby obtaining the full drag trajectory and individual drag trajectories. For touch screen devices, similar functionality can be achieved through the touchstart, touchmove, and touchend events.

[0139] In this embodiment, the drag position deviation is the difference between the position where the tester finally drags the color block and the correct position.

[0140] In this embodiment, a mapping relationship between color gamut distribution and color vision test levels is established. Color vision test level adjustment rules corresponding to different color gamut distribution characteristics are pre-set. For example, if the color gamut distribution shows that the test subject makes many errors in distinguishing low-level colors, the color vision test level is lowered; if the test subject performs well in distinguishing high-level colors, the color vision test level is increased. Programming is used to automatically determine and update the color vision test level based on the generated color gamut distribution. Color vision test questions of the corresponding level are then retrieved from the test question library to continue testing the test subject.

[0141] The beneficial effects of the above technical solution are: accurately capturing the color block dragging behavior, providing a rich and accurate data basis for subsequent analysis of the tester's color vision, generating a color gamut distribution by analyzing the dragging error probability and position deviation, and being able to intuitively present the tester's color vision in a visual manner, and dynamically updating the color vision test level according to the color gamut distribution, so that the test difficulty can better match the tester's color vision ability.

[0142] The present invention provides a functional vision testing method for visual health self-monitoring, which analyzes the dragging error probability of each test color block based on the color block dragging behavior, and generates a color gamut distribution in combination with the dragging position deviation of each test color block, including:

[0143] The time it takes for each unit color block to pass through a point in each full drag trajectory involved in each color vision test question is counted to construct a trajectory vector ,in, Indicates the position of the second unit color block in the full drag trajectory and residence time ; Indicates the position of the mth unit color block in the full drag trajectory and residence time ;

[0144] Filtering positions whose dwell time is greater than a preset time from the trajectory vector as first positions, and splitting the entire dragging trajectory according to the first positions;

[0145] Determine the color difference between the color identifier of the unit color block at the last position in each individual dragging trajectory and the color of the test identifier, and obtain the color difference vector and position difference vector corresponding to the entire dragging trajectory;

[0146] When the number of the individual dragging tracks is 1, the color difference vector and the position difference vector are kept unchanged;

[0147] When there are multiple independent drag trajectories, determining the number of drags at each first position and the drag direction of each drag, and drawing a drag vector diagram;

[0148] determining, based on the drag vector diagram, the invalidity of dragging between the last position in each individual dragging track and the first position in the next individual dragging track, adjusting the color difference and the position difference of the last position, and updating the color difference vector and the position difference vector;

[0149] Determine the number of test color blocks in each color gamut for each color vision test question, and construct a color difference matrix based on the final color difference vector to obtain a color difference feature vector for the corresponding color gamut. Simultaneously, construct a position difference matrix based on the final position difference vector to obtain a position difference feature vector for the corresponding color gamut.

[0150] The drag error probability of each test color block is obtained based on the final color difference vector, and the error vertical value of the corresponding color gamut is obtained by combining the color difference feature vector and the position difference feature vector;

[0151] Generate color gamut distribution based on the error vertical values ​​under different color gamuts.

[0152] In this embodiment, the color vision test questions are designed to test the test taker's color vision ability, typically consisting of a combination of different color blocks. For example, to move a green block A1 from position 1 to position 2, the coordinate values ​​of each touch operation are obtained by overriding the View's onTouchEvent method. These coordinate values ​​are recorded in chronological order, forming a dragging trajectory for the block. A counter is also set to count the number of drags, thereby determining the number of individual drag trajectories.

[0153] In this embodiment, the drag position deviation is the difference between the actual placement of the test block and the ideal placement when the tester drags the test block to the target area. For example, the green block should ideally be placed in the center of the green area, but the tester placed it 2 cm from the center. This 2 cm is the drag position deviation.

[0154] In this embodiment, the color gamut distribution comprehensively depicts the distribution of testers' perception and manipulation of different colors, based on the error probability and position deviation of test blocks of different colors. It can be displayed graphically (such as a heat map or radar chart). In Matplotlib, a bar chart, radar chart, or heat map is drawn with color category as the horizontal axis and the combined score of error probability and position deviation as the vertical axis to display the color gamut distribution. For example, for testers with normal color vision, the average error probability for dragging blue and yellow blocks is 5%, and the average position deviation is 0.5 cm. For testers with abnormal color vision, the average error probability for dragging red and green blocks is 40%, and the average position deviation is 1.5 cm. The generated color gamut distribution graph visually shows that testers with red and green color vision exhibit significantly higher error probabilities and larger position deviations in the red and green areas.

[0155] In this embodiment, when the color gamut distribution shows that the test subject has a low probability of dragging errors for multiple colors and a small position deviation, it is judged that the test subject has good color vision ability, and the test level can be increased, and more difficult color vision test questions can be retrieved from the question bank, such as color block combination questions with similar colors; when the color gamut distribution shows that the test subject has a high probability of dragging errors and a large position deviation, especially in certain key color areas (such as the red and green areas for red-green color blindness detection), it is judged that the test subject has poor color vision ability, and the test level is lowered to provide simpler color block combination questions with large color differences.

[0156] In this embodiment, the preset time is 1 second.

[0157] In this embodiment, Figure 2 As shown, there are color blocks 1, 2, 3, 4, 5 and 6. At this time, it is necessary to move from color block 1 to color block 6 to pass the test. At this time, color block 2 and color block 5 are in the corresponding first positions. At this time, the individual dragging tracks are: color block 1--color block 2, color block 3--color block 4--color block 5, color block 6.

[0158] In this embodiment, the number of separate drag tracks is 1, indicating that the entire drag track is not split.

[0159] At this time: color difference vector = {color difference between color block 2 and color block 6, color difference between color block 5 and color block 6, color difference between color block 6 and color block 6}.

[0160] Position difference vector = {position difference between color block 2 and color block 6, position difference between color block 5 and color block 6, position difference between color block 6 and color block 6}, where the values ​​corresponding to the position difference between color block 6 and color block 6 and the color difference between color block 6 and color block 6 are 0.

[0161] In this embodiment, the drag vector diagram is a diagram constructed by the drag direction and number of times of each first position. This is to avoid invalid dragging due to screen freeze, but time statistics are performed on the time of passing points, resulting in difference analysis errors. Among them, the direction and length of each drag are the corresponding drag vectors, and all drag vectors constitute the drag vector diagram.

[0162] In this embodiment, the drag vector map is a map of the remaining tracks except for the last single drag track.

[0163] In this embodiment, the drag invalidity is: if: the sum of the lengths of each drag direction / (the length of the unit color block × the number of drags in the drag vector diagram of the last position in the individual drag trajectory) is greater than or equal to 1, then the drag invalidity is 1;

[0164] Otherwise, the drag invalidity is: the sum of the lengths in each drag direction / (the length of the unit color block × the number of drags in the drag vector diagram of the last position in the individual drag trajectory).

[0165] In this embodiment, if the drag invalidity is 1, the color difference corresponding to the last position in the independent drag track is reduced by half and the position difference is reduced by half.

[0166] In this embodiment, if the drag invalidity is not 1, then the updated color difference = the original color difference × drag invalidity × (1 / 2); the updated position difference = the original position difference × drag invalidity × (1 / 2).

[0167] In this embodiment, the color gamut may be green, blue, and red.

[0168] In this embodiment, the test color block refers to a designated color block that needs to be moved to a desired location, that is, the final location of the test color block.

[0169] In this embodiment, ,

[0170] Location .

[0171] In this embodiment, the eigenvector is obtained by performing eigenvalue solving on the matrix, which belongs to the prior art.

[0172] In this embodiment, the drag error probability is the ratio of the number of times the color difference in the corresponding last color difference vector is non-zero (that is, the number of times the erroneous operation occurs) to (the total number of elements in the last color difference vector + 1). It should be noted that the matrix is ​​an alignment matrix, and the remaining vectors are filled based on the number of the largest elements in the corresponding vector. If: there are 3 elements and the vector U1 is {0.1, 0}, then after filling, it is: {0.1, 0, 0}; if the vector U1 is {0.1, 0.2}, then after filling, it is: {0.1, 0.2, 0.2}.

[0173] In this embodiment, the error vertical value of the color gamut = ,in, They represent the feature averages of the corresponding color difference feature vectors and position difference feature vectors, The final color difference vector is the average of the sum of the drag error probabilities of each test color block. 、 Indicates that They are mapped to between 0 and 1 respectively, and then the two are added.

[0174] In this embodiment, for example, the error vertical value of the green color gamut is 2, the error vertical value of the red color gamut is 2.3, and the error vertical value of the blue color gamut is 3.1. At this time, a bar graph can be drawn to obtain the color gamut distribution.

[0175] The beneficial effects of the above technical solution are: by capturing the color block dragging behavior, detailed operation information of the tester during the color vision test can be obtained, and by analyzing the dragging error probability and position deviation to generate the color gamut distribution, the tester's color vision situation can be presented in a visual manner. The color vision test level is updated according to the color gamut distribution and the test is continued, which can dynamically adapt the test difficulty to the tester's color vision ability.

[0176] The present invention provides a functional vision testing method for performing self-monitoring of visual health, generating a remote vision assessment report, comprising:

[0177] Obtaining a visual acuity threshold based on all visual acuity test items at different difficulty levels and the fusion features of each visual acuity test item;

[0178] Based on all contrast sensitivity test items at different difficulty levels of the contrast sensitivity test and the fusion features of each contrast sensitivity test item, a contrast sensitivity curve is obtained;

[0179] Averaging each color gamut error vertical value in the color gamut distribution of color vision test questions at different difficulty levels of the color vision test to obtain a color vision arrangement error distribution;

[0180] A remote vision assessment report is generated based on the visual acuity threshold, the contrast sensitivity curve, and the color perception error distribution.

[0181] In this embodiment, fused features are a set of features formed by integrating multiple features of the test content (such as optotype features) and the test-taker's operational behavior (such as response time and click location). For example, the fused features of a visual acuity test question might include information such as optotype size (5' visual angle), color (black), the test-taker's reaction time to click the correct answer (2 seconds), and the deviation between the click location and the correct answer location (0.3 cm).

[0182] In this embodiment, the color vision arrangement error distribution is a result obtained by averaging the error vertical values ​​of each color gamut in the color gamut distribution, and is used to comprehensively reflect the error distribution of the tester in the color vision test.

[0183] In this embodiment, in the contrast sensitivity test software, test questions containing different contrast and spatial frequency combinations are displayed in sequence according to different preset difficulty levels. The Python Matplotlib library is used to draw a contrast sensitivity curve based on the calculated spatial frequency and contrast data.

[0184] First, integrate data such as visual acuity thresholds, contrast sensitivity curves, and color permutation error distribution. Then, using a report generation template (which can be in Word, PDF, or other formats), populate the template with the data in the form of charts (such as a bar chart of visual acuity thresholds, a line chart of contrast sensitivity curves, and a radar chart of color permutation error distribution) and text descriptions. For example, the text description section interprets the test results, analyzes the tester's visual function, and provides an assessment of whether there are vision problems or color vision abnormalities for reference. Recommendations for further examination or correction are also provided as appropriate. Finally, the generated report is sent to the tester or relevant professionals (such as doctors and optometrists) via a network (e.g., email, online platforms, etc.).

[0185] The beneficial effect of the above technical solution is that by obtaining the above-mentioned vision test results, it is possible to systematically understand the tester's errors and characteristics in the aforementioned vision test, generate a remote vision assessment report, and present complex vision test data to relevant personnel in a concise and intuitive format. This not only makes it easier for testers to understand their own visual health status, but also provides professionals with a preliminary basis for rapid evaluation.

[0186] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A functional vision test method for self-monitoring of visual health, characterized in that: include: Load the multi-task adaptive model and the parameters of the self-supervised pre-training editor to the visual tester's smart device. After loading, the test questions for visual acuity test, contrast sensitivity test, and color vision test are presented as required. Capturing the visual tester's operation behavior based on each test question based on the smart device; Input the test content and operation behavior of each test question into the self-supervised pre-training editor to extract fusion features; Based on all fusion features under each test and in combination with the difficulty level of each test item based on the current distance between the display screen and the visual test subject, the visual acuity threshold, contrast sensitivity curve, and color permutation error distribution of the visual test subject are obtained, and a remote visual assessment report is generated; Among them, the visual acuity test includes: When the visual tester triggers the visual acuity test operation on the display screen, a presentation set at a difficulty level matching the current distance is output based on the display screen, wherein the presentation set includes a plurality of visual acuity test questions at an initial recognition difficulty level, and the visual acuity test questions are related to the size of the sight mark and the speed of the sight mark movement; Determining a feedback result of the vision test subject on each visual acuity test item based on the presentation set, wherein the feedback result is related to the operation behavior; After determining the feedback result of the visual test subject for each visual acuity test question, the method further includes: adjusting the sight mark according to the feedback result and the actual distance after the change, specifically including: Obtaining a correct recognition rate based on the feedback results at the difficulty level; If the correct recognition rate is greater than or equal to a preset recognition rate, increasing the difficulty level according to the feedback result; in, Indicates the increased difficulty level; Indicates the current difficulty level; Indicates the visual acuity difficulty adjustment step size; Indicates that based on feedback results , auxiliary functions of key information K and value information V within the multi-task adaptive model; Indicates the preset recognition rate; represents the correct recognition rate; Indicates the rounding symbol; Obtaining a first new set matching the actual distance after the change from the level-distance database, and conducting a visual acuity test of the visual test subject at the same difficulty level again, wherein the first new set includes test questions related to the size of the sight mark and the speed of the sight mark movement at the updated level; If the correct recognition rate is less than the preset recognition rate, the difficulty level is lowered according to the feedback result, and a second new set matching the actual changed distance is obtained from the level-distance database to continue the visual acuity test.

2. The functional vision testing method for self-monitoring of visual health according to claim 1, characterized in that: Also includes: The test results of each loaded smart device are collected, and the federated learning mechanism is used to perform local training on the collected data to update the parameters of the multi-task adaptive model and the self-supervised pre-training editor.

3. The functional vision testing method for self-monitoring of visual health according to claim 1, characterized in that: Contrast sensitivity testing, including: An initial contrast is set, and test questions with the same initial contrast are randomly presented on the display screen based on the initial contrast, and a feedback vector of each contrast test question is obtained from the visual test subject, where the feedback vector f = {Pj, tj}, where Pj represents the judgment result of the j-th contrast test question and tj represents the time taken to answer the j-th contrast test question; Input the feedback vector f into the multi-task adaptive model, fuse the bar information corresponding to the contrast test question and the user interaction information to form a high-dimensional feature; Dynamically updating the initial contrast according to all high-dimensional features under the initial contrast and continuing testing; in, Represents all high-dimensional features under the initial contrast , auxiliary functions of key information K and value information V within the multi-task adaptive model; Indicates the contrast after dynamic update; represents the initial contrast; Indicates the contrast difficulty adjustment step size.

4. The functional vision testing method for self-monitoring of visual health according to claim 1, characterized in that: Color vision tests include: Capturing the visual tester's dragging behavior of a color block for each color vision test question on a display screen, wherein the color block dragging behavior includes: a full dragging trajectory of each test color block and individual dragging trajectories involved in the full dragging trajectory, wherein the individual dragging trajectories are consistent with the number of drags; Analyzing the dragging error probability of each test color block based on the color block dragging behavior, and generating a color gamut distribution in combination with the dragging position deviation of each test color block; The color vision test level is updated according to the color gamut distribution, and the test is continued.

5. The functional vision testing method for self-monitoring of visual health according to claim 4, characterized in that: Analyzing the dragging error probability of each test color block based on the color block dragging behavior and combining the dragging position deviation of each test color block to generate a color gamut distribution, including: The time it takes for each unit color block to pass through a point in each full drag trajectory involved in each color vision test question is counted to construct a trajectory vector ,in, Indicates the position of the second unit color block in the full drag trajectory and residence time ; Indicates the position of the mth unit color block in the full drag trajectory and residence time ; Filtering positions whose dwell time is greater than a preset time from the trajectory vector as first positions, and splitting the entire dragging trajectory according to the first positions; Determine the color difference between the color identifier of the unit color block at the last position in each individual dragging trajectory and the color of the test identifier, and obtain the color difference vector and position difference vector corresponding to the entire dragging trajectory; When the number of the individual dragging tracks is 1, the color difference vector and the position difference vector are kept unchanged; When there are multiple independent drag trajectories, determining the number of drags at each first position and the drag direction of each drag, and drawing a drag vector diagram; determining, based on the drag vector diagram, the invalidity of dragging between the last position in each individual dragging track and the first position in the next individual dragging track, adjusting the color difference and the position difference of the last position, and updating the color difference vector and the position difference vector; Determine the number of test color blocks in each color gamut for each color vision test question, and construct a color difference matrix based on the final color difference vector to obtain a color difference feature vector for the corresponding color gamut. Simultaneously, construct a position difference matrix based on the final position difference vector to obtain a position difference feature vector for the corresponding color gamut. Obtaining a drag error probability of each test color block according to the color difference vector, and obtaining an error vertical value of the corresponding color gamut by combining the color difference feature vector and the position difference feature vector; Generate color gamut distribution based on the error vertical values ​​under different color gamuts.

6. The functional vision testing method for self-monitoring of visual health according to claim 1, characterized in that: Generates a remote visual assessment report including: Obtaining a visual acuity threshold based on all visual acuity test items at different difficulty levels and the fusion features of each visual acuity test item; Based on all contrast sensitivity test items at different difficulty levels of the contrast sensitivity test and the fusion features of each contrast sensitivity test item, a contrast sensitivity curve is obtained; Averaging each color gamut error vertical value in the color gamut distribution of color vision test questions at different difficulty levels of the color vision test to obtain a color vision arrangement error distribution; A remote vision assessment report is generated based on the visual acuity threshold, the contrast sensitivity curve, and the color perception error distribution.

Images