An interactive visual training and control method for strabismus patients

By using real-time eye-tracking acquisition and closed-loop interactive training control, the problem of the inability to dynamically adapt the training difficulty in existing systems has been solved, thus realizing the effectiveness and safety of visual training for strabismus patients and improving the authenticity and efficiency of training.

CN123101366APending Publication Date: 2026-11-17HUBEI QUASI OPTOMETRY TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611568557.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-10
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing digital strabismus visual training systems cannot dynamically and adaptively adjust the intensity of convergence and divergence stimuli based on the patient's real-time training performance. The pace of difficulty increase does not match the patient's actual rehabilitation ability, resulting in limited training efficiency and easily causing visual fatigue or diplopia discomfort.

Method used

By constructing a closed-loop interactive training logic using real-time eye-tracking acquisition, the training process is dynamically adjusted and its effectiveness is verified through training task partitioning, eye-tracking calibration and initialization, real-time interactive control, visual fatigue warning, and multi-dimensional adaptive difficulty adjustment, ensuring that the training difficulty matches the patient's rehabilitation ability.

Benefits of technology

Effectively identify spurious training behaviors, avoid eye strain and double vision, improve the authenticity and efficiency of training, ensure that the training difficulty matches the patient's ability, and improve rehabilitation results.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

This invention discloses an interactive visual training control method for strabismus patients, belonging to the field of visual rehabilitation training technology. It constructs a closed-loop interactive training logic through real-time eye movement acquisition, which can identify false training completion caused by passive viewing and abandons the traditional fixed-sequence playback mode. Combining convergence-divergence fusion, binocular separation perception, and fixation stability zone training, it adaptively adjusts the intensity of convergence and separation stimuli in small steps based on the patient's real-time training performance to match the patient's actual rehabilitation ability. Simultaneously, it introduces a visual fatigue warning and pause mechanism to avoid problems such as excessive difficulty inducing diplopia and visual fatigue, and insufficient stimulation due to insufficient difficulty. Furthermore, it adds training effectiveness verification and iterative correction, as well as training data archiving and evaluation, to achieve a complete closed loop of training, evaluation, and parameter correction, effectively solving the technical defects of rigid training processes, poor interactive realism, inability to dynamically adapt difficulty, and limited rehabilitation efficiency.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of visual rehabilitation training technology, and in particular relates to an interactive visual training control method for strabismus patients. Background Technology

[0002] Strabismus is a common binocular visual dysfunction in clinical practice. Patients cannot keep their visual axes parallel, and their eyes cannot fixate on the same target at the same time. When one eye fixates on a target, the other eye deviates. It is often accompanied by a series of binocular visual impairments such as diplopia, inhibition, abnormal retinal correspondence, and stereopsis defects. It not only affects appearance, but also causes visual fatigue, reading errors, and decreased attention, which seriously restricts patients' learning, work and daily life.

[0003] Most existing digital strabismus visual training systems use a preset fixed process to play training materials and switch training tasks according to a fixed sequence. They simply complete the image output and lack closed-loop interactive control logic based on the patient's real-time eye movement and gaze response.

[0004] The training process cannot dynamically perceive the patient's actual gaze state, nor can it distinguish whether the patient is truly completing binocular fusion response or passively viewing the image, which can easily lead to patients falsely completing training tasks. The training difficulty parameters are mostly manually preset, and cannot dynamically and adaptively adjust the intensity of convergence and divergence stimuli based on the patient's real-time training performance. The pace of difficulty increase does not match the patient's actual rehabilitation ability. If the difficulty is too high, it can easily cause visual fatigue or even induce diplopia discomfort. If the difficulty is too low, it will not provide effective training stimulation, thus limiting rehabilitation efficiency. Therefore, there is an urgent need for an interactive visual training control method for strabismus patients to solve the above problems. Summary of the Invention

[0005] The purpose of this invention is to address the problem that existing training methods cannot dynamically and adaptively adjust the intensity of convergent and divergent stimuli based on the patient's real-time training performance, and that the pace of difficulty increase does not match the patient's actual rehabilitation ability. Therefore, this invention proposes an interactive visual training control method for strabismus patients.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] An interactive visual training and control method for strabismus patients includes the following steps: S1. Patient basic information collection and training task zoning modeling: Input the patient's strabismus type, strabismus degree, and binocular vision function test report, complete the establishment of the patient's basic vision function file, and divide the training scene into three major functional training areas based on the strabismus rehabilitation mechanism: convergence and divergence fusion training area, binocular separation perception training area, and fixation stability training area. Retrieve the strabismus vision function training database and match the basic stimulus parameter thresholds for different training areas. S2. Eye Tracking Calibration and Training Initialization: Synchronize the eye tracking acquisition device with the display terminal, execute the binocular eye tracking calibration process, collect the patient's binocular baseline fixation parameters, load the basic training materials for the corresponding training zone, set the initial convergence and divergence stimulus intensity, separation stimulus intensity, and material switching time window, and complete the training session initialization. S3. Real-time eye movement acquisition and closed-loop interactive training execution: After the training session starts, the pupil position, gaze offset angle and eye movement trajectory data of both eyes are collected synchronously, and real-time interactive control is carried out on the convergence and divergence fusion training area, binocular separation perception training area and fixation stability training area. S4. Instantaneous training status verification and fatigue risk warning: During the training process, the binocular fixation deviation, effective fixation duration, and frequency of high-frequency eye jitter are statistically analyzed in real time to calculate the patient's real-time visual fatigue index. When the visual fatigue index exceeds the preset safety threshold, the current training task is automatically paused and a rest prompt is displayed. The training session is resumed after the rest is confirmed, thus avoiding high-intensity training that may induce diplopia and eye fatigue discomfort. S5. Multi-dimensional training performance evaluation and adaptive difficulty adjustment: The patient's task completion accuracy, average response time and binocular synchronization index are comprehensively evaluated in each training cycle. The convergence and divergence stimulus intensity parameters are dynamically adjusted based on the evaluation results. Multi-level difficulty gradient adjustment is executed in sequence. A buffer transition interval is set during the difficulty increase process to prevent parameter jumps. After each round of difficulty adjustment, the training parameters and patient feedback are recorded, and the training session log is saved. S6. Training effectiveness verification and training parameter iterative correction: Extract all eye movement raw data and task completion data of this training session and compare them with the patient's preset rehabilitation target indicators. If the training effectiveness achievement rate is lower than the set threshold, retrieve the strabismus visual function training database to readjust the basic stimulus parameters and difficulty adjustment step size, and return to S3 to start closed-loop interactive training again. S7. Secondary assessment and archiving of rehabilitation status after the session: After the end of a single training session, retrieve the full-cycle training data to generate a training report, conduct a secondary assessment of convergence and divergence fusion ability, binocular separation perception ability, fixation stability, update the patient's personal visual function profile, and output the recommended initial parameters for the next training session.

[0008] As a further description of the above technical solution: In S1, the training task zones are based on the strabismus clinical rehabilitation guidelines and historical patient rehabilitation cases. The convergence and divergence fusion training area is mainly used to train patients' ability to converge and diverge their eyes. The binocular separation perception training area is used to improve binocular object separation recognition and suppress abnormal monocular fixation. The fixation stabilization training area is used to improve fixation drift, nystagmus, and fixation positioning deviation.

[0009] At the same time, the threshold values ​​of basic stimulation parameters for each region are compensated and corrected based on the patient's strabismus degree and age.

[0010] As a further description of the above technical solution:

[0011] In step S2, the specific procedure for binocular eye movement calibration is as follows:

[0012] Multiple sets of calibration target points are output sequentially, guiding the patient to fixate on the calibration target points with both eyes in turn, and collecting the baseline coordinates of the pupils of both eyes.

[0013] After completing the baseline calibration, load the training material, set the initial convergence and divergence stimulus intensity and the separation stimulus intensity, and lock the synchronization error between the eye-tracking device and the display terminal to ≤5ms.

[0014] As a further description of the above technical solution:

[0015] In S3, the real-time interactive control method is as follows:

[0016] Every 20ms, a set of binocular eye movement data is collected to analyze the patient’s real binocular fixation response and distinguish between real fusion completion and passive viewing false completion.

[0017] When the system determines that the patient has effectively completed the target interactive task, it triggers a dynamic switch of training materials.

[0018] If the completion is determined to be false, the current training stimulus image is maintained without switching, and a prompt is output to guide the patient to complete the binocular coordination response again.

[0019] As a further description of the above technical solution:

[0020] In S3, the convergence and divergence fusion training area collects the convergence and divergence movement offset angles of both eyes. Only when both eyes synchronously complete the corresponding convergence and divergence movements and maintain an effective fixation duration of ≥350ms is it determined to be a true fusion completion, and the training material can be switched. If only one eye fixates in place, it is determined to be a false completion, and the task switch is not executed.

[0021] Binocular separation perception training area: Collect the difference in gaze position between the two eyes, identify whether there is monocular inhibition, and the task is considered completed only when both eyes recognize the corresponding separated stimulus image.

[0022] Fixation stabilization training area: continuously collects eye tremor and drift data. When the eye deviation exceeds the allowable deviation threshold, fixation is judged to have failed and fixation guidance is retried.

[0023] Eye movement data is collected every 20ms, and fixation and time response curves are output. The data are compared with the strabismus visual function training database to identify false training completion behavior.

[0024] As a further description of the above technical solution:

[0025] In S4, the visual fatigue index is calculated based on the frequency of high-frequency eye tremors, the number of times binocular fixation is lost, and subjective feedback. When the visual fatigue index reaches the warning threshold, training is immediately paused, and a forced rest period of 30-120 seconds is set. The training session resumes after the rest period ends and is confirmed.

[0026] As a further description of the above technical solution:

[0027] In S5, multi-level difficulty gradient adjustment control is implemented.

[0028] Difficulty increase adjustment: When the effective completion rate of the task is ≥85% for multiple consecutive training cycles, gradually increase the intensity of convergent and divergent stimuli and increase the training difficulty in small steps.

[0029] Difficulty maintenance adjustment: When the effective completion rate of the task is in the range of 60%-85%, the current training stimulus parameters are kept unchanged;

[0030] Difficulty reduction adjustment: When the effective completion rate of the task is less than 60%, the intensity of convergence and divergence stimuli is automatically reduced to decrease the training load;

[0031] The difficulty adjustment step size should not be too large, and the change in convergence and divergence stimulation in a single adjustment should not exceed 2 prism degrees, and the change in separation stimulation intensity in a single adjustment should not exceed 1 level; avoid stimulation parameter ranges that are easy to induce diplopia during the training process; the refresh rate of the training screen output should be stable, and the screen latency of the training display terminal should be ≤20ms.

[0032] As a further description of the above technical solution:

[0033] In S6, the training preset target indicators are: effective completion rate of convergence and divergence fusion training area ≥80%; effective completion rate of binocular separation perception training area ≥75%; effective fixation ratio of fixation stability training area ≥82%; if the effective achievement rate of training in any area does not meet the target, return to S3 to correct the stimulus parameters and difficulty step size and start training again.

[0034] As a further description of the above technical solution:

[0035] In step S7, a training report is generated while recording the patient's subjective feedback, updating the patient's visual function profile, outputting the recommended initial convergence and divergence stimulus parameters for the next training session, and storing the training data in encrypted form so that doctors can easily view the rehabilitation trend.

[0036] In summary, due to the adoption of the above technical solutions, the beneficial effects of this invention are as follows: By constructing a closed-loop interactive training logic through real-time eye-tracking acquisition, it can identify false training completion caused by patients passively watching, thus abandoning the traditional fixed-sequence playback mode; by combining convergence-divergence fusion, binocular separation perception, and fixation stability zone training, it adaptively adjusts the intensity of convergence and separation stimuli in small steps based on the patient's real-time training performance to match the patient's actual rehabilitation ability; at the same time, it introduces a visual fatigue warning and pause mechanism to avoid the problems of excessive difficulty inducing diplopia and visual fatigue, and insufficient stimulation due to excessive difficulty; and it adds training effectiveness verification, iterative correction, and training data archiving and evaluation links to achieve a complete closed loop of training, evaluation, and parameter correction, effectively solving the technical defects of rigid training process, poor interactive authenticity, inability to dynamically adapt difficulty, and limited rehabilitation efficiency. Detailed Implementation

[0037] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0038] In its specific implementation, the present invention provides a technical solution:

[0039] An interactive visual training and control method for strabismus patients includes the following steps:

[0040] S1. Patient basic information collection and training task zoning modeling: Input the patient's strabismus type, strabismus degree, and binocular vision function test report, complete the establishment of the patient's basic vision function file, and divide the training scene into three major functional training areas based on the strabismus rehabilitation mechanism: convergence and divergence fusion training area, binocular separation perception training area, and fixation stability training area. Retrieve the strabismus vision function training database and match the basic stimulus parameter thresholds for different training areas.

[0041] In S1, the training task zones are based on the strabismus clinical rehabilitation guidelines and historical patient rehabilitation cases.

[0042] The convergence and divergence fusion training area is mainly used to train patients' ability to converge and diverge their eyes.

[0043] The binocular separation perception training area is used to improve binocular object separation recognition and suppress abnormal monocular fixation.

[0044] The fixation stabilization training area is used to improve fixation drift, nystagmus, and fixation positioning deviation.

[0045] At the same time, the threshold values ​​of basic stimulation parameters for each region are compensated and corrected based on the patient's strabismus degree and age.

[0046] S2. Eye Tracking Calibration and Training Initialization: Synchronize the eye tracking acquisition device with the display terminal, execute the binocular eye tracking calibration process, collect the patient's binocular baseline fixation parameters, load the basic training materials for the corresponding training zone, set the initial convergence and divergence stimulus intensity, separation stimulus intensity, and material switching time window, and complete the training session initialization.

[0047] In S2, the specific procedure for binocular eye movement calibration is as follows:

[0048] Multiple sets of calibration target points are output sequentially, guiding the patient to fixate on the calibration target points with both eyes in turn, and collecting the baseline coordinates of the pupils of both eyes.

[0049] After completing the baseline calibration, load the training material, set the initial convergence and divergence stimulus intensity and the separation stimulus intensity, and lock the synchronization error between the eye-tracking device and the display terminal to ≤5ms.

[0050] S3. Real-time eye movement acquisition and closed-loop interactive training execution: After the training session starts, the pupil position, gaze offset angle and eye movement trajectory data of both eyes are collected synchronously, and real-time interactive control is carried out on the convergence and divergence fusion training area, binocular separation perception training area and fixation stability training area.

[0051] In S3, the real-time interactive control method is as follows:

[0052] Every 20ms, a set of binocular eye movement data is collected to analyze the patient’s real binocular fixation response and distinguish between real fusion completion and passive viewing false completion.

[0053] When the system determines that the patient has effectively completed the target interactive task, it triggers a dynamic switch of training materials.

[0054] If the completion is determined to be false, the current training stimulus image is maintained without switching, and a prompt is output to guide the patient to complete the binocular coordination response again.

[0055] In S3, the convergence and divergence fusion training area collects the convergence and divergence movement offset angles of both eyes. Only when both eyes synchronously complete the corresponding convergence and divergence movements and maintain an effective fixation duration of ≥350ms is it considered a true fusion completion, and switching training materials is allowed; if only one eye fixates in place, it is considered a false completion, and task switching is not performed.

[0056] Binocular separation perception training area: Collect the difference in gaze position between the two eyes, identify whether there is monocular inhibition, and the task is considered completed only when both eyes recognize the corresponding separated stimulus image.

[0057] Fixation stabilization training area: continuously collects eye tremor and drift data. When the eye deviation exceeds the allowable deviation threshold, fixation is judged to have failed and fixation guidance is retried.

[0058] Eye movement data is collected every 20ms, and fixation and time response curves are output. The data are compared with the strabismus visual function training database to identify false training completion behavior.

[0059] S4. Instantaneous training status verification and fatigue risk warning: During the training process, the binocular fixation deviation, effective fixation duration, and frequency of high-frequency eye jitter are statistically analyzed in real time to calculate the patient's real-time visual fatigue index. When the visual fatigue index exceeds the preset safety threshold, the current training task is automatically paused and a rest prompt is displayed. The training session is resumed after the rest is confirmed, thus avoiding high-intensity training that may induce diplopia and eye fatigue discomfort.

[0060] In S4, the visual fatigue index is calculated based on the frequency of high-frequency eye tremors, the number of times binocular fixation is lost, and subjective feedback. When the visual fatigue index reaches the warning threshold, training is immediately paused, and a forced rest period of 30-120 seconds is set. The training session resumes after the rest period ends and is confirmed.

[0061] S5. Multi-dimensional training performance evaluation and adaptive difficulty adjustment: The patient's task completion accuracy, average response time and binocular synchronization index are comprehensively evaluated in each training cycle. The convergence and divergence stimulus intensity parameters are dynamically adjusted based on the evaluation results. Multi-level difficulty gradient adjustment is executed in sequence. A buffer transition interval is set during the difficulty increase process to prevent parameter jumps. After each round of difficulty adjustment, the training parameters and patient feedback are recorded, and the training session log is saved.

[0062] In S5, multi-level difficulty gradient adjustment control:

[0063] Difficulty increase adjustment: When the effective completion rate of the task is ≥85% for multiple consecutive training cycles, gradually increase the intensity of convergent and divergent stimuli and increase the training difficulty in small steps.

[0064] Difficulty maintenance adjustment: When the effective completion rate of the task is in the range of 60%-85%, the current training stimulus parameters are kept unchanged;

[0065] Difficulty reduction adjustment: When the effective completion rate of the task is less than 60%, the intensity of convergence and divergence stimuli is automatically reduced to decrease the training load;

[0066] The difficulty adjustment step size should not be too large, and the change in convergence and divergence stimulation in a single adjustment should not exceed 2 prism degrees, and the change in separation stimulation intensity in a single adjustment should not exceed 1 level; avoid stimulation parameter ranges that are easy to induce diplopia during the training process; the refresh rate of the training screen output should be stable, and the screen latency of the training display terminal should be ≤20ms.

[0067] S6. Training effectiveness verification and training parameter iterative correction: Extract all eye movement raw data and task completion data of this training session and compare them with the patient's preset rehabilitation target indicators. If the training effectiveness achievement rate is lower than the set threshold, retrieve the strabismus visual function training database to readjust the basic stimulus parameters and difficulty adjustment step size, and return to S3 to start closed-loop interactive training again.

[0068] In S6, the training preset target indicators are: effective completion rate of tasks in the convergence and divergence fusion training area ≥80%; effective completion rate of tasks in the binocular separation perception training area ≥75%; effective fixation rate in the fixation stability training area ≥82%; if the effective achievement rate of training in any area does not meet the target, return to S3 to correct the stimulus parameters and difficulty step size and start training again.

[0069] S7. Secondary assessment and archiving of rehabilitation status after the session: After the end of a single training session, retrieve the full-cycle training data to generate a training report, conduct a secondary assessment of convergence and divergence fusion ability, binocular separation perception ability, fixation stability, update the patient's personal visual function profile, and output the recommended initial parameters for the next training session.

[0070] In S7, training reports are generated while recording patients' subjective feedback, updating patients' visual function profiles, and outputting recommended initial convergence and divergence stimulus parameters for the next training session. Training data is stored in encrypted form, making it easy for doctors to view rehabilitation trends.

[0071] In all embodiments of this invention, the test subjects were intermittent exotropia subjects with a strabismus degree of 9-14 prism diopters and insufficient fusion function. Before training, standardized binocular vision function tests were completed, and personal vision function files were established. The training was divided into convergence and divergence fusion training area, binocular separation perception training area, and fixation stability training area.

[0072] Example 1:

[0073] S1. Complete the collection of basic patient information and zonal modeling of training tasks, retrieve the strabismus visual function training database, and match the threshold of basic stimulus parameters for each training area.

[0074] S2. Eye Tracking Calibration and Training Initialization: The eye tracking acquisition device and the display terminal are synchronized in time to complete the binocular eye tracking calibration, with the timestamp synchronization error controlled at 4ms; training materials are loaded, and the initial convergence and divergence stimulus parameters are configured to complete the training session initialization;

[0075] S3. Real-time eye tracking acquisition and closed-loop interactive training execution, acquiring binocular eye tracking data every 20ms; effective fixation judgment threshold of 350ms in convergence-divergence training area; distinguishing between real completion and fake completion, without switching the training screen for fake completion, and outputting fixation and time response curves;

[0076] S4. Training instantaneous status verification and fatigue risk warning: Real-time calculation of visual fatigue index, triggering threshold to automatically pause training, force a 60-second rest, and resume training after confirmation;

[0077] S5. Multi-dimensional training performance evaluation and adaptive difficulty adjustment; step size adjustment is limited according to the present invention, single convergent and divergent stimuli are adjusted to 1.5 prism degrees, and separation stimuli are adjusted to level 1; training session logs are saved;

[0078] S6. Training effectiveness verification and training parameter iterative correction; Partition detection: The effective completion rate of the convergence-dispersion fusion training area was 83.7%, the effective completion rate of the binocular separation perception training area was 79.2%, and the effective fixation rate of the fixation stability training area was 85.1%, all of which met the preset indicators and did not require returning to correct parameters.

[0079] S7. After the session ends, a secondary assessment of the rehabilitation status is conducted and archived, a training report is generated, the patient's visual function profile is updated, and recommended initial stimulation parameters for the next training session are output.

[0080] This embodiment has a complete set of mechanisms including eye movement closed-loop discrimination, visual fatigue warning, and small-step adaptive parameter adjustment; the training process can identify false completion behavior, and the subjects do not experience diplopia or obvious eye strain or discomfort; the training difficulty dynamically adapts to the patient's actual visual function performance, and the training stimulus is effective.

[0081] Example 2:

[0082] S1. Complete the collection of basic patient information and zonal modeling of training tasks, retrieve the strabismus visual function training database, and match the threshold of basic stimulus parameters for each training area.

[0083] S2. Eye Tracking Calibration and Training Initialization: The eye tracking acquisition device and the display terminal are synchronized in time to complete the binocular eye tracking calibration, with the timestamp synchronization error controlled at 4ms; training materials are loaded, and the initial convergence and divergence stimulus parameters are configured to complete the training session initialization;

[0084] S3. Cancel the logic for judging the true and false completion of eye movements, do not analyze the binocular fixation response, and automatically switch the training material screen according to a preset fixed time.

[0085] S4. Training instantaneous status verification and fatigue risk warning: Real-time calculation of visual fatigue index, triggering threshold to automatically pause training, force a 60-second rest, and resume training after confirmation;

[0086] S5, multi-dimensional training performance evaluation and adaptive difficulty adjustment; single convergent and divergent stimulus adjustment to 1.5 prism power, separate stimulus adjustment to level 1; saving training session logs;

[0087] S6. Training effectiveness verification and training parameter iterative correction; The statistical system showed a high task completion rate. After removing the passive viewing false samples, the actual effective completion rate in the convergence-dispersion fusion training area was 56.1%, and the actual effective completion rate in the binocular separation perception training area was 53.4%, which did not meet the preset target.

[0088] S7. After the session, a second assessment of the rehabilitation status is conducted and archived to generate a training report.

[0089] This embodiment omits eye-tracking closed-loop interaction discrimination and relies on fixed-sequence switching of materials; the surface completion rate statistically obtained by the system is inflated, with a large number of passive viewing behaviors constituting sham training behavior. Patients do not truly complete binocular coordinated response, resulting in insufficient actual training stimulation and reduced rehabilitation benefits. Verifying eye-tracking closed-loop discrimination is a key step in ensuring the authenticity of training.

[0090] Example 3:

[0091] S1. Complete the collection of basic patient information and zonal modeling of training tasks, retrieve the strabismus visual function training database, and match the threshold of basic stimulus parameters for each training area.

[0092] S2. Eye Tracking Calibration and Training Initialization: The eye tracking acquisition device and the display terminal are synchronized in time to complete the binocular eye tracking calibration, with the timestamp synchronization error controlled at 4ms; training materials are loaded, and the initial convergence and divergence stimulus parameters are configured to complete the training session initialization;

[0093] S3. Real-time eye tracking acquisition and closed-loop interactive training execution, acquiring binocular eye tracking data every 20ms; effective fixation judgment threshold of 350ms in convergence-divergence training area; distinguishing between real completion and fake completion, without switching the training screen for fake completion, and outputting fixation and time response curves;

[0094] S4. Training instantaneous status verification and fatigue risk warning: Real-time calculation of visual fatigue index, triggering threshold to automatically pause training, force a 60-second rest, and resume training after confirmation;

[0095] S5, multi-dimensional training performance evaluation and adaptive difficulty adjustment; set single convergent stimulus adjustment to 3.2 prism power, single separation stimulus adjustment to 3 levels, with significant parameter jumps; save training session logs;

[0096] S6. Verification of training effectiveness and iterative correction of training parameters; Post-training statistics: The effective completion rate of the convergence-dispersion fusion training area was 53.5%, and the effective completion rate of the binocular separation perception training area was 51.8%, which did not meet the preset indicators; During the training process, the subjects reported experiencing transient diplopia and significant eye fatigue;

[0097] S7. After the session, a second assessment of the rehabilitation status is conducted and archived to generate a training report.

[0098] In this embodiment, the difficulty adjustment step size was too large, and the sudden change in stimulus intensity exceeded the patient's tolerance range, inducing diplopia and eye fatigue, and significantly reducing the effective completion rate of the task. This proves that the difficulty adjustment step size must be controlled within the range defined in this invention. Parameter jumps will cause training failure and bring the risk of eye discomfort.

[0099] Comparative example:

[0100] Traditional fixed-process digital strabismus training solution

[0101] The same subjects as in Example 1 were selected, but no training task partitioning was set; there was no real-time closed-loop eye-tracking interaction logic, and the materials were played in a fixed sequence; the training difficulty parameters were manually fixed before training, and there was no adaptive difficulty adjustment during the training process; there was no automatic eye fatigue warning and pause mechanism; and there was no iterative verification of training effectiveness and secondary evaluation and archiving.

[0102] Test results: The test failed to detect fraudulent training behavior throughout the entire training process; some stages were too difficult, causing eye strain and fatigue in the subjects, while others were too easy, resulting in insufficient training stimulation; the actual effective completion rate for the convergence-fusion training area was 50.2%, the actual effective completion rate for the binocular separation perception training area was 48.7%, and the effective fixation rate for the fixation stability training area was 57.3%; these results indicate that traditional fixed-process training cannot adapt to the real-time rehabilitation level of patients, and the training quality is uncontrollable.

[0103] Performance items Example 1 Example 2 Example 3 Comparative Example The actual effective completion rate (%) of the convergence and decentralization training area 83.7 56.1 53.5 50.2 True and effective completion rate (%) of binocular separation perception training area 79.2 53.4 51.8 48.7 Effective fixation percentage (%) in the fixation stability training zone 85.1 61.3 59.6 57.3 Fake task identification capability Can identify fake completion Unable to identify fake completion Can identify fake completion Unable to identify fake completion Subjective experience of the subjects No double vision, mild fatigue No obvious discomfort, but there is sham training. Double vision and significant eye strain Eye strain and fatigue, due to insufficient stimulation during some training sessions.

[0104] In summary, this invention provides a complete solution for training task partitioning modeling, eye movement calibration initialization, real-time eye movement closed-loop interactive training, fatigue risk warning, multi-dimensional adaptive difficulty adjustment, iterative verification of training effectiveness, and secondary archiving of rehabilitation status. This solution enables closed-loop interactive strabismus training based on the patient's actual eye movement fixation response, identifies false completion of training behavior, adaptively adjusts the difficulty according to the patient's training performance, avoids the risks of diplopia and visual fatigue, and significantly improves the real effectiveness of digital visual training for strabismus.

Claims

1. An interactive visual training and control method for strabismus patients, characterized in that, Includes the following steps: S1. Patient basic information collection and training task zoning modeling: Input the patient's strabismus type, strabismus degree, and binocular vision function test report, complete the establishment of the patient's basic vision function file, and divide the training scene into three major functional training areas based on the strabismus rehabilitation mechanism: convergence and divergence fusion training area, binocular separation perception training area, and fixation stability training area. Retrieve the strabismus vision function training database and match the basic stimulus parameter thresholds for different training areas. S2. Eye Tracking Calibration and Training Initialization: Synchronize the eye tracking acquisition device with the display terminal, execute the binocular eye tracking calibration process, collect the patient's binocular baseline fixation parameters, load the basic training materials for the corresponding training zone, set the initial convergence and divergence stimulus intensity, separation stimulus intensity, and material switching time window, and complete the training session initialization. S3. Real-time eye movement acquisition and closed-loop interactive training execution: After the training session starts, the pupil position, gaze offset angle and eye movement trajectory data of both eyes are collected synchronously, and real-time interactive control is carried out on the convergence and divergence fusion training area, binocular separation perception training area and fixation stability training area. S4. Instantaneous training status verification and fatigue risk warning: During the training process, the binocular fixation deviation, effective fixation duration, and frequency of high-frequency eye jitter are statistically analyzed in real time to calculate the patient's real-time visual fatigue index. When the visual fatigue index exceeds the preset safety threshold, the current training task is automatically paused and a rest prompt is displayed. The training session is resumed after the rest is confirmed, thus avoiding high-intensity training that may induce diplopia and eye fatigue discomfort. S5. Multi-dimensional training performance evaluation and adaptive difficulty adjustment: The patient's task completion accuracy, average response time and binocular synchronization index are comprehensively evaluated in each training cycle. The convergence and divergence stimulus intensity parameters are dynamically adjusted based on the evaluation results. Multi-level difficulty gradient adjustment is executed in sequence. A buffer transition interval is set during the difficulty increase process to prevent parameter jumps. After each round of difficulty adjustment, the training parameters and patient feedback are recorded, and the training session log is saved. S6. Training effectiveness verification and training parameter iterative correction: Extract all eye movement raw data and task completion data of this training session and compare them with the patient's preset rehabilitation target indicators. If the training effectiveness achievement rate is lower than the set threshold, retrieve the strabismus visual function training database to readjust the basic stimulus parameters and difficulty adjustment step size, and return to S3 to start closed-loop interactive training again. S7. Secondary assessment and archiving of rehabilitation status after the session: After the end of a single training session, retrieve the full-cycle training data to generate a training report, conduct a secondary assessment of convergence and divergence fusion ability, binocular separation perception ability, fixation stability, update the patient's personal visual function profile, and output the recommended initial parameters for the next training session.

2. The interactive visual training control method for strabismus patients according to claim 1, characterized in that, In S1, the training task zones are based on the strabismus clinical rehabilitation guidelines and historical patient rehabilitation cases. The convergence and divergence fusion training area is mainly used to train patients' ability to converge and diverge their eyes. The binocular separation perception training area is used to improve binocular object separation recognition and suppress abnormal monocular fixation. The fixation stabilization training area is used to improve fixation drift, nystagmus, and fixation positioning deviation. At the same time, the threshold values ​​of basic stimulation parameters for each region are compensated and corrected based on the patient's strabismus degree and age.

3. The interactive visual training and control method for strabismus patients according to claim 1, characterized in that, In step S2, the specific procedure for binocular eye movement calibration is as follows: Multiple sets of calibration target points are output sequentially, guiding the patient to fixate on the calibration target points with both eyes in turn, and collecting the baseline coordinates of the pupils of both eyes. After completing the baseline calibration, load the training material, set the initial convergence and divergence stimulus intensity and the separation stimulus intensity, and lock the synchronization error between the eye-tracking device and the display terminal to ≤5ms.

4. The interactive visual training and control method for strabismus patients according to claim 1, characterized in that, In S3, the real-time interactive control method is as follows: Every 20ms, a set of binocular eye movement data is collected to analyze the patient’s real binocular fixation response and distinguish between real fusion completion and passive viewing false completion. When the system determines that the patient has effectively completed the target interactive task, it triggers a dynamic switch of training materials. If the completion is determined to be false, the current training stimulus image is maintained without switching, and a prompt is output to guide the patient to complete the binocular coordination response again.

5. The interactive visual training control method for strabismus patients according to claim 4, characterized in that, In S3, the convergence and divergence fusion training area collects the convergence and divergence movement offset angles of both eyes. Only when both eyes synchronously complete the corresponding convergence and divergence movements and maintain an effective fixation duration of ≥350ms is it determined to be a true fusion completion, and the training material can be switched. If only one eye fixates in place, it is determined to be a false completion, and the task switch is not executed. Binocular separation perception training area: Collect the difference in gaze position between the two eyes, identify whether there is monocular inhibition, and the task is considered completed only when both eyes recognize the corresponding separated stimulus image. Fixation stabilization training area: continuously collects eye tremor and drift data. When the eye deviation exceeds the allowable deviation threshold, fixation is judged to have failed and fixation guidance is retried. Eye movement data is collected every 20ms, and fixation and time response curves are output. The data are compared with the strabismus visual function training database to identify false training completion behavior.

6. The interactive visual training control method for strabismus patients according to claim 1, characterized in that, In S4, the visual fatigue index is calculated based on the frequency of high-frequency eye tremors, the number of times binocular fixation is lost, and subjective feedback. When the visual fatigue index reaches the warning threshold, training is immediately paused, and a forced rest period of 30-120 seconds is set. The training session resumes after the rest period ends and is confirmed.

7. The interactive visual training control method for strabismus patients according to claim 1, characterized in that, In S5, multi-level difficulty gradient adjustment control is implemented. Difficulty increase adjustment: When the effective completion rate of the task is ≥85% for multiple consecutive training cycles, gradually increase the intensity of convergent and divergent stimuli and increase the training difficulty in small steps. Difficulty maintenance adjustment: When the effective completion rate of the task is in the range of 60%-85%, the current training stimulus parameters are kept unchanged; Difficulty reduction adjustment: When the effective completion rate of the task is less than 60%, the intensity of convergence and divergence stimuli is automatically reduced to decrease the training load; The difficulty adjustment step size should not be too large, and the change in convergence and divergence stimulation in a single adjustment should not exceed 2 prism degrees, and the change in separation stimulation intensity in a single adjustment should not exceed 1 level; avoid stimulation parameter ranges that are easy to induce diplopia during the training process; the refresh rate of the training screen output should be stable, and the screen latency of the training display terminal should be ≤20ms.

8. The interactive visual training control method for strabismus patients according to claim 1, characterized in that, In S6, the training preset target indicators are: effective completion rate of convergence and divergence fusion training area ≥80%; effective completion rate of binocular separation perception training area ≥75%; effective fixation ratio of fixation stability training area ≥82%; if the effective achievement rate of training in any area does not meet the target, return to S3 to correct the stimulus parameters and difficulty step size and start training again.

9. The interactive visual training control method for strabismus patients according to claim 1, characterized in that, In step S7, a training report is generated while recording the patient's subjective feedback, updating the patient's visual function profile, outputting the recommended initial convergence and divergence stimulus parameters for the next training session, and storing the training data in encrypted form so that doctors can easily view the rehabilitation trend.