Eye position control force evaluation method and system for intermittent strabismus

Abstract

The invention discloses an intermittent strabismus-oriented eye position control force evaluation method and system, and belongs to the technical field of vision detection.The method comprises the steps that when preset video content is played, a camera device is used for conducting image collection on the left eye and the right eye of a user, and a plurality of frames of eye position images and the physiological offset amount of the user are obtained; according to the eye position image and the physiological offset, calculating the relative position of the iris of the two eyes in the eye socket to obtain the deviation value of the two eyes of each frame of eye position image; based on the two-eye deviation value, performing potential strabismus frame identification on the eye position image, and outputting a strabismus attack event and a strabismus total duration during a playing period; according to the total effective watching duration and the total strabismus duration of the user, the fluctuation condition of the eye position control force of the user is analyzed, and a control force evaluation result is obtained. Therefore, by implementing the method and the device, objective evaluation of the eye position control force can be realized, and data support is provided for evaluating the development trend of intermittent strabismus and surgical decisions.

Description

Technical Field

[0001] This application belongs to the field of vision testing technology, specifically relating to a method and system for assessing eye position control in intermittent strabismus. Background Technology

[0002] Intermittent strabismus is an abnormal state of eye position control that lies between orthotropia and manifest eye deviation. Its main characteristic is the fluctuation of "eye position control," meaning that users can maintain orthotropia when they are focused or alert, but may experience eye deviation when they are fatigued, distracted, or in poor physical condition. Therefore, quantifying a user's eye position control can help determine whether there is intermittent instability in eye position control.

[0003] Traditional methods for assessing eye alignment control typically involve covering one eye and testing the other. However, in this scenario, users are often highly alert and may appear to be in the correct eye alignment, thus masking potential eye alignment deficiencies and affecting measurement accuracy. Furthermore, current technologies often employ the Newcastle Score to assess eye alignment control. This score relies heavily on subjective recall and lacks continuous, objective quantitative data. Additionally, there is a lack of dynamic analysis methods for tracking trends in eye alignment control, making it difficult to continuously assess eye alignment stability. Summary of the Invention

[0004] This application proposes a method and system for assessing eye position control in intermittent strabismus, enabling objective assessment of eye position control and providing data support for evaluating the development trend of intermittent strabismus and surgical decisions.

[0005] The first aspect of this application provides a method for assessing eye position control for intermittent strabismus, the method comprising: While playing the preset video content, the camera device is used to capture images of the user's left and right eyes respectively, and obtain several frames of eye position images and the user's physiological offset. Based on the eye position image and the physiological bias, the binocular deviation value of each frame of the eye position image is obtained by calculating the relative position of the two irises in the eye socket. Based on the binocular deviation value, the potential strabismus frame is identified in the eye position image, and the strabismus attack events and total strabismus duration during playback are output. Based on the user's total effective fixation time and the total strabismus time, the fluctuation of the user's eye position control is analyzed to obtain the control ability assessment result.

[0006] The aforementioned scheme continuously stimulates the user's eyes with video content of interest to induce changes in eye position, allowing the user to be in a natural and relaxed state. Images are captured from both eyes separately, obtaining eye position images containing eye position information, as well as physiological offset values ​​representing facial structural asymmetry and shooting deviations to eliminate errors in subsequent eye position deviation assessments. Then, the relative positions of the irises in the eye sockets are detected using the captured eye position images, enabling dynamic monitoring of eye position status and identifying image frames where strabismus attacks occur. Based on the number of image frames corresponding to strabismus attacks, the total duration of strabismus during video playback can be determined, providing objective evidence for assessing changes in eye position control. Finally, the proportion of total strabismus duration to total effective fixation time is used to quantitatively assess fluctuations in eye position control, providing objective and repeatable quantitative indicators for eye position control assessment. This comprehensively reflects the stability and changing patterns of the user's eye position control ability, providing strong data support for assessing the development trend of the user's eye position control ability and assisting doctors in developing personalized intervention plans tailored to the user.

[0007] In one possible implementation of the first aspect, while playing preset video content, a camera device is used to capture images of the user's left and right eyes respectively, obtaining several frames of eye position images and the user's physiological offset, specifically as follows: Play the video content pre-selected by the user. Before the playback time reaches a first threshold, use a camera device to collect the user's binocular features and calculate the iris position ratio of the left and right eyes based on the binocular features. Before reaching the first threshold, the user's eyes are assumed to be in a normal position. The difference between the iris position ratio of the left eye and the iris position ratio of the right eye is used as the physiological bias. After the playback time reaches the first threshold, the camera device is used to capture images of the user's left and right eyes respectively to obtain the eye position images.

[0008] The above scheme assumes that the user is not in a relaxed state and that their eyes are in the correct position at the beginning of image acquisition. Therefore, during this period, the user's binocular features are collected to calculate the physiological bias, eliminate the influence of facial anatomical differences (such as eye socket asymmetry, natural iris position deviation) and shooting errors on the detection results, and improve the accuracy of subsequent judgments.

[0009] In one possible implementation of the first aspect, based on the eye position image and the physiological bias, the binocular deviation value of each frame of the eye position image is obtained by calculating the relative positions of the two irises in the eye sockets, specifically as follows: For each frame of the eye position image, the user's eyelid closure degree and head posture are detected, and the eye position images that fail the detection are marked as invalid frames; Invalid frames are removed from the eye position images to obtain valid image frames. For each valid image frame, the facial landmark algorithm is used to locate the user's eyes to obtain the first horizontal distance from the iris center to the inner canthus of the eye socket and the second horizontal distance from the inner canthus of the eye socket to the outer canthus of the eye socket. Based on the first horizontal distance and the second horizontal distance, calculate the ratio of the iris positions of both eyes in each effective image frame; The binocular deviation value is obtained based on the iris position ratio and the physiological offset.

[0010] The above scheme first removes images with non-compliant user postures to improve data quality. In each valid image frame, the relative positions of the left and right irises in the eye sockets are detected, and the difference between the two is calculated as the binocular deviation value. This is used to determine whether the visual axes of the two eyes are separated, providing an objective basis for subsequent judgments on whether the user has strabismus.

[0011] In one possible implementation of the first aspect, the binocular deviation value is obtained based on the iris position ratio and the physiological offset, specifically as follows: Calculate the difference between the ratio of the iris positions of the left and right eyes in each effective image frame, subtract the physiological bias from the absolute value of the difference, and take the absolute value to obtain the binocular deviation value; The binocular deviation value is used to eliminate facial asymmetry of the user and systematic errors caused by the camera equipment during image acquisition.

[0012] In one possible implementation of the first aspect, based on the binocular deviation value, potential strabismus frame identification is performed on the eye position image, and the strabismus attack events and total strabismus duration during playback are output, specifically: The binocular deviation value of each frame of the eye position image is judged using a preset sliding time window. If the binocular deviation value is greater than a second threshold, the corresponding eye position image is marked as a potential strabismus frame. If a third consecutive threshold of potential strabismus frames exists within the sliding time window, the strabismus attack event is considered to have occurred, and the strabismus duration of the strabismus attack event is recorded. The total duration of strabismus is obtained by counting all the strabismus durations.

[0013] To prevent misjudgments caused by rapid eye saccades, the above scheme uses binocular deviation values ​​to determine which eye-position images show binocular visual axis separation (i.e., binocular deviation values ​​exceeding a set physiological tolerance), thereby accurately identifying potential strabismus frames where strabismus attacks may occur. Then, the strabismus duration is determined by the number of image frames, and the total strabismus duration is obtained by accumulating the strabismus duration across all time periods.

[0014] One possible implementation of the first aspect also includes: The cropped eye position image is input into a preset eye position state recognition model to calculate the eye position state confidence of each frame of the eye position image; wherein, the eye position state confidence includes the probability of orthotropia and the probability of strabismus. If the strabismus probability is greater than the fourth threshold, the corresponding eye position image is marked as the strabismus attack event.

[0015] The above scheme can also use end-to-end deep learning to directly output the eye position confidence of each image frame, thereby determining which image frames have strabismus attacks, without the need for complex formula calculations, making the detection method relatively simpler.

[0016] In one possible implementation of the first aspect, the fluctuation of the user's eye position control is analyzed based on the user's total effective fixation time and the total strabismus time to obtain a control ability assessment result, specifically as follows: The total effective gaze duration is obtained by subtracting the duration of all invalid frames from the duration of playing the video content. The ratio of the total strabismus duration to the total effective fixation duration is taken as the current strabismus time ratio; Based on the user ID, the user's historical test data is read. Combining the historical test data with the current strabismus time ratio, the fluctuation of the user's strabismus time ratio in the current time period is calculated, and a strabismus time ratio fluctuation curve is plotted. Based on the strabismus time-proportion fluctuation curve, the changing trend of the user's eye position control is evaluated to obtain the control ability evaluation result.

[0017] The above scheme determines the duration of decreased eye position control in the test based on the proportion of strabismus duration. It also assesses the fluctuation trend by combining the test data with historical test data to obtain the relative fluctuation of eye position control in the recent period. If eye position control declines, a follow-up examination is required, providing a reliable and objective basis for surgical decisions.

[0018] The second aspect of this application provides an eye position control assessment system for intermittent strabismus, the system comprising: a data acquisition module, a binocular deviation value calculation module, a strabismus event recognition module, and an eye position control assessment module; The data acquisition module is used to capture images of the user's left and right eyes using a camera device when playing preset video content, and obtain several frames of eye position images and the user's physiological offset. The binocular deviation value calculation module is used to calculate the binocular deviation value of each frame of eye position image by calculating the relative position of the two irises in the eye socket based on the eye position image and the physiological offset amount. The strabismus event recognition module is used to identify potential strabismus frames in the eye position image based on the binocular deviation value, and output the strabismus attack events and the total duration of strabismus during playback; The eye position control assessment module is used to analyze the fluctuation of the user's eye position control based on the user's total effective fixation time and the total strabismus time, and obtain the control ability assessment result.

[0019] A third aspect of this application provides a terminal device, the device comprising: a terminal device including a processor and a memory, the memory storing a computer program, wherein the processor executes the computer program to implement the steps of the eye position control assessment method for intermittent strabismus as described in any one of the embodiments of this application.

[0020] A fourth aspect of this application provides a storage medium storing computer-readable program code that, when executed, implements the steps of an eye position control assessment method for intermittent strabismus as described in any one of the embodiments of this application. Attached Figure Description

[0021] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0022] Figure 1 This is a schematic flowchart of a method for assessing eye position control in intermittent strabismus, provided in one embodiment of this application. Figure 2 This is a structural diagram of an eye position control assessment system for intermittent strabismus provided in one embodiment of this application; Figure 3 This is a structural diagram of a terminal device provided in an embodiment of this application. Detailed Implementation

[0023] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0024] It should be understood that the step numbers used in the text are for ease of description only and are not intended to limit the order in which the steps are performed.

[0025] First Embodiment The functional characteristic of intermittent strabismus is that users can usually control their eye position to be upright when they are concentrating, but their eye position is prone to deviate when they are visually fatigued or their attention is relaxed. Based on this, the embodiments of this application simulate the user's eye use state in a natural life scenario by playing video content of a certain duration. As the fixation time goes by, the user's ability to control their eye position gradually weakens, thereby exposing the hidden strabismus symptoms. This allows for the accurate quantification of the user's eye position control ability and provides reliable data support for subsequent assessment of the stability of the user's eye position control ability.

[0026] like Figure 1 As shown, to address the problem in existing technologies that lack objective quantification of the user's eye-position control, leading to inaccurate assessments of the development of intermittent strabismus, the first embodiment of this application provides a detailed flowchart of an eye-position control assessment method for intermittent strabismus. This embodiment's eye-position control assessment method for intermittent strabismus includes steps S1 to S4, detailed below: Step S1: While playing the preset video content, use a camera device to capture images of the user's left and right eyes respectively, and obtain several frames of eye position images and the user's physiological offset.

[0027] Before conducting intermittent strabismus testing on each user, a health profile is first established for each new user. Basic information such as the user's age and strabismus type (esotropia / exotropia) is entered, and a unique user ID is generated for each user to link this information. This user ID is used to link the user's intermittent strabismus test data from multiple tests on different dates and time periods, providing data support for subsequent assessments of the stability of the user's eye alignment control.

[0028] First, play video content that interests the user, such as cartoons or reading materials. This video content can be selected by the user or automatically selected based on the user's interests, and its duration can be set by the user.

[0029] After the video starts playing, image acquisition of the user's eyes begins. Initially, before the playback duration reaches a set threshold, the camera captures the user's binocular features. This is because, at the beginning of video playback, the user is assumed to be in a neutral position, focused on the screen. Furthermore, considering that everyone's face is not perfectly symmetrical (e.g., the inner corner of the left eye is larger than the right eye), and the camera angle may not be directly facing the face, the user's physiological offset is calculated based on these binocular features to avoid misinterpreting eye position as strabismus from certain angles. This eliminates systematic errors caused by individual facial anatomical asymmetry (such as differences in epicanthal folds) and non-central shooting angles of the device.

[0030] When calculating the binocular deviation value, the physiological offset amount needs to be subtracted to obtain the accurate eye position deflection amount.

[0031] Based on the aforementioned binocular characteristics, the ratio of iris position in the left and right eyes is calculated, and the difference between the ratios of iris position in the two eyes is set as the physiological bias of the user in this test.

[0032] Then, after the initial playback, the camera device is used to capture images of the user's left and right eyes respectively, resulting in multiple frames of eye position images.

[0033] Optionally, in this embodiment of the application, a monocular camera is used, and the images from the first 5 seconds of playback are used to calculate the physiological bias of the two eyes.

[0034] Step S2: Based on the eye position image and the physiological bias, calculate the relative positions of the irises of both eyes in the eye socket to obtain the binocular deviation value of each frame of the eye position image.

[0035] For each frame of eye position image, the user's eyelid closure and head posture are detected simultaneously. If blinking, eye closure, or large head turning is detected, the corresponding eye position image is marked as an invalid frame and is not included in the subsequent calculation of binocular deviation value.

[0036] This is because substandard eyelid closure and head posture can affect the judgment of eye position, making it easy to mistake closed eyes for strabismus. Therefore, invalid frames are removed first to improve the data quality of the remaining eye position images.

[0037] The data remaining after removing invalid frames are called valid image frames. For each valid image frame, a facial landmark algorithm is used to locate the user's two eyes. Based on the location results, the first horizontal distance from the iris center to the inner canthus of the orbit and the second horizontal distance from the inner canthus to the outer canthus of the orbit are calculated for both the left and right eyes. The quotient of the first horizontal distance and the second horizontal distance is used to obtain the iris position ratio (IPR) of the two eyes in each valid image frame.

[0038] The iris position ratio is the degree of positional deviation of the iris relative to the eye socket, and therefore can be used as one of the indicators for assessing whether strabismus has occurred.

[0039] For each valid image frame, the difference between the left eye iris position ratio and the right eye iris position ratio is calculated, and the physiological bias is subtracted to obtain the binocular deviation value of each eye position image. The specific calculation formula is as follows: Binocular deviation value = |(Left eye IPR - Right eye IPR) - Physiological bias |; Under normal circumstances, when a user gazes at the same target, the relative positions of the left and right eyes within the eye sockets should change synchronously; however, in cases of strabismus, the relative positions of the two eyes do not change synchronously. Therefore, the fluctuation of the binocular deviation value can be used to measure whether a strabismus event has occurred.

[0040] Step S3: Based on the binocular deviation value, perform potential strabismus frame identification on the eye position image, and output the strabismus attack events and total strabismus duration during playback.

[0041] To prevent misjudgments caused by rapid eye saccades, this application introduces a timing verification mechanism, which involves thresholding the binocular deviation value of a single frame image to identify potential strabismus frames that may contain strabismus events.

[0042] First, a fixed-size sliding window is constructed. The sliding window is used to detect the binocular deviation value of each frame of eye position image, identify eye position images that exceed a set threshold, and mark these eye position images as potential strabismus frames. If multiple potential strabismus frames appear consecutively within the sliding window, a strabismus attack event is considered to have occurred, and the duration of the strabismus corresponding to the event is recorded.

[0043] For example, a sliding window of size 200ms is constructed. Since the camera operates at 30fps, the occurrence of 6 consecutive potential strabismus frames within this window size is sufficient to determine that a strabismus attack has occurred. The first frame of the sliding window is marked as the start time of this strabismus attack to obtain the complete duration of the strabismus attack.

[0044] For the timing confirmation mechanism, if the binocular deviation value is greater than a set threshold, the corresponding eye position image is determined to be a potential strabismus frame; if the binocular deviation value is less than or equal to the set threshold, the corresponding eye position image is determined to be an orthostatic frame.

[0045] As an improvement to the above scheme, to simplify the complex geometric calculations, other embodiments may also employ a pre-trained eye position recognition model to process the eye position images. The eye position images are first cropped to a suitable size and then input into the eye position recognition model. The model's classifier directly outputs the eye position confidence score for each frame of the eye position image, including the probability of orthotropia and the probability of strabismus. If the strabismus probability consistently exceeds a set threshold, it is directly determined as a strabismus attack event.

[0046] Finally, the strabismus duration of all strabismus attacks is calculated to obtain the total strabismus duration, which is the sum of the time segments of all confirmed strabismus attacks during the video playback process, thus achieving the statistical analysis of the strabismus attack duration throughout the entire playback process.

[0047] Step S4: Based on the user's total effective fixation time and the total strabismus time, analyze the fluctuation of the user's eye position control ability to obtain the control ability assessment result.

[0048] Subtract the invalid frame time from the total duration of the video playback to obtain the user's effective gaze duration.

[0049] The ratio of total strabismus duration to total effective fixation duration is taken as the current strabismus time proportion in this test, and is abbreviated as STR. The specific calculation formula is as follows: STR = (Total duration of strabismus / Total duration of effective fixation) × 100%; Then, historical test data of intermittent strabismus is read using the user ID, and the standard deviation or coefficient of variation of STR is calculated by combining it with the current strabismus time ratio to determine the trend of the user's eye position control. The standard deviation represents the absolute fluctuation of eye position control, and the coefficient of variation represents the relative fluctuation of eye position control (the ratio of the standard deviation to the mean).

[0050] For example, the standard deviation or coefficient of variation of the STR values ​​for the most recent 7 days is calculated using historical test data and the current strabismus time ratio. If a user's STR value is 50%, it indicates severe strabismus, and the corresponding absolute value of 2% indicates stable strabismus; if a user's STR value is 2%, it indicates mild strabismus, and the corresponding absolute value of 2% indicates significant strabismus fluctuation.

[0051] In addition, historical test data can be combined to plot the recent strabismus time ratio fluctuation curve.

[0052] If the slope of the strabismus time-proportional fluctuation curve is greater than the set value, it is considered that the user's eye position control is declining and the user will be advised to have a follow-up examination; if the strabismus time-proportional fluctuation curve remains stable at a low level, it is considered that the eye position control is relatively stable.

[0053] Optionally, if the strabismus time-proportion fluctuation curve remains below 15% for a long period of time, and the single fluctuation amplitude does not exceed ±5%, then the eye position control is considered to be relatively stable.

[0054] In addition, for the quantitative assessment of eye position control, if the STR value is <1%, the eye position control is considered excellent; if 1% ≤ STR value < 15%, the eye position control is considered good; if 15% ≤ STR value < 50%, the eye position control is considered weakened; if the STR value is ≥ 50%, the eye position control is considered to be gradually lost, and a warning needs to be issued.

[0055] Therefore, through the above-mentioned quantitative evaluation methods, the control ability evaluation results of this test were obtained, realizing the objective quantification of the user's eye position control ability.

[0056] Implementing the embodiments of this application has the following beneficial effects: This application embodiment continuously stimulates the user's eyes with video content of interest to induce changes in eye position, allowing the user to be in a natural and relaxed state. Images are captured from both eyes separately, obtaining eye position images containing eye position information and physiological offset values ​​characterizing facial structural asymmetry and shooting deviations to eliminate errors in subsequent eye position deviation judgments. Then, the relative positions of the irises in the eye sockets are detected using the captured eye position images, achieving dynamic monitoring of eye position status and identifying image frames where strabismus attacks occur. Based on the number of image frames corresponding to strabismus attacks, the total duration of strabismus during video playback can be determined, providing objective evidence for assessing changes in eye position control. Finally, the fluctuation of eye position control is quantitatively assessed by the proportion of total strabismus duration to total effective fixation time, providing an objective and repeatable quantitative indicator for eye position control assessment. This comprehensively reflects the stability and variation patterns of the user's eye position control ability, providing strong data support for assessing the development trend of eye position control in users with intermittent strabismus and assisting doctors in developing personalized intervention plans tailored to the user.

[0057] Second Embodiment Furthermore, in order to implement the eye position control assessment system for intermittent strabismus corresponding to the above method embodiments, and to achieve the corresponding functions and technical effects, Figure 2 A structural diagram of an eye position control assessment system for intermittent strabismus is provided. For ease of explanation, only the parts relevant to this embodiment are shown. The eye position control assessment system for intermittent strabismus provided in this application embodiment includes: The data acquisition module 201 is used to acquire images of the user's left and right eyes using a camera device when playing preset video content, so as to obtain several frames of eye position images and the user's physiological bias.

[0058] In this embodiment, before conducting intermittent strabismus testing on each user, a health record is first established for each new user, recording basic information such as the user's age and strabismus type (esotropia / exotropia), and a unique user ID is generated for each user to associate this information. The user ID is used to associate the user's multiple intermittent strabismus test data at different dates and time periods, providing data support for subsequent evaluation of the stability of the user's eye position control.

[0059] First, play video content that interests the user, such as cartoons or reading materials. This video content can be selected by the user or automatically selected based on the user's interests, and its duration can be set by the user.

[0060] After the video starts playing, image acquisition of the user's eyes begins. Initially, before the playback duration reaches a set threshold, the camera captures the user's binocular features. This is because, at the beginning of video playback, the user is assumed to be in a neutral position, focused on the screen. Furthermore, considering that everyone's face is not perfectly symmetrical (e.g., the inner corner of the left eye is larger than the right eye), and the camera angle may not be directly facing the face, the user's physiological offset is calculated based on these binocular features to avoid misinterpreting eye position as strabismus from certain angles. This eliminates systematic errors caused by individual facial anatomical asymmetry (such as differences in epicanthal folds) and non-central shooting angles of the device.

[0061] When calculating the binocular deviation value, the physiological offset amount needs to be subtracted to obtain the accurate eye position deflection amount.

[0062] Based on the aforementioned binocular characteristics, the ratio of iris position in the left and right eyes is calculated, and the difference between the ratios of iris position in the two eyes is set as the physiological bias of the user in this test.

[0063] Then, after the initial playback, the camera device is used to capture images of the user's left and right eyes respectively, resulting in multiple frames of eye position images.

[0064] Optionally, in this embodiment of the application, a monocular camera is used, and the images from the first 5 seconds of playback are used to calculate the physiological bias of the two eyes.

[0065] The binocular deviation value calculation module 202 is used to obtain the binocular deviation value of each frame of eye position image by calculating the relative position of the two irises in the eye socket based on the eye position image and the physiological offset amount.

[0066] In this embodiment, for each frame of eye position image, the user's eyelid closure degree and head posture are detected simultaneously. If blinking, eye closure, or large head turning is detected, the corresponding eye position image is marked as an invalid frame and is not included in the subsequent calculation of binocular deviation value.

[0067] This is because substandard eyelid closure and head posture can affect the judgment of eye position, making it easy to mistake closed eyes for strabismus. Therefore, invalid frames are removed first to improve the data quality of the remaining eye position images.

[0068] The data remaining after removing invalid frames are called valid image frames. For each valid image frame, a facial landmark algorithm is used to locate the user's two eyes. Based on the location results, the first horizontal distance from the iris center to the inner canthus of the orbit and the second horizontal distance from the inner canthus to the outer canthus of the orbit are calculated for both the left and right eyes. The quotient of the first horizontal distance and the second horizontal distance is used to obtain the iris position ratio (IPR) of the two eyes in each valid image frame.

[0069] The iris position ratio is the degree of positional deviation of the iris relative to the eye socket, and therefore can be used as one of the indicators for assessing whether strabismus has occurred.

[0070] For each valid image frame, the difference between the left eye iris position ratio and the right eye iris position ratio is calculated, and the physiological bias is subtracted to obtain the binocular deviation value of each eye position image. The specific calculation formula is as follows: Binocular deviation value = |(Left eye IPR - Right eye IPR) - Physiological bias |; Under normal circumstances, when a user gazes at the same target, the relative positions of the left and right eyes within the eye sockets should change synchronously; however, in cases of strabismus, the relative positions of the two eyes do not change synchronously. Therefore, the fluctuation of the binocular deviation value can be used to measure whether a strabismus event has occurred.

[0071] The strabismus event recognition module 203 is used to identify potential strabismus frames in the eye position image based on the binocular deviation value, and output the strabismus attack events and the total duration of strabismus during playback.

[0072] In this embodiment of the application, in order to prevent misjudgment caused by rapid eye saccades, this embodiment of the application introduces a timing confirmation mechanism, that is, to perform threshold judgment on the binocular deviation value of a single frame image and identify potential strabismus frames that may have strabismus events.

[0073] First, a fixed-size sliding window is constructed. The sliding window is used to detect the binocular deviation value of each frame of eye position image, identify eye position images that exceed a set threshold, and mark these eye position images as potential strabismus frames. If multiple potential strabismus frames appear consecutively within the sliding window, a strabismus attack event is considered to have occurred, and the duration of the strabismus corresponding to the event is recorded.

[0074] For example, a sliding window of size 200ms is constructed. Since the camera operates at 30fps, the occurrence of 6 consecutive potential strabismus frames within this window size is sufficient to determine that a strabismus attack has occurred. The first frame of the sliding window is marked as the start time of this strabismus attack to obtain the complete duration of the strabismus attack.

[0075] For the timing confirmation mechanism, if the binocular deviation value is greater than a set threshold, the corresponding eye position image is determined to be a potential strabismus frame; if the binocular deviation value is less than or equal to the set threshold, the corresponding eye position image is determined to be an orthostatic frame.

[0076] As an improvement to the above scheme, to simplify the complex geometric calculations, other embodiments may also employ a pre-trained eye position recognition model to process the eye position images. The eye position images are first cropped to a suitable size and then input into the eye position recognition model. The model's classifier directly outputs the eye position confidence score for each frame of the eye position image, including the probability of orthotropia and the probability of strabismus. If the strabismus probability consistently exceeds a set threshold, it is directly determined as a strabismus attack event.

[0077] Finally, the strabismus duration of all strabismus attacks is calculated to obtain the total strabismus duration, which is the sum of the time segments of all confirmed strabismus attacks during the video playback process, thus achieving the statistical analysis of the strabismus attack duration throughout the entire playback process.

[0078] The eye position control assessment module 204 is used to analyze the fluctuation of the user's eye position control based on the user's total effective fixation time and the total strabismus time, and obtain the control assessment result.

[0079] In this embodiment of the application, the total duration of the video content playback is subtracted from the invalid frame time to obtain the user's effective gaze duration.

[0080] The ratio of total strabismus duration to total effective fixation duration is taken as the current strabismus time proportion in this test, and is abbreviated as STR. The specific calculation formula is as follows: STR = (Total duration of strabismus / Total duration of effective fixation) × 100%; Then, historical test data of intermittent strabismus is read using the user ID, and the standard deviation or coefficient of variation of STR is calculated by combining it with the current strabismus time ratio to determine the trend of the user's eye position control. The standard deviation represents the absolute fluctuation of eye position control, and the coefficient of variation represents the relative fluctuation of eye position control (the ratio of the standard deviation to the mean).

[0081] For example, the standard deviation or coefficient of variation of the STR values ​​for the most recent 7 days is calculated using historical test data and the current strabismus time ratio. If a user's STR value is 50%, it indicates severe strabismus, and the corresponding absolute value of 2% indicates stable strabismus; if a user's STR value is 2%, it indicates mild strabismus, and the corresponding absolute value of 2% indicates significant strabismus fluctuation.

[0082] In addition, historical test data can be combined to plot the recent strabismus time ratio fluctuation curve.

[0083] If the slope of the strabismus time-proportional fluctuation curve is greater than the set value, it is considered that the user's eye position control is declining and the user will be advised to have a follow-up examination; if the strabismus time-proportional fluctuation curve remains stable at a low level, it is considered that the eye position control is relatively stable.

[0084] Optionally, if the strabismus time-proportion fluctuation curve remains below 15% for a long period of time, and the single fluctuation amplitude does not exceed ±5%, then the eye position control is considered to be relatively stable.

[0085] In addition, for the quantitative assessment of eye position control, if the STR value is <1%, the eye position control is considered excellent; if 1% ≤ STR value < 15%, the eye position control is considered good; if 15% ≤ STR value < 50%, the eye position control is considered weakened; if the STR value is ≥ 50%, the eye position control is considered to be gradually lost, and a warning needs to be issued.

[0086] Therefore, through the above-mentioned quantitative evaluation methods, the control ability evaluation results of this test were obtained, realizing the objective quantification of the user's eye position control ability.

[0087] Implementing the embodiments of this application has the following beneficial effects: This application embodiment continuously stimulates the user's eyes with video content of interest to induce changes in eye position, allowing the user to be in a natural and relaxed state. Images are captured from both eyes separately, obtaining eye position images containing eye position information and physiological offset values ​​characterizing facial structural asymmetry and shooting deviations to eliminate errors in subsequent eye position deviation judgments. Then, the relative positions of the irises in the eye sockets are detected using the captured eye position images, achieving dynamic monitoring of eye position status and identifying image frames where strabismus attacks occur. Based on the number of image frames corresponding to strabismus attacks, the total duration of strabismus during video playback can be determined, providing objective evidence for assessing changes in eye position control. Finally, the fluctuation of eye position control is quantitatively assessed by the proportion of total strabismus duration to total effective fixation time, providing an objective and repeatable quantitative indicator for eye position control assessment. This comprehensively reflects the stability and variation patterns of the user's eye position control ability, providing strong data support for assessing the development trend of eye position control in users with intermittent strabismus and assisting doctors in developing personalized intervention plans tailored to the user.

[0088] Furthermore, Figure 3 This is a structural diagram of a terminal device provided in one embodiment of this application. Figure 3 As shown, the terminal device 3 of this embodiment includes: at least one processor 30 (in... Figure 3 (Only one is shown in the present invention) and a memory 31 and a computer program 32 stored in the memory 31 and executable on the at least one processor, wherein the processor 30 executes the computer program 32 to implement the steps of an eye position control assessment method for intermittent strabismus as described in any one of the embodiments of this application.

[0089] The terminal device 3 may be a computing device such as a desktop computer, a cloud server, or a laptop computer, and the computing device may include, but is not limited to, a processor 30 and a memory 31. Figure 3 This is merely an example of terminal device 3 and does not constitute a limitation on terminal device 3. It may include more or fewer components than those shown in the figure.

[0090] This application provides a storage medium that stores computer-readable program code, which, when executed, implements the steps of the above-described method for assessing eye position control for intermittent strabismus.

[0091] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this application. It should be understood that the above descriptions are merely specific embodiments of this application and are not intended to limit the scope of protection of this application. In particular, it should be noted that any modifications, equivalent substitutions, or improvements made by those skilled in the art within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for assessing eye position control in intermittent strabismus, characterized in that, include: While playing the preset video content, the camera device is used to capture images of the user's left and right eyes respectively, and obtain several frames of eye position images and the user's physiological offset. Based on the eye position image and the physiological bias, the binocular deviation value of each frame of the eye position image is obtained by calculating the relative position of the two irises in the eye socket. Based on the binocular deviation value, the potential strabismus frame is identified in the eye position image, and the strabismus attack events and total strabismus duration during playback are output. Based on the user's total effective fixation time and the total strabismus time, the fluctuation of the user's eye position control is analyzed to obtain the control ability assessment result.

2. The method for assessing eye position control for intermittent strabismus according to claim 1, characterized in that, When playing the preset video content, the camera device is used to capture images of the user's left and right eyes respectively, obtaining several frames of eye position images and the user's physiological offset. Specifically: Play the video content pre-selected by the user. Before the playback time reaches a first threshold, use a camera device to collect the user's binocular features and calculate the iris position ratio of the left and right eyes based on the binocular features. Before reaching the first threshold, the user's eyes are assumed to be in a normal position. The difference between the iris position ratio of the left eye and the iris position ratio of the right eye is used as the physiological bias. After the playback time reaches the first threshold, the camera device is used to capture images of the user's left and right eyes respectively to obtain the eye position images.

3. The method for assessing eye position control for intermittent strabismus according to claim 1, characterized in that, The binocular deviation value for each frame of the eye position image is obtained by calculating the relative positions of the irises of both eyes in the eye socket based on the eye position image and the physiological bias amount. Specifically: For each frame of the eye position image, the user's eyelid closure degree and head posture are detected, and the eye position images that fail the detection are marked as invalid frames; Invalid frames are removed from the eye position images to obtain valid image frames. For each valid image frame, the facial landmark algorithm is used to locate the user's eyes to obtain the first horizontal distance from the iris center to the inner canthus of the eye socket and the second horizontal distance from the inner canthus of the eye socket to the outer canthus of the eye socket. Based on the first horizontal distance and the second horizontal distance, calculate the ratio of the iris positions of both eyes in each effective image frame; The binocular deviation value is obtained based on the iris position ratio and the physiological offset.

4. The method for assessing eye position control for intermittent strabismus according to claim 3, characterized in that, The binocular deviation value is obtained based on the iris position ratio and the physiological offset, specifically as follows: Calculate the difference between the ratio of the iris positions of the left and right eyes in each effective image frame, subtract the physiological bias from the absolute value of the difference, and take the absolute value to obtain the binocular deviation value; The binocular deviation value is used to eliminate facial asymmetry of the user and systematic errors caused by the camera equipment during image acquisition.

5. The method for assessing eye position control for intermittent strabismus according to claim 1, characterized in that, The process involves identifying potential strabismus frames based on the binocular deviation value, and outputting the strabismus attack events and total strabismus duration during playback. Specifically: The binocular deviation value of each frame of the eye position image is judged using a preset sliding time window. If the binocular deviation value is greater than a second threshold, the corresponding eye position image is marked as a potential strabismus frame. If a third consecutive threshold of potential strabismus frames exists within the sliding time window, the strabismus attack event is considered to have occurred, and the strabismus duration of the strabismus attack event is recorded. The total duration of strabismus is obtained by counting all the strabismus durations.

6. The method for assessing eye position control for intermittent strabismus according to claim 1, characterized in that, Also includes: The cropped eye position image is input into a preset eye position state recognition model to calculate the eye position state confidence of each frame of the eye position image; wherein, the eye position state confidence includes the probability of orthotropia and the probability of strabismus. If the strabismus probability is greater than the fourth threshold, the corresponding eye position image is marked as the strabismus attack event.

7. The method for assessing eye position control for intermittent strabismus according to claim 3, characterized in that, The fluctuation of the user's eye position control is analyzed based on the user's total effective fixation time and total strabismus time to obtain the control ability assessment result, specifically: The total effective gaze duration is obtained by subtracting the duration of all invalid frames from the duration of playing the video content. The ratio of the total strabismus duration to the total effective fixation duration is taken as the current strabismus time ratio; Based on the user ID, the user's historical test data is read. Combining the historical test data with the current strabismus time ratio, the fluctuation of the user's strabismus time ratio in the current time period is calculated, and a strabismus time ratio fluctuation curve is plotted. Based on the strabismus time-proportion fluctuation curve, the changing trend of the user's eye position control is evaluated to obtain the control ability evaluation result.

8. A system for assessing eye position control in intermittent strabismus, characterized in that, include: The module includes a data acquisition module, a binocular deviation value calculation module, a strabismus event recognition module, and an eye position control assessment module. The data acquisition module is used to capture images of the user's left and right eyes using a camera device when playing preset video content, and obtain several frames of eye position images and the user's physiological offset. The binocular deviation value calculation module is used to calculate the binocular deviation value of each frame of eye position image by calculating the relative position of the two irises in the eye socket based on the eye position image and the physiological offset amount. The strabismus event recognition module is used to identify potential strabismus frames in the eye position image based on the binocular deviation value, and output the strabismus attack events and the total duration of strabismus during playback; The eye position control assessment module is used to analyze the fluctuation of the user's eye position control based on the user's total effective fixation time and the total strabismus time, and obtain the control ability assessment result.

9. A terminal device, characterized in that, It includes a processor and a memory, the memory storing a computer program, and the processor executing the computer program to implement the steps of the eye position control assessment method for intermittent strabismus as described in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium stores computer-readable program code that, when executed, implements the steps of the eye position control assessment method for intermittent strabismus as described in any one of claims 1 to 7.

Images